CN120177408A - A digging shovel for real-time detection of soil organic matter content - Google Patents
A digging shovel for real-time detection of soil organic matter content Download PDFInfo
- Publication number
- CN120177408A CN120177408A CN202510660814.6A CN202510660814A CN120177408A CN 120177408 A CN120177408 A CN 120177408A CN 202510660814 A CN202510660814 A CN 202510660814A CN 120177408 A CN120177408 A CN 120177408A
- Authority
- CN
- China
- Prior art keywords
- soil
- organic matter
- matter content
- data
- real
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/3563—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing solids; Preparation of samples therefor
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N1/00—Sampling; Preparing specimens for investigation
- G01N1/02—Devices for withdrawing samples
- G01N1/04—Devices for withdrawing samples in the solid state, e.g. by cutting
- G01N1/08—Devices for withdrawing samples in the solid state, e.g. by cutting involving an extracting tool, e.g. core bit
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/359—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using near infrared light
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/47—Scattering, i.e. diffuse reflection
- G01N21/4738—Diffuse reflection, e.g. also for testing fluids, fibrous materials
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
- G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
- G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
- G01S19/42—Determining position
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Pathology (AREA)
- Immunology (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Biochemistry (AREA)
- Engineering & Computer Science (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
本发明的提供一种实时检测土壤有机质含量的挖掘铲，包括可拆卸式挖掘铲，前端为挖掘刃，后端设有传感器集成腔；数据采集与通信模块，集成于所述挖掘铲内部，包括近红外光谱传感器，采用V型光纤漫反射结构采集土壤反射率；高分辨率摄像头，用于采集土壤表面纹理图像；GPS模块，记录采样点经纬度及高程数据；主控单元，作为核心控制器协调数据采集、处理与通信；数据处理单元基于光谱与图像信息融合实现精准的土壤有机质含量数据，帮助工作者实时掌握土壤状况，提高挖掘生产效率。
The present invention provides an excavating shovel for real-time detection of soil organic matter content, comprising a detachable excavating shovel with an excavating blade at the front end and a sensor integrated cavity at the rear end; a data acquisition and communication module integrated inside the excavating shovel, comprising a near-infrared spectrum sensor, which adopts a V-shaped optical fiber diffuse reflection structure to collect soil reflectivity; a high-resolution camera for collecting soil surface texture images; a GPS module for recording the latitude and longitude and elevation data of the sampling point; a main control unit, which serves as a core controller to coordinate data acquisition, processing and communication; and a data processing unit to realize accurate soil organic matter content data based on the fusion of spectrum and image information, so as to help workers grasp the soil condition in real time and improve the excavation production efficiency.
Description
Technical Field
The invention relates to the technical field of agricultural machinery, and provides a digging shovel for detecting the organic matter content of soil in real time.
Background
The organic matter content of the soil is a core index for measuring the soil fertility, and the growth efficiency of crops is directly affected. The traditional detection method (such as a high-temperature combustion method and a wet acidolysis method) needs to rely on laboratory chemical analysis (refer to national standards GB 9834-1988, NY/T1121.6-2006 and the like), and has the obvious defects that firstly, a sample needs to be pretreated, chemically decomposed and detected by an instrument, and the single detection period is as long as several hours to days, so that the field real-time operation requirement cannot be met. 2. The skilled technician is required to operate a precision instrument (such as an elemental analyzer), the equipment maintenance cost is high, and the equipment is not suitable for a non-laboratory environment. 3. Laboratory detection only supports discrete point sampling, and a continuous soil organic matter distribution map is difficult to generate, so that a fertilization decision lacks global data support.
In recent years, near infrared spectroscopy (NIRS) technology and image recognition technology are introduced into the field of soil detection, but the technical bottleneck is that the existing portable spectrometer only supports single spectrum data acquisition and is not integrated with a high-resolution image sensor, a GPS module and an agricultural implement adaptation structure, so that the field operation efficiency is low. For example, an operator needs to manually collect samples and detect the samples independently after the operation of an agricultural machine, an integrated process of 'operation while analysis' cannot be realized, a data model is single, a linear regression model (such as partial least squares regression PLSR) is adopted to process spectrum data in the prior art, multi-source information such as soil surface textures (such as gray level co-occurrence matrix characteristics) and color distribution (RGB histogram statistics) is not fused, the predicted error rate is generally equal to or more than 8%, the precision agricultural requirements are difficult to meet, the three-instantaneity is insufficient, the end-to-end response time of a scheme relying on a cloud server for data processing is more than 10 seconds, the real-time operation decision (such as dynamic adjustment of fertilization amount) of the agricultural machine cannot be supported, the four-dimensional environment is poor, the high-dust and high-humidity environment is prone to cause sensor pollution, the existing equipment protection design (such as a bag filter) needs to be cleaned frequently (every 2 hours/times), the equipment reliability is not optimized for the working conditions such as vibration and impact of the agricultural machine, the existing agricultural machine integrated sensor (such as soil humidity and temperature sensor) is mostly an independent module, the multi-mode data fusion architecture is lack, and the dust-proof efficiency is low in the traditional design (such as 92% dust-proof dust removal scheme is not required). The problems seriously restrict the large-scale application of the soil organic matter detection technology in agricultural mechanized operation.
In summary, the prior art has not solved the core problems of low field real-time detection precision, weak multisource data fusion capability, poor environmental adaptability of an embedded system of an agricultural implement and the like, and an innovative scheme with laboratory-level precision and field operation efficiency is needed.
Disclosure of Invention
The invention aims to provide the excavating shovel for detecting the organic matter content of the soil in real time, which realizes accurate soil organic matter content data based on the fusion of spectrum and image information, grasps the soil condition in real time in actual operation, further makes more scientific and reasonable fertilization decisions, realizes the improvement of production efficiency, avoids resource waste and realizes more accurate soil nutrient management.
