CN119904815B

CN119904815B - Landscape design recognition method and system based on neural network model

Info

Publication number: CN119904815B
Application number: CN202510377152.1A
Authority: CN
Inventors: 孔祥川; 原英东; 方坤; 刘周
Original assignee: Qingdao Municipal Engineering Design Institute Co ltd
Current assignee: Qingdao Municipal Engineering Design Institute Co ltd
Priority date: 2025-03-28
Filing date: 2025-03-28
Publication date: 2025-06-10
Anticipated expiration: 2045-03-28
Also published as: CN119904815A

Abstract

The application provides a landscape design identification method and system based on a neural network model. The method comprises the steps of obtaining landscape design data, lighting equipment data and energy supply network real-time load data in a target area, establishing an optical model to analyze vegetation shielding and artificial structure reflectivity to illuminate the space attenuation effect of light beams, calculating the energy consumption mapping relation between lamp power distribution and energy network load based on an energy transfer model, generating a dynamic dimming strategy by combining the space attenuation effect and the energy consumption mapping relation, coupling the dynamic dimming strategy with tourist distribution thermodynamic map and environment illumination intensity data acquired in real time, inputting the dynamic dimming strategy into a neural network model to perform iterative optimization, dynamically adjusting lamp power distribution, light beam angle coverage and light source color temperature parameters, achieving dynamic optimization of landscape lighting effect and minimization of energy consumption, and improving the visual experience and comfort of tourists. The application provides an intelligent dynamic optimization technical scheme for landscape lighting effect.

Description

Landscape design recognition method and system based on neural network model

Technical Field

The application relates to the technical field of data processing, in particular to a landscape design recognition method and system based on a neural network model.

Background

In modern urban landscape lighting projects, technical challenges are faced on how to efficiently manage energy consumption and optimize lighting effects. As urban scale increases and demand for environmentally friendly solutions increases, a technical solution is needed that can intelligently adjust lighting device parameters based on real-time environmental data as well as dynamically changing people stream densities. This solution not only needs to support basic control functions for the lighting device, but also has to be able to process complex input data sets to achieve accurate lighting regulation, ensure that an optimal lighting experience is provided in different usage scenarios and reduce unnecessary energy consumption as much as possible.

The existing scheme provides an intelligent lighting system based on a sensor network. The system monitors the light level and people flow conditions of the surrounding environment in real time by deploying a series of environmental sensors. Based on these data, the system can automatically adjust the operating state of the light fixture, such as reducing the brightness or changing the color temperature of the light source during periods of low traffic to conserve power.

However, intelligent lighting systems based on sensor networks rely on fixed algorithms or preset rules to make decisions, and lack sufficient adaptive capacity to cope with complex and varied practical application environments, which affects their overall performance and user satisfaction in different application scenarios.

Disclosure of Invention

The embodiment of the application provides a landscape design identification method and system based on a neural network model, which are used for solving the problems of large energy consumption and poor landscape lighting effect caused by poor accuracy of landscape lighting effect and energy consumption management in a complex and changeable practical application environment in the prior art.

In a first aspect, an embodiment of the present application provides a landscape design recognition method based on a neural network model, including:

Acquiring landscape design data, lighting equipment data and real-time load data of an energy supply network in a target area, wherein the landscape design data comprises vegetation shielding outline and artificial structure surface reflectivity, and the lighting equipment data comprises lamp power distribution, beam angle coverage and light source color temperature parameters;

establishing an optical model based on the landscape design data and the lighting equipment data, and analyzing the vegetation shading profile and the space attenuation effect of the light beam emitted by the lighting equipment through the optical model;

establishing an energy transfer model based on the lighting equipment data and the energy supply network real-time load data, and calculating an energy consumption mapping relation between the lamp power distribution and the energy supply network real-time load data through the energy transfer model;

And generating a dynamic dimming strategy by combining the space attenuation effect and the energy consumption mapping relation, coupling the dynamic dimming strategy with the guest distribution thermodynamic diagram and the environmental illumination intensity data acquired in real time to obtain a coupling result, and inputting the coupling result into a neural network model so as to iteratively optimize the lamp power distribution, the light beam angle coverage range and the light source color temperature parameter through the neural network model, thereby realizing the dynamic optimization of the landscape lighting effect and the minimization of energy consumption, and improving the visual experience and comfort of the guest.

Optionally, the generating a dynamic dimming policy by combining the spatial attenuation effect and the energy consumption mapping relationship, coupling the dynamic dimming policy with a guest distribution thermodynamic diagram and ambient light intensity data acquired in real time to obtain a coupling result, and inputting the coupling result into a neural network model to iteratively optimize the lamp power distribution, the beam angle coverage and the light source color temperature parameter through the neural network model, where the generating includes:

Constructing an objective function based on a mapping relation between a space attenuation effect and energy consumption, and generating a dynamic dimming strategy through an optimization algorithm based on the objective function;

Coupling the dynamic dimming strategy with a tourist distribution thermodynamic diagram and ambient illumination intensity data acquired in real time to obtain a coupling result, wherein the coupling result is a multidimensional feature vector;

Inputting the multidimensional feature vector into a neural network model, calculating a predicted value of the lighting effect through a forward propagation algorithm of the neural network model, and iteratively optimizing parameters of the power distribution, the coverage range of the light beam angle and the color temperature of the light source based on a preset loss function and a backward propagation algorithm of the neural network model;

Outputting the optimized lamp power distribution, the light beam angle coverage range and the light source color temperature parameters to a lighting control system so as to adjust the lighting equipment data in the target area in real time according to the optimized lamp power distribution, the light beam angle coverage range and the light source color temperature parameters.

Optionally, the inputting the multidimensional feature vector into a neural network model, calculating a predicted value of the lighting effect through a forward propagation algorithm of the neural network model, and iteratively optimizing parameters of the power distribution, the coverage range of the beam angle and the color temperature of the light source based on a preset loss function and a backward propagation algorithm of the neural network model, including:

carrying out standardization processing on the multidimensional feature vector to obtain a preprocessed multidimensional feature vector, wherein the multidimensional feature comprises high-density region coordinates and dynamic change trend of a tourist distribution thermodynamic diagram, time sequence features of environmental illumination intensity data, vegetation shielding contours and space distribution features of surface reflectivity of an artificial structure;

Inputting the preprocessed multidimensional feature vector into a neural network model, and generating a predicted value of the lighting effect at an output layer by combining a forward propagation algorithm in the neural network model and performing nonlinear transformation through a hidden layer of the neural network model;

And calculating a function value of a preset loss function based on the difference between the predicted value of the illumination effect and the preset target value, and iteratively optimizing and adjusting parameters of the power distribution, the light beam angle coverage and the color temperature of the light source by using the back propagation algorithm based on the function value until the function value of the loss function is converged to the preset threshold value.

Optionally, the inputting the preprocessed multidimensional feature vector into a neural network model, combining a forward propagation algorithm in the neural network model, performing nonlinear transformation through a hidden layer of the neural network model, and generating a predicted value of the lighting effect at an output layer, including:

Designing a structure of a neural network model, wherein the structure of the neural network model comprises an input layer, a hidden layer and an output layer, the node number of the input layer is consistent with the dimension of the multidimensional feature vector, the hidden layer adopts a multilayer structure, and the node number of the output layer is consistent with the dimension of the illumination effect predicted value;

Inputting the preprocessed multidimensional feature vector to the input layer, carrying out weighted summation on multidimensional features corresponding to the preprocessed multidimensional feature vector in the hidden layer by combining a forward propagation algorithm in the neural network model, carrying out nonlinear transformation on a weighted summation result through an activation function, and calculating the multidimensional feature mapping result layer by layer;

And in the output layer, carrying out weighted summation on the multidimensional feature mapping result output by the last hidden layer, and generating a lighting effect predicted value through an activation function of the output layer, wherein the lighting effect predicted value comprises illumination intensity, total energy consumption and visual comfort score of each position in a target area.