The aim of the invention is realized by the following technical scheme:
a digger blade for real-time detection of soil organic matter content, comprising:
the front end of the digging shovel is provided with a digging blade, the rear end of the digging shovel is provided with a sensor integrated cavity, and the digging shovel is connected with an agricultural implement through a fixing bolt and a connecting rod;
the dustproof assembly comprises a dustproof cover and an air pressure balance valve arranged in the dustproof cover;
the main control unit coordinates data acquisition, communication and processing;
a Global Positioning System (GPS) module for recording longitude, latitude and elevation data of the sampling point;
The data acquisition and communication unit is integrated in the sensor integration cavity and comprises a near infrared spectrum sensor, a camera, a light source and a telescopic light screen, wherein the near infrared spectrum sensor is used for acquiring the reflectivity of soil and covering the wave band of 400-2500 nm;
A data processing unit comprising:
The spectrum data processing module adopts a de-envelope curve method to eliminate spectrum baseline drift, then adopts a three-dimensional correlation coefficient method to screen sensitive wave bands which are strongly correlated with organic matters, and the correlation between the soil reflectivity and the organic matter content is strong correlation between 0.6 and 1;
the image data processing module is used for extracting contrast, energy, entropy and color histogram statistic of a gray level co-occurrence matrix (GLCM);
A multi-source data fusion model, which adopts a three-branch convolutional neural network (3D CNN) to fuse spectrum, texture and color characteristics and outputs a predicted value of soil organic matter content;
and the result output module is used for generating a soil organic matter distribution map, displaying the soil organic matter distribution map in real time through a display screen and synchronously uploading the soil organic matter distribution map to the management platform.
As a more preferable technical scheme of the invention, the sensitive wave band which is strongly related to the organic matter comprises 918 nm, 580.8 nm and 831.8 nm.
As a better technical scheme of the invention, the color histogram statistic comprises the mean value, standard deviation, skewness and kurtosis of RGB three channels.
As a preferred technical solution of the present invention, the three-branch convolutional neural network includes:
a spectrum characteristic branch, processing spectrum data screened by a three-dimensional correlation coefficient method by a 1D convolution layer;
b, texture feature branches, namely processing texture feature parameters of the soil image extracted by the gray level co-occurrence matrix by the 1D convolution layer;
c color feature branches, wherein the 1D convolution layer processes the color features of the soil image extracted by the color histogram;
And carrying out multi-source data fusion on the characteristics after the three-linear pooling, and establishing a 3D CNN model.
As a better technical scheme of the invention, the soil organic matter distribution map generated by the result output module meets the following conditions:
Spatial interpolation, namely generating 10m x 10m raster data based on a Kriging interpolation method;
Visual mapping, namely adopting an HSL (hue, saturation and brightness) color space, wherein the gradient from purple to red represents 0 to 5 percent of organic matter.
As a better technical scheme of the invention, the main control unit is a Raspberry Pi, a TensorRT inference engine is carried, the end-to-end response time is less than or equal to 2 seconds, and the continuous spectrum data stream is managed through a circulating buffer zone.
As a better technical scheme of the invention, the adjusting range of the adjustable angle connecting rod is 0-45 degrees, the adjustable angle connecting rod is suitable for interfaces of rotary tillers and sowers, and the carbon steel bearing of the fixing bolt is more than or equal to 50 kg.
As a better technical scheme of the invention, the near infrared spectrum sensor is detachably connected with the camera and the sensor integration cavity, and comprises the following components:
(a) The near infrared spectrum sensor is connected with the sensor integration cavity through a base with a threaded interface;
(b) The camera is connected with the sensor integration cavity through the slide rail.
As a preferred embodiment of the present invention, the power supply module is a replaceable lithium ion battery pack or a solar thin film battery, and the battery management system BMS includes:
a, dynamic power management, wherein power supply of the spectrum sensor and the GPS module is preferentially ensured;
and b, predicting the residual electric quantity based on the operation track of the agricultural machinery, and early warning for 30 minutes in advance.
As a better technical scheme of the invention, the annular dust cover is internally provided with the air pressure balance valve, and the balance air flow is automatically started when the air pressure difference between the annular dust cover and the external environment is less than or equal to 10 Pa.
The beneficial effects are as follows:
The three-dimensional dynamic weight distribution of spectrum, image texture and space position is introduced into the soil detection field for the first time through the cooperative breakthrough of a multi-mode data fusion architecture (spectrum+image+GPS) and an agricultural implement embedded real-time detection system, the unification of laboratory-level precision and field operation efficiency is realized through a detachable structure and a modularized design, the algorithm-hardware cooperative optimization is realized through the combination of a three-branch CNN model and a TensorRT acceleration engine, and a new real-time detection standard is defined.
Drawings
Fig. 1 is a schematic view of the structure of a shovel.
FIG. 2 is a block diagram of the spectrum module inside the digger blade.
Fig. 3 is a three-dimensional correlation coefficient screening example.
Fig. 4 is a graph showing the results of the prediction by the soil organic matter prediction model, which can be used to map the distribution image of the organic matter content of the soil.
The device comprises a display screen 1, a power supply module 2, a near infrared spectrum sensor 3, a light shielding plate 4, a GPS module 5, a connecting rod 6, a main control unit 7, a transmission line 8, a data acquisition and communication module 9, a camera 10, a digger shovel 11, a fixing bolt 12 and a dust cover 13.
Detailed Description
The present invention will be described in further detail with reference to specific examples.
According to the deep scarification digging shovel integrated with the spectrum sensor, the camera, the control unit and the GPS positioning module, spectrum data and image information of soil can be collected while digging the soil in actual operation, and the content of organic matters in the soil can be inverted in real time through a data fusion algorithm. In addition, the digging shovel provided by the invention can record the geographical position of each sampling point while collecting data, generate a soil organic matter content distribution map, and provide accurate soil quality data for agricultural management staff to assist in agricultural decisions such as fertilization, irrigation, soil improvement and the like. The method solves the problems of poor timeliness, complex operation, heavy equipment and the like in the existing soil organic matter monitoring technology, can rapidly and accurately acquire soil organic matter content data in the agricultural production process, and is efficiently applied in the field environment.
Example 1 digger blade structure and hardware configuration.
As shown in figure 1, the excavating shovel for detecting the organic matter content of soil in real time comprises a detachable excavating shovel 11, wherein the front end of the detachable excavating shovel is an excavating blade made of tungsten carbide alloy, the rear end of the detachable excavating shovel is provided with a sensor integrated cavity, the size of the cavity is 120 mm multiplied by 80 multiplied by mm multiplied by 50mm (length multiplied by width multiplied by height), and an IP67 seal design is adopted and a damping bracket with the spring rate coefficient of 5N/mm is arranged in the detachable excavating shovel. The quick-release fixing bolt 12 (M8 bolt, pretightening torque 15-20N M) and the angle-adjustable connecting rod 6 (the adjusting range is 0-45 degrees, and the locking force is more than or equal to 200N) are connected with an agricultural implement. The data acquisition and communication module 9 is a near infrared spectrum sensor 3, a V-shaped optical fiber diffuse reflection structure is adopted, the wave band is covered with 400-2500 nm, a light source is a 10W halogen lamp, and the light source is calibrated through a Labsphere Spectralon standard whiteboard every 30 minutes. The optical fiber is configured to form an included angle of 60 degrees between a core diameter of 400 mu m (light source optical fiber) and 600 mu m (detection optical fiber), and the effective collection depth is 2-5 mm soil layers.
The high-resolution camera 10 is provided with a 200-ten-thousand-pixel CMOS sensor, an annular LED light source and a telescopic light screen 4, supports + -5 cm axial slide rail adjustment and collects soil surface texture images (resolution 1920 multiplied by 1080).
And the GPS module 5 adopts an Ublox NEO-M8N module, the positioning accuracy is less than or equal to 2 cm, and the longitude, latitude and elevation data of the sampling point are recorded in real time.
The main control unit 7 is realized based on raspberry pie, a TensorRT reasoning engine is carried, the end-to-end response time is less than or equal to 2 seconds, the continuous data flow is managed through the circulating buffer zone, and the data acquisition and communication module 9 transmits data to the main control unit 7 through the transmission line 8.
The dustproof component comprises an annular dustproof cover 13 with a hydrophobic nano coating and an air pressure balance valve (a nylon 66 valve body and a silica gel membrane with the thickness of 0.5 mm), wherein the dustproof component is automatically opened when the internal and external air pressure difference is more than or equal to 8Pa, and the dust blocking rate is more than or equal to 99.2%.
Example 2 data processing flow.
Spectral data preprocessing includes:
Reflectance calculation by eliminating ambient light interference by the formula CR (λ) = (I (λ) -min (I))/(max (I) -min (I)), where I (λ) is the value of the spectral signal at wavelength λ, min (I) and max (I) are the minimum and maximum values of the signal, respectively, CR (λ) is the envelope value at wavelength λ;
sensitive band screening, namely 918 nm (C-H bond expansion vibration), 580.8 nm (N-H bond deformation) and 831.8 nm (O-H bond bending) are selected as characteristic bands based on the Pearson correlation coefficient (r > 0.85).
Extracting image features:
Texture features, namely calculating contrast, energy, entropy, homogeneity, correlation and angular second moment by using GLCM;
color characteristics, namely extracting the mean value, standard deviation, skewness and kurtosis of three RGB channels.
Example 3 System Performance verification.
Under the experimental conditions, the sample is carried on a tractor in the northeast black soil area (the organic matter content is 2.1% -4.8%), the operation speed is 5 km/h, and 1200 groups of samples are continuously collected.
Analysis of the results, prediction accuracy: R 2 =0.94, rmse=0.31 g/kg (42% reduction in error compared to the conventional method).
Real-time-end-to-end response time 1.6 seconds (95% confidence interval), supporting 100 data point treatments per second.
The anti-interference performance is that the fluctuation rate of the prediction error is less than or equal to 3 percent (the traditional model is more than or equal to 12 percent) under the working condition of 5-8G vibration.
Example 4 soil organic matter distribution map generation.
Spatial interpolation, which is to generate 10 m x 10 m raster data by adopting a kriging interpolation method (spherical model, variation 35 m).
Visual mapping, namely, expressing organic matter content gradient (purple to red corresponds to 0% -5%) by using an HSL color space, and deriving PDF or CSV format reports.
Example 5 Modular expansion application.
And the power supply module 2 supports a replaceable lithium ion battery pack (with 8 hours of endurance) or a solar film battery, and the BMS predicts the electric quantity according to the track of the agricultural machinery and gives an early warning in advance for 30 minutes.
And (3) algorithm adaptation, namely dynamically adjusting the acquisition frequency of the sensor to 2 times (when the soil humidity is more than 60%) aiming at the red soil area, and optimizing model parameters through migration learning.
Example 6 dustproof and durability test.
After continuous operation for 8 hours, the dust content in the sensor cavity is less than or equal to 0.1 mg/m < 3 >, and the contact angle of the surface of the dust cover is more than or equal to 150 degrees, which indicates that the hydrophobic performance is not attenuated. The polyurethane damping pad (Shore hardness 70A) of the connecting rod effectively absorbs 83% of high-frequency vibration, and the service life of the part is prolonged to 2000 hours.
The embodiment shows that the method realizes the efficient and accurate detection of the organic matter content of the soil through multi-sensor integration, advanced algorithm fusion and modularized design, and provides reliable technical support for accurate agriculture.
According to the invention, by combining the near infrared spectrum sensor and the image acquisition system, the spectrum and image information of the soil can be acquired in real time on site, so that the real-time monitoring of the organic matter content of the soil is realized. The problems of time delay and data feedback lag in the traditional method are avoided, and the agricultural workers are helped to acquire soil information in time and quickly take appropriate agricultural measures.