Optionally, the building an optical model based on the landscape design data and the lighting equipment data, analyzing the vegetation shading profile and the spatial attenuation effect of the light beam emitted by the lighting equipment by the artificial structure surface reflectivity through the optical model, including:

Performing digital processing on the vegetation shielding outline to obtain vegetation spatial distribution characteristics and shielding intensity parameters, and performing regional division on the reflectivity of the surface of the artificial structure to generate a reflectivity distribution diagram;

Generating an optical model based on the spatial distribution characteristics, the shielding intensity parameters and the reflectivity distribution map of the vegetation and combining the power distribution of the lamp, the light beam angle coverage range and the color temperature parameters of the light source;

Based on the optical model, simulating the propagation process of the light beam emitted by the lighting equipment in the target area, calculating the influence of vegetation shielding outline and the reflectivity of the surface of the artificial structure on the light beam, and generating attenuation parameters of the light beam;

And generating a spatial attenuation effect distribution map based on the attenuation parameters of the light beams, wherein the spatial attenuation effect distribution map is used for representing the spatial attenuation effect of the vegetation shading profile and the light beams emitted by the lighting equipment on the basis of the surface reflectivity of the artificial structure.

Optionally, based on the optical model, simulating a propagation process of a light beam emitted by the lighting device in the target area, calculating an influence of a vegetation shielding profile and a reflectivity of a surface of an artificial structure on the light beam, and generating attenuation parameters of the light beam, including:

constructing a light beam propagation path model according to the optical model, the lamp power distribution, the light beam angle coverage range and the light source color temperature parameter, and dividing the target area into a plurality of grid units based on the light beam propagation path model;

In the light beam propagation path model, calculating attenuation effects of the vegetation shielding outline on light beams by grid units, generating vegetation shielding attenuation coefficients, and updating illumination intensity values of each grid unit according to the vegetation shielding attenuation coefficients;

identifying grid cells intersecting with the surface of the artificial structure in the light beam propagation path model, calculating the reflection path and reflection intensity of the light beam on the surface of the artificial structure, and updating the illumination intensity value of the grid cells covered by the reflection path according to the reflection intensity;

and generating the light beam attenuation parameter of each grid cell according to the illumination intensity value of each grid cell.

Optionally, in the light beam propagation path model, calculating the attenuation effect of the vegetation shading profile on the light beam by grid units, generating a vegetation shading attenuation coefficient, and updating the illumination intensity value of each grid unit according to the vegetation shading attenuation coefficient, including:

extracting parameters of the height, density and distribution range of the vegetation from the vegetation shielding outline, and mapping the parameters of the height, density and distribution range of the vegetation into grid cells of a target area;

In the light beam propagation path model, calculating attenuation coefficients of light beams penetrating through vegetation according to mapping data corresponding to vegetation shielding intensity values, vegetation height, vegetation density and vegetation distribution range parameters respectively for each grid unit, and generating vegetation shielding attenuation coefficients of each grid unit;

Combining the scattering effect of vegetation on the light beams, simulating a scattering path of the light beams after penetrating through the vegetation, calculating the propagation direction and the intensity distribution of the scattering light beams, and updating the illumination intensity value of the grid cells covered by the scattering path according to the propagation direction and the intensity distribution;

and applying the vegetation shielding attenuation coefficient of each grid cell and the illumination intensity value of the grid cell covered by the scattering path to the light beam propagation path model so as to update the illumination intensity value of each grid cell through the light beam propagation path model.

In a second aspect, an embodiment of the present application provides a landscape design recognition system based on a neural network model, including:

The system comprises an acquisition module, a control module and a control module, wherein the acquisition module is used for acquiring landscape design data, lighting equipment data and energy supply network real-time load data in a target area, the landscape design data comprise vegetation shielding profiles and artificial structure surface reflectivity, and the lighting equipment data comprise lamp power distribution, light beam angle coverage and light source color temperature parameters;

The analysis module is used for establishing an optical model based on the landscape design data and the lighting equipment data, and analyzing the vegetation shielding outline and the space attenuation effect of the light beam emitted by the lighting equipment through the optical model;

The building calculation module is used for building an energy transfer model based on the lighting equipment data and the energy supply network real-time load data, and calculating an energy consumption mapping relation between the lamp power distribution and the energy supply network real-time load data through the energy transfer model;

The generation coupling module is used for generating a dynamic dimming strategy by combining the spatial attenuation effect and the energy consumption mapping relation, coupling the dynamic dimming strategy with the tourist distribution thermodynamic diagram and the environmental illumination intensity data acquired in real time to obtain a coupling result, and inputting the coupling result into a neural network model so as to iteratively optimize the lamp power distribution, the light beam angle coverage range and the light source color temperature parameter through the neural network model, thereby realizing the dynamic optimization of the landscape lighting effect and the minimization of energy consumption and improving the visual experience and comfort of the tourist.

In a third aspect, an embodiment of the present application provides a computing device, including a processing component and a storage component, where the storage component stores one or more computer instructions, and the one or more computer instructions are used to be invoked and executed by the processing component to implement a landscape design recognition method based on a neural network model according to any one of the first aspect.

In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program, where the computer program when executed by a computer implements a landscape design recognition method based on a neural network model according to any one of the first aspects.

The embodiment of the application provides a landscape design identification method based on a neural network model, which comprises the steps of obtaining landscape design data, lighting equipment data and energy supply network real-time load data in a target area, wherein the landscape design data comprises vegetation shielding profiles and artificial structure surface reflectivities, the lighting equipment data comprises lamp power distribution, beam angle coverage and light source color temperature parameters, establishing an optical model based on the landscape design data and the lighting equipment data, analyzing the space attenuation effect of light beams emitted by lighting equipment through the optical model, establishing an energy transfer model based on the lighting equipment data and the artificial structure surface reflectivities, calculating an energy consumption mapping relation between the lamp power distribution and the energy supply network real-time load data through the energy transfer model, generating a dynamic dimming strategy by combining the space attenuation effect and the energy consumption mapping relation, coupling the dynamic dimming strategy with the acquired passenger distribution thermodynamic diagram and environment illumination intensity data in real time to obtain a coupling result, and inputting the coupling result into the neural network model so as to optimize the lamp power distribution, the beam angle coverage and the light source color temperature parameters through the neural network model, and achieve the optimal dynamic energy consumption and the customer comfort experience.

According to the embodiment of the application, through obtaining the vegetation shielding profile, the surface reflectivity of the artificial structure, the power distribution of the lamp, the coverage range of the light beam angle, the color temperature parameter of the light source and the real-time load data of the energy supply network, a comprehensive data base is provided for subsequent modeling and optimization, and the regulation and control of the lighting system can be ensured to be accurately adapted to the environmental characteristics and the energy requirements. The vegetation shielding outline and the space attenuation effect of the illumination light beam are analyzed through the optical model, the influence of environmental factors on the illumination effect can be quantified, and a basis is provided for the generation of a follow-up dimming strategy. The energy consumption mapping relation between the power distribution of the lamp and the real-time load data of the energy supply network is calculated through the energy transfer model, so that the accurate prediction and optimization of the energy consumption of the lighting system can be realized, and support is provided for the efficient utilization of energy. By comprehensively considering the space attenuation effect and the energy consumption mapping relation, a dynamic dimming strategy is generated, the preliminary balance of the illumination effect and the energy consumption can be realized, and a foundation is provided for follow-up fine optimization. Through coupling visitor distribution thermodynamic diagram and ambient light intensity data, can dynamic adjustment illumination strategy, make it more laminate actual scene demand, promote visitor's visual experience and comfort level. Through iterative optimization of the neural network model, fine adjustment of illumination parameters can be realized, dynamic optimization of illumination effect and energy consumption are minimized, and meanwhile, the intelligent level of the system is improved. Further, through objective function construction, multidimensional feature vector generation and iterative optimization of a neural network model, dynamic fine adjustment of illumination parameters is achieved, the illumination effect and energy consumption are optimally balanced, meanwhile, the visual experience and comfort of tourists are improved, and the high efficiency and adaptability of intelligent illumination regulation are embodied.