The present invention is directed to a shovel that is attachable to an agricultural implement by integrating a spectrum sensor, an image acquisition device, and a control unit. The agricultural worker can directly detect soil in the field, and does not need to carry heavy laboratory equipment, so that the agricultural production efficiency is greatly improved.
The invention adopts a multiband spectral sensor and a high-resolution camera, and combines an advanced data processing algorithm (a three-dimensional correlation coefficient screening method and a three-branch convolution neural network fused with multi-source information) to realize high-precision prediction of the organic matter content of soil. By means of the data fusion technology, spectrum data are complementary with image information, the real condition of soil can be reflected more comprehensively, and measurement errors are reduced.
The invention combines with the GPS positioning module, can record the geographic position of each sampling point in real time and generate a soil organic matter distribution map. The charts not only provide visual soil quality information for agricultural managers, but also can be used for monitoring and analyzing the change trend of the soil quality, and support precise fertilization and soil improvement decision.
The design of the digging shovel has high modularization and customization, can be configured according to different soil types, crop requirements and agricultural environments, is suitable for various agricultural operation scenes, and has wide application prospects in the field of precise agriculture.
The digging shovel can be connected to an agricultural implement in operation through mechanical connection, the digging shovel head acquires a sample through digging soil, and the spectrum sensor and the camera acquire spectrum information and surface images of the soil at the same time. The control unit transmits the data to the processing unit, and the spectrum and the image characteristics are analyzed through an algorithm, so that the predicted data of the organic matter content are extracted. The GPS module records sampling positions, and the system combines the soil data and the position information to generate a soil organic matter distribution map. Eventually, the control unit will store this data and derive a report or profile.
The invention provides a portable soil organic matter detection tool which is convenient for agricultural workers to use by combining the spectrum analysis, image processing, data fusion and GPS positioning technology. The tool can collect spectral data and image information of soil in real time while excavating the soil by optimally designing the subsoiler, and accurately predicts the organic matter content of the soil by a built-in control unit and an advanced algorithm.
The invention provides a digging shovel for detecting the organic matter content of soil in real time, which comprises a detachable digging shovel 11, wherein the front end of the detachable digging shovel is provided with a digging blade, the rear end of the detachable digging shovel is provided with a sensor integrated cavity, and the detachable digging shovel is connected with an agricultural implement through a quick-detachable fixing bolt 12 and an angle-adjustable connecting rod 6; the data acquisition and communication module is integrated in the excavating shovel and comprises a near infrared spectrum sensor 3, a high-resolution camera 10, a GPS module 5, a data processing unit and a dust control module, wherein the near infrared spectrum sensor 3 is used for acquiring soil reflectivity by adopting a V-shaped optical fiber diffuse reflection structure and covering wave bands 400-2500 nm, the high-resolution camera 10 is provided with an annular LED light source and a telescopic light screen 4 and is used for acquiring soil surface texture images, the GPS module 5 is used for recording longitude and latitude and elevation data of sampling points, positioning accuracy is less than or equal to 2 cm, the Raspberry P is used as a core controller and is used for coordinating data acquisition, processing and communication, the data processing unit comprises a spectrum data processing module, an annular dust cover 13 and a balance valve, wherein the annular dust cover 13 is used for preprocessing and extracting characteristics of spectrum data by adopting a decoiling method and three-dimensional correlation coefficient screening, the image data processing module is used for extracting contrast, energy, entropy and color histogram statistics of a gray level symbiotic matrix, the multi-source data fusion module is used for integrating spectrum, texture and color characteristics by adopting a three-branch convolution neural network and outputting a predicted value of soil organic matter content, and a result output module is used for generating a soil organic matter distribution map and displaying in real time through a display screen and synchronously uploading the soil organic matter distribution map to a management platform, and a dust control module is used for preventing pollution caused by the dust control module comprises a hydrophobic nano coating.
The design of the digger blade is based on a detachable subsoiler blade, and the digger blade is combined with a spectrum sensor, an image camera and a control unit, and has the following characteristics:
The digging shovel head part is a normal digging shovel, and a plurality of sensors including a near infrared spectrum sensor, a visible light image camera and the like are integrated at the rear part of the digging shovel head part. The design of the digging shovel head ensures that the spectral data and the surface image information of the soil can be collected simultaneously when the soil is dug. The spectrum module comprises a light source, an optical fiber transmission system, a photoelectric sensor and the like, wherein the spectrum module comprises a plurality of V-shaped optical fibers for transmitting light, the light is diffusely reflected by soil and then transmitted to the photoelectric sensor, the sensor converts an optical signal into an electric signal, and the signal processing unit performs amplification, filtering and A/D conversion to ensure high quality and low noise of data. The module can provide spectral information for inverting soil organic matters. The camera is also provided with a high-resolution visible light camera and is responsible for collecting image information of the soil surface. The camera extracts the characteristics of the soil surface through an image processing algorithm, extracts the texture characteristics of the soil surface by using a gray level co-occurrence matrix, extracts the color characteristics of a soil image by using a color histogram and the like, and helps to improve the accuracy of the prediction of the organic matter content of the soil. The control unit is a core part of the digging shovel and is responsible for acquisition, processing and calculation of sensor data. The control unit adopts Raspberry Pi as a main control unit and processes signals acquired from a spectrum sensor, a camera and other equipment. The Raspberry Pi has stronger computing power, can process a large amount of data in real time, and is effectively connected with other modules. And the software part is used for realizing data acquisition and processing, image processing, data fusion and soil organic matter prediction by using a Python programming language. The soil spectrum data acquisition device comprises the following functions of acquiring soil spectrum data and image data from a spectrum sensor and a camera, and transmitting the acquired data to a control unit. The spectrum data processing comprises the steps of preprocessing the spectrum data by using a decomplexing method, selecting a high-correlation wave band by adopting a three-dimensional correlation coefficient screening method, and calculating a partial index formula and a screening result as follows:
CR(λ)=(I(λ)-min(I))/(max(I)- min(I))
Where I (λ) is the value of the spectral signal at wavelength λ, min (I) and max (I) are the minimum and maximum values of the signal, respectively, and CR (λ) is the envelope value at wavelength λ.
。
R1, R2 and R3 are reflectivity data selected from different spectral bands, the selection of which is based on a correlation analysis of Soil Organic Carbon (SOC) content. The purpose of selecting the wave bands is to find the wavelength combination with the strongest correlation with the SOC content, and after the data is transformed by TDI, the correlation coefficient of the data column and the soil organic matter content column is calculated by using the Pearson correlation coefficient method.
R is the pearson correlation coefficient. x i and y i are observations of two variables, respectively.AndThe average of x and y is given respectively. n is the number of samples.
Fig. 3 is a three-dimensional coefficient correlation graph obtained after screening by different TDI formulas, in each graph, three screened sensitive bands can be seen, for example, three sensitive bands in fig. 3 (a) are 918mm, 580.8nm and 831.8nm, after spectral data are obtained each time, an algorithm model in a control unit can calculate various TDIs and select a TDI model with highest correlation as one of input parts of a subsequent multi-source data model.
And (3) image data processing, namely extracting soil surface characteristics by using an image recognition algorithm. And extracting texture characteristic indexes (contrast, energy, entropy, homogeneity, correlation and angular second moment) of the soil surface by using a gray level symbiotic matrix method, and extracting color characteristic indexes (red average value, green average value, blue average value, red standard deviation, green standard deviation, blue standard deviation, red skewness, green skewness, blue skewness, red kurtosis, green kurtosis and blue kurtosis) of the soil image by using a color histogram method.
And (3) carrying out multi-source information (spectrum characteristics, texture characteristics and color characteristics) fusion on the spectrum data and the image characteristics, substituting the fused data into a Support Vector Machine (SVM), a random forest model (RF), a partial least squares model (PLS) and a 3D CNN model, wherein the model results are shown in figure 4.
RMSE represents the absolute prediction error, R2 represents the correlation coefficient, and in regression analysis, the closer R2 is to 1, the lower the RMSE value is, which generally indicates that the model fitting capability is strong, the prediction error is small, and the model has better effectiveness and reliability.
And (3) soil prediction, namely drawing a distribution image of the organic matter content of the soil according to a result predicted by the soil organic matter prediction model and the recorded geographic coordinates.
In order to ensure accurate association of soil data and geographic positions, the invention designs an integrated GPS module which is responsible for recording the geographic position of each sampling point in real time. Assuming that the precision of the GPS module is +/-5 meters, the geographic coordinates of each sampling point can be accurately associated with the acquired soil data by recording the position data in real time, so that the accuracy of a soil organic matter content distribution map is improved. For example, if 100 soil samples are collected in a 10 hectare farmland, agricultural management personnel can track the soil quality distribution situation in real time through accurate GPS positioning, so that accurate fertilization and soil management can be performed. The technology can provide high-efficiency spatial data support by combining high-precision GPS positioning (such as Ublox NEO-M8N module, with the precision of +/-2.5 meters).
All collected soil data (including spectrum data, image data, GPS positioning data, etc.) are stored in the control unit and can be displayed in real time by the display screen 1. Assuming the system collects 100 data points per second, after 1 hour of soil detection, a total of 360,000 data points are collected, which will be stored and processed in real time. The agricultural workers can check the soil organic matter content prediction result in real time through the control unit, and can make more accurate adjustment according to the data. The system also supports exporting data into CSV or PDF format, which is convenient for generating soil organic matter distribution map and forecast report, and further helps agricultural management decision.
The system provided by the invention has good user interactivity, and a user can intuitively check the data acquired in real time and the predicted value of the organic matter content of the soil through the display screen. The system design adopts a concise and visual user interface, and a user can select different soil samples through the touch screen to check or modify acquisition parameters. Assuming the system is capable of completing data display and updating within 1 second, when the predicted soil organic matter content value deviates from a preset threshold value (such as more than +/-10%), the system can remind an agricultural worker to adjust the fertilization scheme in real time. Through the visual user interaction mode, the agricultural workers can react more quickly, and the efficiency and the accuracy of agricultural production are improved.
The digger blade design of the present invention is highly modular and customizable, and a user can select the appropriate sensor and algorithm configuration based on soil type, crop requirements, and the needs of a particular agricultural environment. For example, the system can dynamically adjust the sensitivity and data acquisition frequency of the sensor according to the information of the pH value, humidity, temperature and the like of the soil. When the soil humidity exceeds 60%, the acquisition frequency of the spectrum sensor can be doubled to ensure the accuracy of the data. The modularized design enables the device to adapt to different agricultural operation scenes, and particularly has wide applicability in the fields of precise agriculture (such as crop yield prediction), intelligent agriculture (such as automatic irrigation), environment monitoring (such as polluted soil detection) and the like. By adopting a standardized interface, the device can expand a new sensor module or algorithm according to different requirements, so that the flexibility and adaptability of the system are further improved.
In some embodiments, the texture features extracted by the image data processing module comprise contrast, energy, entropy, homogeneity, correlation and angular second moment of the gray level co-occurrence matrix, and the color features comprise mean, standard deviation, skewness and kurtosis of the RGB three channels.
In some embodiments, the 3D CNN comprises a spectrum characteristic branch, a texture characteristic branch and a color characteristic branch, wherein the spectrum characteristic branch is screened by a 1D convolution layer processing target three-dimensional correlation coefficient method, the texture characteristic branch is used for processing texture characteristic parameters of a soil image extracted by a gray level symbiotic matrix by the 1D convolution layer, and the color characteristic branch is used for processing color characteristics of the soil image extracted by a color histogram by the 1D convolution layer. The characteristics are subjected to three-linear pooling and then subjected to multi-source data fusion, a 3D CNN model is established, and model evaluation meets the conditions that a determination coefficient R 2 is more than or equal to 0.92 and a Root Mean Square Error (RMSE) is less than or equal to 0.35g/kg.