These and other aspects of the application will be more readily apparent from the following description of the embodiments.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a flow chart of a landscape design recognition method based on a neural network model provided by an embodiment of the application;

FIG. 2 is a schematic structural diagram of a landscape design recognition system based on a neural network model according to an embodiment of the present application;

fig. 3 is a schematic structural diagram of a computing device according to an embodiment of the present application.

Detailed Description

In order to enable those skilled in the art to better understand the present application, the following description will make clear and complete descriptions of the technical solutions according to the embodiments of the present application with reference to the accompanying drawings.

In some of the flows described in the specification and claims of the present application and in the foregoing figures, a plurality of operations occurring in a particular order are included, but it should be understood that the operations may be performed out of order or performed in parallel, with the order of operations such as 11, 12, etc., being merely used to distinguish between the various operations, the order of the operations themselves not representing any order of execution. In addition, the flows may include more or fewer operations, and the operations may be performed sequentially or in parallel. It should be noted that, the descriptions of "first" and "second" herein are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, and are not limited to the "first" and the "second" being different types.

The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to fall within the scope of the application.

The problems of large energy consumption and poor landscape lighting effect caused by poor accuracy of landscape lighting effect and energy consumption management in complex and changeable practical application environments in the prior art are solved. The method specifically comprises the steps of establishing an optical model to analyze the light beam space attenuation effect by acquiring landscape design data, lighting equipment data and energy supply network real-time load data, and calculating an energy consumption mapping relation by an energy transfer model. And combining a dynamic dimming strategy with real-time environment data, iteratively optimizing lamp parameters by utilizing a neural network model, realizing the dynamic optimization of a landscape lighting effect and the minimization of energy consumption, and improving the visual experience and comfort of tourists.

Fig. 1 is a flowchart of a landscape design recognition method based on a neural network model according to an embodiment of the present application, as shown in fig. 1, the method includes:

S11, acquiring landscape design data, lighting equipment data and real-time load data of an energy supply network in a target area, wherein the landscape design data comprises vegetation shielding outline and artificial structure surface reflectivity, and the lighting equipment data comprises lamp power distribution, beam angle coverage and light source color temperature parameters.

The target area may be, among other things, a city park, city square, green road, business block, tourist attraction, industrial park, etc. The energy supply network real-time load data comprises electric parameters such as voltage, current, power factor and the like, and the real-time load condition of the whole network.

While vegetation cover profiles refer to the distribution of vegetation (e.g., trees, shrubs, etc.) in space and its blocking effect on light. Specifically, the vegetation cover profile includes parameters of vegetation height, density, distribution range, and the like, which together determine the attenuation degree of the light beam after penetrating through the vegetation. The vegetation shielding outline is digitally processed, so that the spatial distribution characteristics and shielding intensity parameters of the vegetation can be obtained, and data support is provided for quantifying the influence of the vegetation on intensity attenuation after light beam penetration. For example, in a night-time lighting scene, vegetation may result in partial areas that are under-lit, knowing their occlusion profile can help adjust the position or power of the light fixtures to compensate for this effect. And, the reflectivity of the surface of an artificial structure relates to the light reflecting capability of the surface material of an unnatural building object (such as a building, a sculpture, a road and the like). Different materials have different reflective properties, and therefore, when light of the same intensity is irradiated onto different surfaces, the intensity of the reflected light produced will also be different. The power distribution of the lamp refers to a power configuration scheme of each lighting device in different spatial positions in a target area, and specifically comprises a power value of a single lamp, a power adjustment range and a cooperative allocation relation of the power adjustment range in an overall lighting network.

S12, an optical model is established based on landscape design data and lighting equipment data, and the vegetation shading profile and the surface reflectivity of the artificial structure are analyzed through the optical model to achieve the spatial attenuation effect of the light beam emitted by the lighting equipment.

The optical model is used for simulating the light beam to start from the lamp, and the influence of factors such as vegetation shielding, artificial structure reflection and the like is considered in the process of transmitting the light beam in the target area. The spatial attenuation effect refers to a phenomenon that the intensity of a light beam emitted by a lighting device gradually decreases with the increase of a spatial distance due to the influence of environmental factors in the propagation process of the light beam. The spatial attenuation effect may be displayed by a spatial attenuation effect profile.

And S13, establishing an energy transfer model based on the lighting equipment data and the energy supply network real-time load data, and calculating an energy consumption mapping relation between the lamp power distribution and the energy supply network real-time load data through the energy transfer model.

The energy transfer model may include an electric power calculation formula, an active power calculation formula, and an energy consumption calculation formula. The energy transfer model can determine the power requirement of each lamp in the target area, such as the power value of a single lamp and the adjustment range of the power value according to the illumination design requirement and environmental conditions (such as vegetation shielding outline, artificial structure reflectivity and the like). And calculating the actual power consumption of each lamp by utilizing the electric power calculation formula and combining the working voltage and the working current of each lamp. If the power factor is considered, the active power formula is used for correction. The energy loss from the power source to the light fixture is calculated based on the wire resistance and other electrical characteristics on the power transmission path. And adding up the power consumption of all the lamps and adding up the loss in the power transmission process to obtain the total energy consumption of the whole lighting system. Combining the calculated total energy consumption with the real-time load data of the energy supply network to establish an energy consumption mapping relation between the power distribution of the lamp and the real-time load data of the energy supply network.

By way of example, the energy transfer model may be expressed using the following formula:

;

Wherein, N is the number of lamps in the target area, which is the total energy consumption; The actual power for the ith lamp, determined by the operating current of the lamp, Is the working voltage of the lamp and is the working voltage of the lamp,Is the internal resistance of the lamp, t is the working time length, M is the number of line sections on the power transmission path,I _j is the current of the j-th line on the power transmission path, which is the j-th line resistance on the power transmission path.

S14, generating a dynamic dimming strategy by combining the space attenuation effect and the energy consumption mapping relation, coupling the dynamic dimming strategy with the guest distribution thermodynamic diagram and the environmental illumination intensity data acquired in real time to obtain a coupling result, and inputting the coupling result into a neural network model to iteratively optimize the lamp power distribution, the beam angle coverage range and the light source color temperature parameters through the neural network model so as to realize the dynamic optimization of the landscape lighting effect and the minimization of energy consumption, and simultaneously improve the visual experience and the comfort level of the guest.

The tourist distribution thermodynamic diagram is used for reflecting the crowds in a specific area and the distribution situation. Such a chart typically uses different colors to represent people flow density in different areas, with darker colors (e.g., red or warm tone) representing more people flow in the area and lighter colors (e.g., blue or cool tone) representing less people flow. The coupling may be achieved by a data fusion technique. Ambient light intensity data is used to reflect the brightness level of natural or artificial light sources at a particular point in time for a particular location or area. The dynamic dimming strategy refers to an intelligent regulation scheme for dynamically regulating the power distribution of the lamp, the coverage range of the angle of the light beam and the color temperature parameters of the light source according to real-time environmental data (such as guest distribution, ambient illumination intensity, vegetation shielding outline, artificial structure reflectivity and the like) and preset optimization targets (such as energy consumption minimization and visual experience improvement).

In a landscape lighting system of an urban park, vegetation cover profiles and artificial structure surface reflectivity are obtained by laser Detection AND RANGING (Light Detection AND RANGING) scanning and spectral analysis, and lighting equipment data and energy load data are obtained from lamp manufacturers and smart meters. Based on the data, an optical model and an energy transfer model are established, and the spatial attenuation effect and the energy consumption mapping relation of the light beam are respectively analyzed. Then, combining the tourist distribution thermodynamic diagram acquired in real time (through an infrared sensor and a camera) and the ambient illumination intensity data, generating a dynamic dimming strategy and inputting the dynamic dimming strategy into a neural network model for optimization. Finally, the system dynamically adjusts the power, the beam angle and the color temperature of the lamp according to the optimization result, and provides comfortable visual experience for tourists while realizing energy-saving illumination at night.