Sensitive bands verify pearson correlation coefficient matrices at 918nm (C-H bond stretching vibration), 580.8nm (N-H bond deformation), 831.8nm (O-H bond bending), as shown in Table 1.
TABLE 1
The test was compared with the NY/T1121.6-2006 method (n=500).
4.3-Fold acceleration was achieved by means of TensorRT INT8 quantization for a 200-group representative spectral image pair.
The quantization error is less than or equal to 0.02g/kg (based on the RMSE less than or equal to 0.35 g/kg).
Compared with the traditional dual-branch model (such as 1D+2D CNN), the structure realizes dynamic weight distribution through a multi-head attention mechanism, and is compared with the image Net test set as shown in table 2.
TABLE 2
The channel attention mechanism (SE module) is adopted to enhance the robustness of vibration noise, the fluctuation rate of a prediction error is less than or equal to 3 percent (the traditional model is more than or equal to 12 percent) under the working condition of 5-8G acceleration, and the dynamic weight distribution of the characteristic channel is realized through the following three-stage operation.
The Squeeze phase (global information compression):
inputting a feature map dimension U: (h=224, w=224, c=512 corresponds to ResNet-18 final residual block outputs).
Global average pooling operation: generating channel statistics vectors Where z c represents the global spatial feature of the c-th channel.
The expression phase (modeling of channel correlation):
double full connection layer structure:
dimension reduction full connection layer (r=16 is compression ratio, and is the optimal balance point through experimental verification).
ReLU activation function (prevent gradient from disappearing):
The full connection layer of the upkeep (resume the primitive channel number);
Sigmoid function (output weight range [0,1 ]);
Stage Reweight (characteristic channel calibration):
Channel level multiplication: Wherein s c represents importance weight of the c-th channel, and noise suppression and key feature enhancement are realized.
The compression ratio r is determined as shown in table 3.
TABLE 3 Table 3
When r=16 is selected, the model achieves pareto optimum on the parameter quantity-precision-speed three-dimensional index.
Vibration noise suppression by channel weights.
The weight of the high-frequency vibration noise related channel is dynamically reduced (experiments show that the weight attenuation of the noise channel under the vibration working condition is more than or equal to 70%).
Feature robustness is enhanced by giving 1.2-1.5 times weight gain to the soil surface texture (GLCM features) and color distribution (RGB histogram) related channels.
The full link layer weights of the SE modules are converted from FP32 to INT8, achieving a 4.3-fold acceleration (delay down from 85ms to 20 ms) on raspberry group 4B.
The SE weight vector is managed by using a ring buffer, so that the memory occupation is reduced by 62% (from 8.2MB to 3.1 MB).
The channel attention validity verification is shown in table 4.
TABLE 4 Table 4
The computational overhead analysis is shown in table 5.
TABLE 5
The circulation buffer area management supports continuous transmission of the maximum 30 seconds breakpoint, and is suitable for intermittent loss of GPS signals of agricultural machinery (average packet loss rate is less than or equal to 2%).
In some embodiments, the soil organic matter profile generated by the result output module satisfies:
Spatial interpolation, namely generating 10m x 10m raster data based on a Kriging interpolation method;
visual mapping adopts an HSL color space, and the gradient from purple to red represents 0 to 5 percent of organic matter.
In some embodiments, the main control unit of the data processing unit is a raspberry pie, a TensorRT inference engine is carried, the end-to-end response time is less than or equal to 2 seconds, and the continuous spectrum data stream is managed through a circular buffer.
In some embodiments, the adjusting range of the adjustable angle connecting rod 6 is 0-45 degrees, the adjustable angle connecting rod is suitable for rotary tillers and seeder interfaces, and the carbon steel bearing of the fixing bolt (12) is more than or equal to 50 kg.
In some embodiments, the near infrared spectrum sensor 3 and the camera 10 are in a modular quick-release design, including:
(a) The spectrum sensor base is provided with an M12 threaded interface and a waterproof sealing ring;
(b) And the camera slide rail bracket supports the axial position adjustment + -5 cm.
In some embodiments, the power supply module 2 is a replaceable lithium ion battery pack or a solar thin film battery, and a Battery Management System (BMS) thereof includes:
(a) Dynamic power management, preferably guaranteeing power supply of the spectrum sensor and the GPS module;
(b) And predicting the residual electric quantity based on the operation track of the agricultural machinery, and early warning for 30 minutes in advance.
In some embodiments, the annular dust cover 13 of the dust-proof assembly is internally provided with an air pressure balance valve, and the balance air flow is automatically opened when the air pressure difference between the annular dust cover and the external environment is less than or equal to 10 Pa.
In some embodiments, the quick release securing bolt 12 is an M8 bolt with a pre-tightening torque of 15-20N M, which is adapted to an ISO 2320 standard agricultural machine interface.
In some embodiments, a polyurethane damping pad (shore hardness 70A) is additionally arranged between the connecting rod and the fixing bolt, so that the polyurethane damping pad can absorb over 80% of high-frequency vibration (frequency >50 Hz) in the operation of the agricultural machinery.
In some embodiments, the end of the adjustable angle connecting rod 6 is provided with an ISO 5675 standard quick coupler, a universal interface of a rotary cultivator (such as John Deere 5E series) and a seeder (such as Case IH 2150) is supported, the angle precision is adjusted by +/-1 DEG, and the locking force is adjusted to be more than or equal to 200N.
In some embodiments, the sensor integration cavity has a cavity size of 120 mm ×80× 80 mm ×50× 50 mm (length×width×height), and an IP67 seal design is used, and a shock mount (spring rate 5N/mm) is preset inside.
In some embodiments, the source fiber (core diameter 400 μm, NA 0.22) is at a 60 ° angle with the detection fiber (core diameter 600 μm, NA 0.24), spacing 0.5mm, and effective collection depth is 2-5 mm soil layers.
In some embodiments, the halogen light source (wavelength range 400-2500 nm, power 10W, color temperature 2856K) is automatically calibrated once every 30 minutes (using Labsphere Spectralon standard white board).
In some embodiments, the base is fixed by an M12 threaded interface (according to IEC 61076-2-101 standard), and the waterproof sealing ring is made of fluororubber (resistant to temperature of-20 ℃ to 120 ℃).
In some embodiments, the spectral signal is normalized to the reflectivity, R (λ) = ((λ) -D (λ))/(W (λ) -D (λ)).
In some embodiments, the envelope junction selection rule is a piecewise linear fit with the absorption peak minima (as at 1410 nm, 1910, nm) as nodes.
In some embodiments, the band (400-2500 nm), space (1 m2 area mean around the sample point), time (sliding window of 10 consecutive samples).
In some embodiments, based on pearson correlation coefficient (r > 0.85) screening, experiments demonstrated that the combined prediction error is minimal (rmse=0.28 g/kg) for 918 nm (C-H bond stretching vibration), 580.8 nm (NO 3 -absorption peak), 831.8 nm (organic aromatic c=c feature).
In some embodiments, the spectral branching is a 3-layer 1D convolution (kernel size [3,5,3], step size 1, channel number [32,64,128 ]), followed by a global average pooling layer.
In some embodiments, the branches are fused, a multi-headed attention mechanism (4-headed, embedded dimension 256), with fusion weights dynamically assigned by a learnable parameter.
In some embodiments, the model training process is as follows:
the data set comprises 5000 groups of soil samples (the organic matter content is 0.5% -5.2%), 70% training set, 15% verification set and 15% test set;
super-parameters, initial learning rate 0.001 (Cosine decay), batch size 32, loss function Huber Loss (delta=0.5), optimizer AdamW;
And adopting FP16 quantification in TensorRT deployment, wherein the model reasoning time is actually measured to be 1.3-1.8 seconds (raspberry group 4B, and the ambient temperature is 25 ℃).
In some embodiments, the interpolation parameters are a semi-variational function, a spherical model (0.12 for the block, 1.05 for the base, 35 for the range, m), an anisotropy ratio of 1.5 (30℃in azimuth). The searching strategy is that the maximum radius is 50m, the minimum number is 8, the interpolation error is controlled by cross validation (average absolute error MAE is less than or equal to 0.2 g/kg).
In some embodiments, the valve body of the air pressure balance valve is made of nylon 66 (temperature resistant to-40 ℃ to 120 ℃), a built-in silica gel membrane (thickness 0.5 mm, elastic modulus 1.5 MPa) and spring rate 0.8N/mm. The differential pressure sensor (model Honeywell HSC series) monitors the internal and external air pressure in real time, and when the delta P is more than or equal to 8 Pa, the valve is triggered to open (response time is less than or equal to 0.5 seconds).
The northeast black soil area (the organic matter content is 2.1% -4.8%) of the embodiment of the invention is mounted on a Leiwo M904 tractor, and the operation speed is 5 km/h. Prediction accuracy r2=0.94, rmse=0.31 g/kg (n=1200). Response time 1.6 seconds (95% confidence interval), dust-proof effect: 8 hours of continuous operation without dust (dust blocking rate 99.2%).
In summary, the invention aims to provide a soil organic matter detection tool with high real-time performance, convenience and accuracy, which helps agricultural producers make scientific and reasonable decisions by acquiring soil data in real time and promotes agricultural production to develop towards the intelligent and accurate directions.
The shovel head part of the shovel integrates a plurality of sensors (such as a near infrared spectrum sensor, an image camera and the like), and has normal digging function. Through optimizing the structural design of the excavating shovel head, the soil spectrum data and the surface image information can be synchronously collected when the soil excavating task is executed. The design makes the agriculture worker not need to carry heavy equipment, and can directly acquire soil information in the field, thereby improving convenience and high efficiency of soil detection. The three-dimensional sensitive wave band screening and the 3D CNN mainly comprise the steps of firstly, selecting a wave band highly related to the soil organic matter content from a plurality of spectrum wave bands by a three-dimensional correlation coefficient screening method, ensuring that spectrum data with the most predictive value is extracted, and improving the prediction precision of a model. And then, processing the fused spectral data and image characteristics by adopting a three-branch convolutional neural network, respectively utilizing three branches of the spectral characteristics, the texture characteristics and the color characteristics to independently process, and fusing the results, thereby realizing more comprehensive prediction of the organic matter content of the soil. The method effectively combines the advantages of multi-source data, reduces errors and improves the accuracy and reliability of prediction. The system can record the geographic position of each data sampling point, combines the collected soil data with the geographic position to generate a soil organic matter distribution map, and helps an agricultural manager to carry out precise fertilization, irrigation and soil improvement decisions.