By executing S11-S14, the embodiment of the application provides a comprehensive data base for subsequent modeling and optimization by acquiring vegetation shielding outline, artificial structure surface reflectivity, lamp power distribution, beam angle coverage, light source color temperature parameters and real-time load data of an energy supply network, and ensures that the regulation and control of a lighting system can accurately adapt to environmental characteristics and energy requirements. The vegetation shielding outline and the space attenuation effect of the illumination light beam are analyzed through the optical model, the influence of environmental factors on the illumination effect can be quantified, and a basis is provided for the generation of a follow-up dimming strategy. The energy consumption mapping relation between the power distribution of the lamp and the real-time load data of the energy supply network is calculated through the energy transfer model, so that the accurate prediction and optimization of the energy consumption of the lighting system can be realized, and support is provided for the efficient utilization of energy. By comprehensively considering the space attenuation effect and the energy consumption mapping relation, a dynamic dimming strategy is generated, the preliminary balance of the illumination effect and the energy consumption can be realized, and a foundation is provided for follow-up fine optimization. Through coupling visitor distribution thermodynamic diagram and ambient light intensity data, can dynamic adjustment illumination strategy, make it more laminate actual scene demand, promote visitor's visual experience and comfort level. Through iterative optimization of the neural network model, fine adjustment of illumination parameters can be realized, dynamic optimization of illumination effect and energy consumption are minimized, and meanwhile, the intelligent level of the system is improved.

In a possible embodiment, S14, generating a dynamic dimming policy by combining the spatial attenuation effect and the energy consumption mapping relationship, coupling the dynamic dimming policy with a guest distribution thermodynamic diagram and environmental illumination intensity data acquired in real time to obtain a coupling result, and inputting the coupling result into a neural network model to iteratively optimize the lamp power distribution, the beam angle coverage and the light source color temperature parameter through the neural network model, so as to achieve dynamic optimization of a landscape lighting effect and minimization of energy consumption, and simultaneously improve visual experience and comfort of a guest, including:

and 141, constructing an objective function based on the mapping relation between the space attenuation effect and the energy consumption, and generating a dynamic dimming strategy through an optimization algorithm based on the objective function.

Where the objective function typically contains a number of variables and constraints for minimizing or maximizing certain key metrics (e.g., energy consumption, visual comfort score, etc.). For example, in a landscape lighting scenario, the objective function may be designed to maximize the guest's visual experience while minimizing total energy consumption. Or in order to reduce the energy consumption as much as possible on the premise of keeping a certain illumination level, the objective function may comprehensively consider the factors such as the power distribution of the lamp, the coverage range of the beam angle, the color temperature parameter of the light source, the ambient illumination intensity and the like, and calculate the overall score through a weighted formula. The optimization algorithm may be a genetic algorithm, an ant colony algorithm, a gradient descent method, a simulated annealing algorithm, or the like. The dynamic dimming strategy is customized according to the characteristics of different areas (such as a pavement area, a sculpture area and the like), and the special dimming strategy is customized to ensure that each area can obtain a proper lighting effect.

The optimization algorithm may be multi-objective particle swarm optimization, and the objective function is Etotal-Scomfort, where Etotal is the energy consumption, scomfort is the visual comfort score, the energy consumption is the total energy consumption, and the total energy consumption can be calculated by the energy transfer model, and the visual comfort score is Σ|i _pred(x,y)−I_target(x,y)||²,I_pred (x, y) is the predicted illumination intensity, and I _target (x, y) is the target illumination intensity.

And 142, coupling the dynamic dimming strategy with the tourist distribution thermodynamic diagram and the ambient illumination intensity data acquired in real time to obtain a coupling result, wherein the coupling result is a multidimensional feature vector.

The multidimensional feature vector refers to integrating various different types of data (such as dynamic dimming strategies, guest distribution thermodynamic diagrams, ambient illumination intensity data, vegetation shielding profiles, artificial structure surface reflectivity and the like) into a high-dimensional mathematical representation form through a data fusion technology.

Step 143, inputting the multidimensional feature vector into a neural network model, calculating a predicted value of the illumination effect through a forward propagation algorithm of the neural network model, and iteratively optimizing parameters of the power distribution, the beam angle coverage and the color temperature of the light source of the lamp based on a preset loss function and a backward propagation algorithm of the neural network model.

The preset loss function is a mathematical expression used for quantifying the difference between the model predicted value and the real target value in the neural network training process.

Step 144, outputting the optimized lamp power distribution, the beam angle coverage and the light source color temperature parameters to the illumination control system, so as to adjust the illumination equipment data in the target area in real time according to the optimized lamp power distribution, the beam angle coverage and the light source color temperature parameters.

Where the beam angle coverage is the range of horizontal/vertical divergence angles formed in space by the light rays emitted by the illumination device, typically expressed in full or half angle form.

The following is a specific example:

In a landscape lighting system of an urban square, an objective function is firstly constructed based on a spatial attenuation effect and an energy consumption mapping relation, and a dynamic dimming strategy (such as pavement area power 60W and sculpture area power 120W) is generated. Subsequently, the dimming strategy is coupled with a real-time acquired tourist distribution thermodynamic diagram (dense pedestrian traffic in the pavement area) and environmental illumination intensity data (night illumination intensity of 0 lux) to generate a multidimensional feature vector. The multidimensional feature vectors are then input into a neural network model, and the luminaire parameters are iteratively optimized by forward and backward propagation algorithms (e.g., the pavement area power is increased to 70W, and the sculptured area power is decreased to 100W). Finally, the optimized parameters are output to an illumination control system, and parameters of the power of the lamp, the angle of the light beam and the color temperature of the light source are adjusted in real time, so that balance between energy-saving illumination and visual experience of tourists is realized.

By executing steps 141-144, the embodiment of the application realizes the intelligent dynamic dimming of the landscape lighting system through objective function construction, data coupling, neural network optimization and real-time control. By combining the space attenuation effect, the energy consumption mapping relation and the real-time environment data, the dynamic adaptability and the energy utilization efficiency of the lighting effect are improved, the visual experience and the comfort level of tourists are optimized, and an efficient and intelligent solution is provided for landscape lighting management in a smart city.

In one possible embodiment, step 143, inputting the multidimensional feature vector into the neural network model, calculating the illumination effect predicted value through the forward propagation algorithm of the neural network model, and iteratively optimizing the lamp power distribution, the beam angle coverage and the light source color temperature parameters based on the preset loss function and the backward propagation algorithm of the neural network model, includes:

Step a1, carrying out standardization processing on the multidimensional feature vector to obtain a preprocessed multidimensional feature vector, wherein the multidimensional feature comprises high-density area coordinates and dynamic change trend of a tourist distribution thermodynamic diagram, time sequence features of environmental illumination intensity data, vegetation shielding contours and spatial distribution features of surface reflectivity of an artificial structure.

The time sequence features may refer to statistical rules, periodic patterns, and dynamic trends presented by the change of the environmental illumination intensity data over time.

And a2, inputting the preprocessed multidimensional feature vector into a neural network model, and combining a forward propagation algorithm in the neural network model, performing nonlinear transformation through a hidden layer of the neural network model, and generating a predicted value of the lighting effect at an output layer.

Wherein, nonlinear transformation refers to a process of converting a linear combination of input data into nonlinear output by a mathematical function. In an embodiment of the application, this process occurs at the hidden layer of the neural network.

Illustratively, the forward propagation algorithm in the neural network model uses the following formula:

Z^(l)=W^(l)a^(l−1)+b^(l),a^(l)=σ(z^(l));

Wherein Z ^(l) is a predicted value of the lighting effect output by the neural network model, a ^(l−1) is a multidimensional feature vector extracted by the first layer-1, a ⁽⁰⁾ is a preprocessed multidimensional feature vector, l is the first layer in the neural network model, W ^(l) is the first layer weight, and b ^(l) is the first layer bias. Sigma is an activation function (e.g., reLU).

And a3, calculating a function value of a preset loss function based on the difference between the predicted value of the illumination effect and the preset target value, and iteratively and optimally adjusting the lamp power distribution, the beam angle coverage and the light source color temperature parameters through a back propagation algorithm based on the function value until the function value of the loss function is converged to a preset threshold value.

The convergence of the function value to the preset threshold value may mean that in the neural network training process, the model parameters are continuously adjusted through a back propagation algorithm, so that the value of the loss function gradually approaches a preset target value or target range, and finally, the value is stabilized in the target range, which indicates that the model training is completed.

By way of example, the preset loss function may employ the following formula:

;

Wherein, For the function value of the preset loss function,For the predicted energy consumption of the neural network,For the predicted intensity of illumination of the neural network,For the predicted comfort level of the neural network,In order to be a real energy consumption, the energy consumption,For the intensity of the real light to be illuminated,For the target comfort score, λ1, λ2, λ3 are weights corresponding to each loss term.

The following is a specific example:

In a landscape lighting system of an urban park, a tourist distribution thermodynamic diagram of a multidimensional feature vector processing and optimizing process is that the pedestrian flow in a pavement area is dense (density is 0.9), and the pedestrian flow in a sculpture area is sparse (density is 0.3). Ambient light intensity, night light intensity was 0 lux. Vegetation shielding outline, namely the pavement area shielding degree is 50%. Reflectivity data, namely, pavement area reflectivity 0.6 and sculpture area reflectivity 0.8. Normalization processing, namely normalizing all data to the [0,1] interval to generate a feature vector. And inputting the feature vector into a neural network model, and generating a predicted value of the lighting effect through forward propagation. And calculating a loss function value, and optimizing lamp parameters through a back propagation algorithm. Outputting the optimized parameters to a lighting control system, and adjusting the power of the lamp and the angle of the light beam in real time.

By executing the steps a 1-a 3, the embodiment of the application realizes the intelligent dynamic dimming of the landscape lighting system through the standardized processing of the multidimensional feature vector and the forward propagation and backward propagation optimization of the neural network model. By combining real-time environment data and a preset optimization target, the dynamic adaptability and the energy utilization efficiency of the lighting effect are improved, the visual experience and the comfort level of tourists are optimized, and an efficient and intelligent solution is provided for landscape lighting management in a smart city.

In a possible embodiment, step a2 inputs the preprocessed multidimensional feature vector into a neural network model, and combines a forward propagation algorithm in the neural network model to perform nonlinear transformation through a hidden layer of the neural network model, and generates a predicted value of the lighting effect at an output layer, which includes:

And b1, designing a structure of a neural network model, wherein the structure of the neural network model comprises an input layer, a hidden layer and an output layer, the node number of the input layer is consistent with the dimension of the multidimensional feature vector, the hidden layer adopts a multilayer structure, and the node number of the output layer is consistent with the dimension of the illumination effect predicted value.

The hidden layer is a parameterizable layer positioned between the input layer and the output layer in the neural network, and the core function of the hidden layer is to extract a high-order abstract representation from input characteristics through nonlinear transformation so as to capture a complex mapping relation between input data and an output target.

And b2, inputting the preprocessed multidimensional feature vector into an input layer, carrying out weighted summation on multidimensional features corresponding to the preprocessed multidimensional feature vector in a hidden layer by combining a forward propagation algorithm in a neural network model, carrying out nonlinear transformation on a weighted summation result through an activation function, and calculating a multidimensional feature mapping result layer by layer.

The neural network model is a machine learning model based on a multidimensional characteristic input-nonlinear transformation-output prediction framework, and the core function of the neural network model is to learn a complex mapping relation from input data through hierarchical abstraction and output optimized lighting equipment parameters.

And b3, in the output layer, carrying out weighted summation on the multidimensional feature mapping result output by the last hidden layer, and generating a lighting effect predicted value through an activation function of the output layer, wherein the lighting effect predicted value comprises illumination intensity, total energy consumption and visual comfort score of each position in the target area.

The total energy consumption refers to the total electric energy consumed by all lighting devices in the target area in a specific time period.

The following is a specific example of the design and prediction process of a neural network model in a landscape lighting system of an urban square. Input layer 8 nodes (corresponding to 8-dimensional feature vectors). And the hidden layer is 2 layers, namely 16 nodes and 8 nodes respectively. Output layer, 3 nodes (corresponding to illumination intensity, total energy consumption, visual comfort score). And inputting the feature vector to an input layer, and performing nonlinear transformation through a hidden layer to generate a multidimensional feature mapping result. And carrying out weighted summation on the hidden layer output to generate a predicted value of the lighting effect. The illumination intensity is [0.8, 0.6, 0.9], the total energy consumption is 0.7, and the visual comfort score is 0.85. And outputting the predicted value to a lighting control system, and adjusting the parameters of the lamp in real time.

By executing the steps b 1-b 3, the embodiment of the application realizes the intelligent prediction and optimization of the landscape lighting effect through the structural design of the neural network model, forward propagation calculation and output layer prediction. The multi-dimensional feature vector and the real-time environment data are combined, the dynamic adaptability and the energy utilization efficiency of the lighting system are improved, the visual experience and the comfort level of tourists are optimized, and an efficient and intelligent solution is provided for landscape lighting management in a smart city.

In one possible embodiment, S12, creating an optical model based on the landscape design data and the lighting device data, analyzing the vegetation shading profile and the spatial attenuation effect of the light beam emitted by the lighting device by the artificial structure surface reflectivity through the optical model, including:

And 121, performing digital processing on the vegetation shielding outline to obtain the spatial distribution characteristics and shielding intensity parameters of vegetation, and performing regional division on the reflectivity of the surface of the artificial structure to generate a reflectivity distribution diagram.

The reflectivity distribution diagram is a two-dimensional visualization diagram generated by a region division technology and is used for representing the reflectivity spatial distribution characteristics of the surface of the artificial structure in the target region. In the reflectance profile, higher reflectance means more light is reflected back into the environment, which is important to increase the brightness of certain areas or create a specific visual effect.

Step 122, generating an optical model based on the spatial distribution characteristics and the shielding intensity parameters of the vegetation and the reflectivity distribution map by combining the lamp power distribution, the beam angle coverage range and the light source color temperature parameters.

The optical model is a light propagation simulation framework constructed by a numerical simulation method based on landscape design data of a target area and parameters of lighting equipment.

Step 123, based on the optical model, simulating the propagation process of the light beam emitted by the lighting equipment in the target area, calculating the influence of the vegetation shielding profile and the reflectivity of the surface of the artificial structure, and generating the attenuation parameters of the light beam.

The attenuation parameter of the light beam is a physical quantity of intensity attenuation caused by vegetation shielding, artificial structure reflection and scattering effects when the light beam propagates in the target area through the optical model quantification. The key purpose of the dynamic dimming system is to provide accurate constraint of spatial light intensity change for a dynamic dimming strategy and ensure balance of lighting effect and energy consumption.

And 124, generating a spatial attenuation effect distribution diagram based on the attenuation parameters of the light beams, wherein the spatial attenuation effect distribution diagram is used for representing the vegetation shading profile and the spatial attenuation effect of the light beams emitted by the lighting equipment on the basis of the surface reflectivity of the artificial structure.

The space attenuation effect refers to the phenomenon that light intensity is non-linearly attenuated along with space distance or environmental characteristics due to physical effects such as vegetation shielding, artificial structure reflection and scattering in the propagation process of light beams emitted by lighting equipment.

The following is a specific example:

in a landscape lighting system of an urban park, vegetation shielding outline is obtained through LiDAR scanning in the construction and application process of an optical model, and shielding intensity parameters (such as pavement area shielding rate 50%) are calculated. The reflectivity of the artificial structure is measured by a spectrum analyzer to generate a reflectivity distribution map (such as the reflectivity of the pavement area is 0.6 and the reflectivity of the sculpture area is 0.8). And combining the vegetation shielding outline, the reflectivity distribution diagram and lamp parameters (such as power 60W and beam angle 90 degrees) to construct an optical model. And simulating the light beam propagation path, calculating the illumination intensity attenuation of the pavement area by 50%, and increasing the reflection illumination intensity of the sculpture area by 20%. Generating a distribution map, displaying the illumination intensity attenuation of the pavement area, and enhancing the reflection illumination intensity of the sculpture area. Outputting the distribution diagram to an illumination control system, adjusting the power of the lamp in the pavement area to 80W, and compensating the illumination intensity attenuation.

By executing the steps 121-124, the embodiment of the application generates a spatial attenuation effect distribution map through vegetation shading profile digitalization, reflectivity region division, optical model construction and light beam propagation simulation, and provides a basis for a dynamic dimming strategy of a landscape lighting system. By combining the real-time environment data and the optical model, the dynamic adaptability and the energy utilization efficiency of the lighting effect are improved, the visual experience and the comfort level of tourists are optimized, and an efficient and intelligent solution is provided for landscape lighting management in a smart city.

In one possible embodiment, step 123, based on the optical model, simulates the propagation process of the light beam emitted by the lighting device in the target area, calculates the influence of the vegetation cover profile and the reflectivity of the surface of the artificial structure, and generates the attenuation parameter of the light beam.

Step 1231, constructing a beam propagation path model according to the optical model, the lamp power distribution, the beam angle coverage and the light source color temperature parameters, and dividing the target area into a plurality of grid cells based on the beam propagation path model.

The beam propagation path model is a numerical calculation model for simulating the propagation process of a beam in a target area based on a gridding method. The method has the core functions of generating a beam attenuation parameter and a spatial attenuation effect distribution map by analyzing vegetation shielding, surface reflection and scattering effects on a grid unit by grid unit, and providing physical constraint for a dynamic dimming strategy.

In step 1232, in the beam propagation path model, the attenuation effect of the vegetation shading profile on the beam is calculated on a grid cell-by-grid cell basis, a vegetation shading attenuation coefficient is generated, and the illumination intensity value of each grid cell is updated according to the vegetation shading attenuation coefficient.

The vegetation shielding attenuation coefficient is a physical quantity for quantifying the intensity attenuation of vegetation after the vegetation penetrates light beams, and reflects the absorption, scattering and blocking effects of the characteristics of vegetation density, height and the like on light energy.

And step 1233, identifying grid cells intersecting with the surface of the artificial structure in the light beam propagation path model, calculating the reflection path and the reflection intensity of the light beam on the surface of the artificial structure, and updating the illumination intensity value of the grid cells covered by the reflection path according to the reflection intensity.

The illumination intensity value of the grid cell covered by the reflection path refers to the illumination intensity of the grid cell transmitted to the target area through the reflection path after the light beam is reflected on the surface of the artificial structure. The method is used for calculating the illumination intensity value of the grid unit covered by the reflection path, which is needed to be combined with the initial intensity of the light beam, the surface reflectivity, the incident angle, the effective area of the reflection surface, the propagation distance of the reflection path and the like, and is used for correcting the influence of the secondary light path in the light beam propagation model on the illumination distribution.

Illustratively, the equation used to calculate the reflected intensity of a light beam at the surface of an artificial structure is as follows:

I(x,y)=I₀×γ(x,y)×cosθ(x,y)×[Ar(x,y)/r2(x,y)];

wherein I (x, y) is the reflection intensity of the light beam on the surface of the artificial structure, I ₀ is the initial intensity of the incident light beam, gamma (x, y) is the surface reflectivity, theta (x, y) is the incident angle, ar (x, y) reflects the effective area of the surface, and r (x, y) is the transmission distance of the reflection path, and is in meters.

Step 1234, generating a beam attenuation parameter for each grid cell according to the illumination intensity value of each grid cell.

The light beam attenuation parameter is a quantitative index for representing the intensity attenuation of the light beam in the process of propagation in the target area after comprehensive vegetation shielding, artificial structure reflection, scattering effect and other factors.

The following is a specific example:

In a landscape lighting system of an urban square, a beam propagation path model is constructed according to lamp parameters (power 60W, beam angle 90 DEG) in the construction and application process of the beam propagation path model, and a target area is divided into grid cells of 1m multiplied by 1 m. The vegetation cover attenuation coefficient (such as 0.5) of the pavement area grid unit is calculated, and the illumination intensity value is updated (from 100 lux to 50 lux). The grid cells intersecting the sculptured surface are identified, the reflection path and reflection intensity (e.g., reflectance 0.8, reflection intensity 40 lux) are calculated, and the illumination intensity value of the reflection path covering the grid cells is updated (from 50 lux to 90 lux). The beam attenuation parameters (such as walk area attenuation rate 50%, sculptured area attenuation rate 10%) of each grid cell are calculated.

By executing the steps 1231-1234, the embodiment of the application realizes the accurate modeling and attenuation effect quantification of the light beam propagation path in the landscape lighting system through the construction of the light beam propagation path model, the vegetation shielding attenuation effect calculation, the reflection effect calculation and the light beam attenuation parameter generation. By combining the real-time environment data and the optical model, the dynamic adaptability and the energy utilization efficiency of the lighting effect are improved, the visual experience and the comfort level of tourists are optimized, and an efficient and intelligent solution is provided for landscape lighting management in a smart city.

In one possible embodiment, step 1232, in the beam propagation path model, calculating the attenuation effect of the vegetation shading profile on the beam from grid cell to grid cell, generating a vegetation shading attenuation coefficient, and updating the illumination intensity value of each grid cell according to the vegetation shading attenuation coefficient, includes:

and c1, extracting parameters of the height, the density and the distribution range of the vegetation from the vegetation shielding outline, and mapping the parameters of the height, the density and the distribution range of the vegetation into grid cells of a target area.

Step c2, in the light beam propagation path model, calculating attenuation coefficients of light beams penetrating through vegetation according to mapping data corresponding to vegetation shielding intensity values, vegetation height, vegetation density and vegetation distribution range parameters in each grid unit, and generating vegetation shielding attenuation coefficients of each grid unit.

Illustratively, the attenuation coefficient of the beam after penetrating vegetation is calculated as follows:

α(x,y)=1−exp(−β×h(x,y)×ρ(x,y)×d(x,y));

Where α (x, y) is a vegetation shading attenuation coefficient of the grid unit (x, y), β is a vegetation shading intensity value, which can be understood as a vegetation type coefficient, and exemplary, arbor β=0.05, shrub β=0.03, h (x, y) is mapping data corresponding to vegetation height of the grid unit (x, y), ρ (x, y) is mapping data corresponding to density in meters, and d (x, y) is mapping data corresponding to a distribution range parameter, for example, a vegetation distribution range duty ratio.

The distribution range parameter refers to a parameter set for describing the spatial distribution characteristics of vegetation in a target area, and comprises a horizontal distribution range and a vertical distribution range.

And c3, applying the vegetation shielding attenuation coefficient of each grid cell and the illumination intensity value of the grid cell covered by the scattering path to the light beam propagation path model so as to update the illumination intensity value of each grid cell through the light beam propagation path model.

The attenuation coefficient after the light beam penetrates through the vegetation is a parameter for quantifying the intensity attenuation after the light beam penetrates through the vegetation and reflects the absorption and scattering effects of the vegetation density, height and scattering characteristics on the light energy.

The following is a specific example:

In a landscape lighting system of an urban park, vegetation height (2 m), density (0.6) and distribution range (50%) are obtained through LiDAR scanning in the calculating and applying process of vegetation shielding attenuation effect and are mapped into grid cells of 1m multiplied by 1 m. The occlusion intensity value (e.g., 0.6) of the pavement area grid cell is calculated to generate an attenuation coefficient (e.g., 0.45). The illumination intensity values of the pavement area grid cells are updated according to the attenuation coefficients (from 100 lux down to 55 lux). The scattering path of the beam in the vegetation is calculated and the illumination intensity value of the scattering path covering the grid cell is updated (from 55 lux to 70 lux).

By executing the steps c 1-c 3, the embodiment of the application realizes accurate modeling and quantification of vegetation shading effect in a landscape lighting system through vegetation parameter extraction and mapping, vegetation shading attenuation coefficient calculation and illumination intensity updating. By combining the light beam propagation path model and real-time environment data, the dynamic adaptability and the energy utilization efficiency of the lighting effect are improved, the visual experience and the comfort level of tourists are optimized, and an efficient and intelligent solution is provided for landscape lighting management in a smart city.

Fig. 2 is a schematic structural diagram of a landscape design recognition system based on a neural network model according to an embodiment of the present application, as shown in fig. 2, the system includes:

The acquisition module 21 is configured to acquire landscape design data, lighting equipment data and real-time load data of an energy supply network in a target area, where the landscape design data includes vegetation shielding profiles and surface reflectivity of an artificial structure, and the lighting equipment data includes lamp power distribution, beam angle coverage and light source color temperature parameters.

The analysis module 22 is configured to establish an optical model based on the landscape design data and the lighting device data, and analyze the vegetation shading profile and the surface reflectivity of the artificial structure through the optical model to obtain the spatial attenuation effect of the light beam emitted by the lighting device.

The building calculation module 23 is configured to build an energy transfer model based on the lighting device data and the energy supply network real-time load data, and calculate an energy consumption mapping relationship between the luminaire power distribution and the energy supply network real-time load data through the energy transfer model.

The generating coupling module 24 is configured to generate a dynamic dimming policy by combining a spatial attenuation effect and an energy consumption mapping relationship, couple the dynamic dimming policy with a guest distribution thermodynamic diagram and environmental illumination intensity data acquired in real time to obtain a coupling result, and input the coupling result into a neural network model to iteratively optimize parameters of lamp power distribution, beam angle coverage and light source color temperature through the neural network model, so as to achieve dynamic optimization of a landscape lighting effect and minimization of energy consumption, and improve visual experience and comfort of the guest.

The landscape design recognition system based on the neural network model shown in fig. 2 may execute the landscape design recognition method based on the neural network model shown in the embodiment shown in fig. 1, and its implementation principle and technical effects are not repeated. The specific manner in which the various modules and units perform operations in the neural network model-based landscape design recognition system in the above embodiments has been described in detail in the embodiments related to the method, and will not be described in detail here.

In one possible design, the neural network model-based landscape design recognition system of the embodiment shown in FIG. 2 may be implemented as a computing device that may include a storage component 31 and a processing component 32, as shown in FIG. 3.

The storage component 31 stores one or more computer instructions for execution by the processing component 32.

The processing component 32 is configured to obtain landscape design data, lighting equipment data and real-time energy supply network load data in a target area, wherein the landscape design data comprises vegetation shading profiles and artificial structure surface reflectances, the lighting equipment data comprises lamp power distribution, beam angle coverage and light source color temperature parameters, establish an optical model based on the landscape design data and the lighting equipment data, analyze the spatial attenuation effect of the light beams emitted by lighting equipment through the vegetation shading profiles and the artificial structure surface reflectances by the optical model, establish an energy transfer model based on the lighting equipment data and the real-time energy supply network load data, calculate an energy consumption mapping relation between the lamp power distribution and the real-time energy supply network load data by the energy transfer model, generate a dynamic dimming strategy by combining the spatial attenuation effect and the energy consumption mapping relation, input the dynamic dimming strategy into a neural network model to obtain coupling results, and iteratively optimize the power distribution, the beam angle coverage and the light source color temperature parameters by the neural network model so as to achieve dynamic optimization of the lighting effect and minimization of energy consumption, and improve the visual experience and comfort of tourists.

Wherein the processing component 32 may include one or more processors to execute computer instructions to perform all or part of the steps of the methods described above. Of course, the processing component may also be implemented as one or more Application-specific integrated circuits (ASICs), digital signal processors (DIGITAL SIGNAL processes, DSPs), digital signal processing devices (DIGITAL SIGNAL Process devices, DSPDs), programmable logic devices (Programmable Logic Device, PLDs), field programmable gate arrays (Field Programmable GATE ARRAY, FPGA), controllers, microcontrollers, microprocessors, or other electronic elements for performing the above method.

The storage component 31 is configured to store various types of data to support operations at the terminal. The Memory component may be implemented by any type or combination of volatile or nonvolatile Memory devices such as Random Access Memory (Random Access Memory, RAM), static Random-Access Memory (SRAM), electrically erasable programmable Read-Only Memory (EEPROM), erasable programmable Read-Only Memory (Erasable Programmable Read Only Memory, EPROM), programmable Read-Only Memory (Programmable Read Only Memory, PROM), read Only Memory (ROM), magnetic Memory, flash Memory, magnetic or optical disk.

Of course, the computing device may necessarily include other components as well, such as input/output interfaces, display components, communication components, and the like.

The input/output interface provides an interface between the processing component and a peripheral interface module, which may be an output device, an input device, etc.

The communication component is configured to facilitate wired or wireless communication between the computing device and other devices, and the like.

The computing device may be a physical device or an elastic computing host provided by the cloud computing platform, and at this time, the computing device may be a cloud server, and the processing component, the storage component, and the like may be a base server resource rented or purchased from the cloud computing platform.

The embodiment of the application also provides a computer storage medium which stores a computer program, and the computer program can realize the landscape design recognition method based on the neural network model in the embodiment shown in the figure 1 when being executed by a computer.

It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, which are not repeated herein.

The apparatus embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.

From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.

It should be noted that the above-mentioned embodiments are merely for illustrating the technical solution of the present application, and not for limiting the same, and although the present application has been described in detail with reference to the above-mentioned embodiments, it should be understood by those skilled in the art that the technical solution described in the above-mentioned embodiments may be modified or some technical features may be equivalently replaced, and these modifications or substitutions do not make the essence of the corresponding technical solution deviate from the spirit and scope of the technical solution of the embodiments of the present application.

Claims

1. A landscape design recognition method based on a neural network model, characterized by comprising:

Acquire landscape design data, lighting equipment data and real-time load data of the energy supply network in the target area, wherein the landscape design data includes vegetation shading outline and artificial structure surface reflectivity, and the lighting equipment data includes lamp power distribution, beam angle coverage and light source color temperature parameters;

An optical model is established based on the landscape design data and the lighting equipment data, and the spatial attenuation effect of the vegetation shielding contour and the artificial structure surface reflectivity on the light beam emitted by the lighting equipment is analyzed by the optical model;

Establishing an energy transfer model based on the lighting equipment data and the real-time load data of the energy supply network, and calculating the energy consumption mapping relationship between the power distribution of the lamps and the real-time load data of the energy supply network through the energy transfer model;

Generate a dynamic dimming strategy by combining the spatial attenuation effect with the energy consumption mapping relationship, couple the dynamic dimming strategy with the real-time collected tourist distribution heat map and ambient light intensity data to obtain a coupling result, and input the coupling result into a neural network model to iteratively optimize the lamp power distribution, the beam angle coverage range and the light source color temperature parameters through the neural network model, so as to achieve dynamic optimization of landscape lighting effects and minimization of energy consumption, while improving the visual experience and comfort of tourists;

The optical model is established based on the landscape design data and the lighting equipment data, and the spatial attenuation effect of the vegetation shielding contour and the artificial structure surface reflectivity on the light beam emitted by the lighting equipment is analyzed by the optical model, including:

The vegetation shielding contour is digitally processed to obtain the spatial distribution characteristics and shielding intensity parameters of the vegetation, and the surface reflectivity of the artificial structure is divided into regions to generate a reflectivity distribution map;

Based on the spatial distribution characteristics and shading intensity parameters of the vegetation, and the reflectivity distribution diagram, combined with the lamp power distribution, the beam angle coverage range and the light source color temperature parameters, an optical model is generated;

Based on the optical model, the propagation process of the light beam emitted by the lighting device in the target area is simulated, the influence of the vegetation blocking contour and the reflectivity of the surface of the artificial structure on the light beam is calculated, and the attenuation parameter of the light beam is generated;

Based on the attenuation parameters of the light beam, a spatial attenuation effect distribution diagram is generated, and the spatial attenuation effect distribution diagram is used to characterize the spatial attenuation effect of the vegetation shielding contour and the artificial structure surface reflectivity on the light beam emitted by the lighting device.

2. The method according to claim 1 is characterized in that the dynamic dimming strategy is generated by combining the spatial attenuation effect with the energy consumption mapping relationship, the dynamic dimming strategy is coupled with the real-time collected tourist distribution heat map and ambient light intensity data to obtain a coupling result, and the coupling result is input into a neural network model to iteratively optimize the lamp power distribution, the beam angle coverage range and the light source color temperature parameters through the neural network model, including:

Building an objective function based on the mapping relationship between the spatial attenuation effect and the energy consumption, and generating a dynamic dimming strategy through an optimization algorithm based on the objective function;

The dynamic dimming strategy is coupled with the real-time collected tourist distribution heat map and ambient light intensity data to obtain a coupling result, wherein the coupling result is a multi-dimensional feature vector;

The multidimensional feature vector is input into a neural network model, a lighting effect prediction value is calculated by a forward propagation algorithm of the neural network model, and the lamp power distribution, beam angle coverage and light source color temperature parameters are iteratively optimized based on a preset loss function and a back propagation algorithm of the neural network model;

The optimized lamp power distribution, beam angle coverage and light source color temperature parameters are output to the lighting control system to adjust the lighting equipment data in the target area in real time according to the optimized lamp power distribution, beam angle coverage and light source color temperature parameters.

3. The method according to claim 2, characterized in that the multidimensional feature vector is input into a neural network model, a lighting effect prediction value is calculated by a forward propagation algorithm of the neural network model, and a lamp power distribution, a beam angle coverage range and a light source color temperature parameter are iteratively optimized based on a preset loss function and a back propagation algorithm of the neural network model, comprising:

The multidimensional feature vector is standardized to obtain a preprocessed multidimensional feature vector, wherein the multidimensional features include the high-density area coordinates and dynamic change trends of the tourist distribution heat map, the time series characteristics of the ambient light intensity data, and the spatial distribution characteristics of the vegetation shading contour and the surface reflectivity of the artificial structure;

Inputting the preprocessed multidimensional feature vector into a neural network model, combining the forward propagation algorithm in the neural network model, performing nonlinear transformation through the hidden layer of the neural network model, and generating a lighting effect prediction value at the output layer;

Based on the difference between the lighting effect prediction value and the preset target value, the function value of the preset loss function is calculated. Based on the function value, the lamp power distribution, beam angle coverage and light source color temperature parameters are iteratively optimized and adjusted through the back propagation algorithm until the function value of the loss function converges to a preset threshold.

4. The method according to claim 3, characterized in that the preprocessed multidimensional feature vector is input into a neural network model, combined with a forward propagation algorithm in the neural network model, a nonlinear transformation is performed through a hidden layer of the neural network model, and a lighting effect prediction value is generated at an output layer, comprising:

Designing a structure of a neural network model, wherein the structure of the neural network model includes an input layer, a hidden layer and an output layer, the number of nodes in the input layer is consistent with the dimension of the multidimensional feature vector, the hidden layer adopts a multi-layer structure, and the number of nodes in the output layer is consistent with the dimension of the lighting effect prediction value;

The preprocessed multidimensional feature vector is input into the input layer, and in combination with the forward propagation algorithm in the neural network model, the multidimensional features corresponding to the preprocessed multidimensional feature vector are weighted summed in the hidden layer, and the weighted summation result is nonlinearly transformed by an activation function, and the multidimensional feature mapping result is calculated layer by layer;

In the output layer, the multidimensional feature mapping results output by the last hidden layer are weighted and summed, and the lighting effect prediction value is generated through the activation function of the output layer. The lighting effect prediction value includes the light intensity, total energy consumption and visual comfort score of each location in the target area.

5. The method according to claim 1, characterized in that the step of simulating the propagation process of the light beam emitted by the lighting device in the target area based on the optical model, calculating the influence of the vegetation blocking contour and the surface reflectivity of the artificial structure on the light beam, and generating the attenuation parameter of the light beam comprises:

Constructing a beam propagation path model according to the optical model, the lamp power distribution, the beam angle coverage range and the light source color temperature parameter, and dividing the target area into a plurality of grid units based on the beam propagation path model;

In the light beam propagation path model, the attenuation effect of the vegetation shading contour on the light beam is calculated grid unit by grid unit, a vegetation shading attenuation coefficient is generated, and the light intensity value of each grid unit is updated according to the vegetation shading attenuation coefficient;

Identify the grid cells that intersect with the surface of the artificial structure in the light beam propagation path model, calculate the reflection path and reflection intensity of the light beam on the surface of the artificial structure, and update the light intensity value of the grid cells covered by the reflection path according to the reflection intensity;

According to the light intensity value of each grid cell, a beam attenuation parameter of each grid cell is generated.

6. The method according to claim 5, characterized in that, in the light beam propagation path model, the attenuation effect of the vegetation occlusion contour on the light beam is calculated grid unit by grid unit, a vegetation occlusion attenuation coefficient is generated, and the light intensity value of each grid unit is updated according to the vegetation occlusion attenuation coefficient, comprising:

Extracting the height, density and distribution range parameters of vegetation from the vegetation occlusion contour, and mapping the height, density and distribution range parameters of vegetation to the grid cells of the target area;

In the beam propagation path model, for each grid unit, according to the vegetation shielding intensity value, the mapping data corresponding to the vegetation height, density and distribution range parameters, the attenuation coefficient of the beam after penetrating the vegetation is calculated to generate the vegetation shielding attenuation coefficient of each grid unit;

Combined with the scattering effect of vegetation on the light beam, the scattering path of the light beam after penetrating the vegetation is simulated, the propagation direction and intensity distribution of the scattered light beam are calculated, and the light intensity value of the grid unit covered by the scattering path is updated according to the propagation direction and the intensity distribution;

The vegetation shading attenuation coefficient of each grid unit and the light intensity value of the grid unit covered by the scattering path are applied to the light beam propagation path model to update the light intensity value of each grid unit through the light beam propagation path model.

7. A landscape design recognition system based on a neural network model, characterized by comprising:

An acquisition module is used to acquire landscape design data, lighting equipment data and real-time load data of the energy supply network in the target area, wherein the landscape design data includes vegetation shading contours and artificial structure surface reflectivity, and the lighting equipment data includes lamp power distribution, beam angle coverage and light source color temperature parameters;

Establishing an analysis module for establishing an optical model based on the landscape design data and the lighting equipment data, and analyzing the spatial attenuation effect of the vegetation shielding contour and the artificial structure surface reflectivity on the light beam emitted by the lighting equipment through the optical model;

Establishing a calculation module, for establishing an energy transfer model based on the lighting equipment data and the real-time load data of the energy supply network, and calculating the energy consumption mapping relationship between the power distribution of the lamps and the real-time load data of the energy supply network through the energy transfer model;

Generate a coupling module for combining the spatial attenuation effect with the energy consumption mapping relationship to generate a dynamic dimming strategy, couple the dynamic dimming strategy with the real-time collected tourist distribution heat map and ambient light intensity data to obtain a coupling result, and input the coupling result into a neural network model to iteratively optimize the lamp power distribution, the beam angle coverage range and the light source color temperature parameters through the neural network model to achieve dynamic optimization of landscape lighting effects and minimization of energy consumption, while improving the visual experience and comfort of tourists;

Wherein, the establishment of the parsing module is also used for:

8. A computing device, characterized in that it includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a landscape design recognition method based on a neural network model as described in any one of claims 1 to 6.

9. A computer storage medium, characterized in that a computer program is stored therein, and when the computer program is executed by a computer, the landscape design recognition method based on a neural network model as described in any one of claims 1 to 6 is implemented.