The spectral sensor and the image sensor of the digger blade of the present invention can be configured according to different soil types and environmental conditions. The spectrum sensor can automatically adjust according to the humidity of soil and other environmental factors, so that the accuracy of data is ensured. Meanwhile, the algorithm is optimized according to different characteristics of the soil, and accurate organic matter prediction is provided.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510660814.6A CN120177408B (en) | 2025-05-22 | 2025-05-22 | A digging shovel for real-time detection of soil organic matter content |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510660814.6A CN120177408B (en) | 2025-05-22 | 2025-05-22 | A digging shovel for real-time detection of soil organic matter content |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN120177408A true CN120177408A (en) | 2025-06-20 |
| CN120177408B CN120177408B (en) | 2025-09-12 |
Family
ID=96043704
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510660814.6A Active CN120177408B (en) | 2025-05-22 | 2025-05-22 | A digging shovel for real-time detection of soil organic matter content |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN120177408B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA2836642A1 (en) * | 2013-12-17 | 2015-06-17 | Asaad A.A. Ahmed | All-purpose gardening tool |
| US20160307312A1 (en) * | 2015-04-15 | 2016-10-20 | Ingrain, Inc. | Method For Determining Fabric And Upscaled Properties Of Geological Sample |
| CN212031475U (en) * | 2020-04-22 | 2020-11-27 | 南京信息工程大学 | Intelligent shovel |
| CN115171080A (en) * | 2022-06-14 | 2022-10-11 | 安徽工程大学 | Compressed video driver behavior identification method for vehicle-mounted equipment |
| US20230160182A1 (en) * | 2021-11-22 | 2023-05-25 | Minesense Technologies Ltd. | Compositional multispectral and hyperspectral imaging systems for mining shovels and associated methods |
| CN116310881A (en) * | 2023-04-24 | 2023-06-23 | 中国科学院空天信息创新研究院 | Soil organic matter content estimation method, device, electronic equipment and storage medium |
| CN116932996A (en) * | 2023-07-10 | 2023-10-24 | 中国科学院长春光学精密机械与物理研究所 | Estimation method of soil organic matter content based on improved Hapke model |
| CN118858169A (en) * | 2023-04-28 | 2024-10-29 | 中国农业大学 | Soil organic matter content detection method, device, equipment and storage medium |
-
2025
- 2025-05-22 CN CN202510660814.6A patent/CN120177408B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA2836642A1 (en) * | 2013-12-17 | 2015-06-17 | Asaad A.A. Ahmed | All-purpose gardening tool |
| US20160307312A1 (en) * | 2015-04-15 | 2016-10-20 | Ingrain, Inc. | Method For Determining Fabric And Upscaled Properties Of Geological Sample |
| CN212031475U (en) * | 2020-04-22 | 2020-11-27 | 南京信息工程大学 | Intelligent shovel |
| US20230160182A1 (en) * | 2021-11-22 | 2023-05-25 | Minesense Technologies Ltd. | Compositional multispectral and hyperspectral imaging systems for mining shovels and associated methods |
| CN115171080A (en) * | 2022-06-14 | 2022-10-11 | 安徽工程大学 | Compressed video driver behavior identification method for vehicle-mounted equipment |
| CN116310881A (en) * | 2023-04-24 | 2023-06-23 | 中国科学院空天信息创新研究院 | Soil organic matter content estimation method, device, electronic equipment and storage medium |
| CN118858169A (en) * | 2023-04-28 | 2024-10-29 | 中国农业大学 | Soil organic matter content detection method, device, equipment and storage medium |
| CN116932996A (en) * | 2023-07-10 | 2023-10-24 | 中国科学院长春光学精密机械与物理研究所 | Estimation method of soil organic matter content based on improved Hapke model |
Non-Patent Citations (2)
| Title |
|---|
| 丁宇雪等: "《东北地区自然资源监测与黑土退化》", 30 September 2021, 中国地质大学出版社, pages: 190 - 193 * |
| 童超等: "《人工智能在肺癌诊断中的应用》", 28 February 2023, 北京航空航天大学出版社, pages: 82 - 83 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN120177408B (en) | 2025-09-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Roy et al. | Precision farming: A step towards sustainable, climate-smart agriculture | |
| Hall et al. | Optical remote sensing applications in viticulture‐a review | |
| US6937939B1 (en) | Soil measuring instrument, soil measurement assisting device and method, recorded medium on which a program is recorded, recorded medium on which data is recorded, application amount controller, application amount determining device, method for them, and farm working determination assisting system | |
| CN115665690B (en) | River buffer zone soil restoration feedback system and restoration method | |
| Sui et al. | Ground-based sensing system for cotton nitrogen status determination | |
| CN101975764B (en) | Device and method for multi-band soil nitrogen detection based on near-infrared spectroscopy | |
| CN101930662A (en) | Real-time monitoring system and method of farmland information based on remote monitoring | |
| CN109557003B (en) | Pesticide deposition amount detection method and device and data acquisition combination device | |
| JPWO2001004627A1 (en) | Soil measuring device, soil measuring support device and method, recording medium with program recorded thereon, recording medium with data recorded thereon, spraying rate control device, spraying rate determination device and methods therefor, and agricultural work decision support system | |
| CN102256102A (en) | Agricultural condition information real-time monitoring method and system based on remote monitoring | |
| CN118968351A (en) | A system and method for preventing and controlling citrus pests and diseases | |
| CN101666741B (en) | High density nondestructive collection method of airborne crop nitrogen information | |
| CN105136732A (en) | Field crop dual band imaging NDVI measurement apparatus | |
| CN100480681C (en) | Plant growth information acquisition device based on near infrared spectrum | |
| CN120298146A (en) | A smart tea garden planting management system and method | |
| CN216132492U (en) | Vegetation root system state monitoring device and system | |
| CN120177408B (en) | A digging shovel for real-time detection of soil organic matter content | |
| Cherry et al. | Calibrating the LI-COR LAI-2000 for estimating leaf area index in eucalypt plantations | |
| Mazzetto et al. | Comparing commercial optical sensors for crop monitoring tasks in precision viticulture | |
| CN105181632B (en) | NDVI measuring device is imaged in network-type various dimensions plant | |
| CN209525221U (en) | Pesticide deposit amount characteristic wave data acquisition and pesticide deposit amount detection device | |
| CN119438099B (en) | A method and system for inversion of grassland nitrogen content | |
| Sun et al. | China | |
| Picchiotti et al. | Enhancing precision agriculture through cyber-physical systems: a functional monitoring platform as decision support tool | |
| CN117665249A (en) | Wireless monitoring system for high-quality grain field soil heavy metal pollution |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |