[go: up one dir, main page]

CN119338695B - Urban building image processing method and system based on unmanned aerial vehicle inspection - Google Patents

Urban building image processing method and system based on unmanned aerial vehicle inspection Download PDF

Info

Publication number
CN119338695B
CN119338695B CN202411884141.4A CN202411884141A CN119338695B CN 119338695 B CN119338695 B CN 119338695B CN 202411884141 A CN202411884141 A CN 202411884141A CN 119338695 B CN119338695 B CN 119338695B
Authority
CN
China
Prior art keywords
detail
urban building
image
building image
pixel point
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.)
Active
Application number
CN202411884141.4A
Other languages
Chinese (zh)
Other versions
CN119338695A (en
Inventor
杨弦
艾霹雳
黄瑜
利彩青
刘扬
陈睿
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong Zhiyi Data Co ltd
Guangzhou Institute of Technology of Xidian University
Original Assignee
Guangdong Zhiyi Data Co ltd
Guangzhou Institute of Technology of Xidian University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guangdong Zhiyi Data Co ltd, Guangzhou Institute of Technology of Xidian University filed Critical Guangdong Zhiyi Data Co ltd
Priority to CN202411884141.4A priority Critical patent/CN119338695B/en
Publication of CN119338695A publication Critical patent/CN119338695A/en
Application granted granted Critical
Publication of CN119338695B publication Critical patent/CN119338695B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00Image enhancement or restoration
    • G06T5/50Image enhancement or restoration using two or more images, e.g. averaging or subtraction
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00Image enhancement or restoration
    • G06T5/20Image enhancement or restoration using local operators
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00Indexing scheme for image analysis or image enhancement
    • G06T2207/20Special algorithmic details
    • G06T2207/20212Image combination
    • G06T2207/20224Image subtraction

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Image Processing (AREA)

Abstract

本发明属于图像处理技术领域,具体涉及一种基于无人机巡检的城市建筑图像处理方法及系统,其方法包括:对城市建筑图像分解获得多个方向的不同尺寸的细节子图,根据不同尺寸对细节特征的表现度以及城市建筑图像中的目标像素点在各方向的不同尺寸的细节子图中对应的像素点的像素值,获得目标像素点在各方向的细节表现程度,综合多个方向的细节表现程度获得综合细节向量,根据综合细节向量获得目标像素点的高斯滤波器;根据高斯滤波器对各像素点进行滤波处理,将城市建筑图像与获得的背景图像相减,得到受光均匀的城市建筑图像。本发明获得的受光均匀的城市建筑图像同时具有较好的清晰度和较好的匀光效果。

The present invention belongs to the field of image processing technology, and specifically relates to a method and system for processing urban building images based on drone inspection, the method comprising: decomposing urban building images to obtain detail sub-images of different sizes in multiple directions, obtaining the detail expression degree of the target pixel in each direction according to the expression degree of detail features of different sizes and the pixel values of the corresponding pixels of the detail sub-images of different sizes in each direction of the target pixel in the urban building image, obtaining a comprehensive detail vector by integrating the detail expression degrees of multiple directions, and obtaining a Gaussian filter of the target pixel according to the comprehensive detail vector; filtering each pixel according to the Gaussian filter, subtracting the urban building image from the obtained background image, and obtaining an urban building image with uniform light. The urban building image with uniform light obtained by the present invention has both good clarity and good uniform light effect.

Description

Urban building image processing method and system based on unmanned aerial vehicle inspection
Technical Field
The invention relates to the technical field of image processing. More particularly, the invention relates to an urban building image processing method and system based on unmanned aerial vehicle inspection.
Background
In the process of inspection and acquisition of urban building images based on unmanned aerial vehicles, uneven brightness distribution occurs in the obtained urban building images due to the influence of factors such as illumination conditions, shooting angles and atmospheric conditions, so that the difficulty in follow-up accurate analysis of urban buildings is increased.
In the related art, for example, a Chinese patent application document with the application publication number of CN115588087A discloses an unmanned aerial vehicle mapping data processing system and method for modeling, wherein the unmanned aerial vehicle mapping data processing system comprises an image acquisition module, an image processing module, an elevation precision detection module, an image dodging processing module, an air-three adjustment processing module, a joint adjustment module, a texture mapping module and a photo stabbing module, the image acquisition module acquires images by utilizing unmanned aerial vehicle oblique photography and transmits acquired image data to the image processing module, the oblique photography technology in the image acquisition module is utilized to increase the simultaneous exposure mode of four-angle oblique view lenses, simultaneously acquire images of lower view and oblique view angles, can further access surface information with characteristics, and the images contain abundant real environment information by relative positioning and absolute positioning in the oblique photography measurement process.
In the image dodging processing module in the related art, the image dodging processing module is formed byThe light homogenizing algorithm performs light homogenizing treatment on the image with uneven illumination to obtain an image with uniform light receiving; The light homogenizing algorithm carries out low-pass filtering on the original image through a Gaussian filter so as to obtain a background image simulating brightness distribution, and the original image and the background image are subjected to difference so as to obtain an image with uniform light receiving, wherein the selection of the Gaussian filter determines the light homogenizing effect and definition of the image with uniform light receiving, so that the accurate analysis of urban buildings is affected.
Disclosure of Invention
To solve the above problemsThe selection of the Gaussian filter in the dodging algorithm determines the dodging effect and definition of the uniformly-received image, so that the technical problem of accurately analyzing the urban building is affected.
In a first aspect, the invention provides a city building image processing method based on unmanned aerial vehicle inspection, which comprises the steps of inspecting and collecting city building images through unmanned aerial vehicle inspection, decomposing the city building images through wavelet transformation to obtain detail subgraphs with different sizes in multiple directions, calculating the representation degree of detail features with different sizes according to the ratio of the different sizes to the sizes of the city building images, taking any pixel point in the city building images as a target pixel point, weighting and summing pixel values of corresponding pixel points in the detail subgraphs with different sizes of the target pixel point in all directions according to the representation degree of the detail features with different sizes to obtain the detail representation degree of the target pixel point in all directions, taking the detail representation degree of the target pixel point in all directions as the model length of a detail vector of the target pixel point in all directions, calculating the sum of the detail vectors of the target pixel point in all directions as a comprehensive detail vector of the target pixel point, and obtaining a two-dimensional Gaussian model according to the comprehensive detail vector,The directions of the two dimensions of (a) are respectivelyAndStandard deviation of two dimensions is equal toAndWherein, the method comprises the steps of, wherein,In order to integrate the direction of the detail vector,The length of the projection of the comprehensive detail vector in the first dimension and the second dimension respectively according to a two-dimensional Gaussian modelThe method comprises the steps of obtaining a Gaussian filter of a target pixel point, carrying out filtering treatment on each pixel point according to the Gaussian filter of each pixel point in an urban building image to obtain a background image, subtracting the urban building image from the background image, and obtaining the urban building image with uniform light receiving.
According to the method, the detail expression degree of the target pixel point in each direction is obtained according to the expression degree of the detail features of different sizes and the pixel values of the corresponding pixel points in the detail subgraphs of different sizes of the target pixel point in the urban building image, the detail expression degree of the target pixel point in each direction is obtained, the comprehensive detail vector is obtained according to the detail expression degrees of a plurality of directions, the Gaussian filter of the target pixel point is obtained according to the comprehensive detail vector, for the pixel point with the larger detail expression degree, the standard deviation of the two-dimensional Gaussian model is larger, the detail features contained in the local area of the pixel point in the finally obtained background image are fewer, the definition of the obtained urban building image with uniform light is enhanced, for the pixel point with the smaller detail expression degree, the standard deviation of the two-dimensional Gaussian model is smaller, the effect of the local area of the pixel point in the finally obtained background image is good, and therefore the uniform light homogenizing effect of the obtained urban building image with uniform light is enhanced.
Preferably, the decomposing the city building image by wavelet transformation to obtain detail subgraphs with different sizes in multiple directions comprises the steps of when the decomposing the city building image by wavelet transformation is carried outSub-decomposition, the number of kinds of all sizes is equal toWherein, the method comprises the steps of,First, theThe size of the detail subgraph of each direction obtained by sub-decomposition isWill beIs marked as the firstSize, thenIs the firstThe length of the dimension is such that,Is the firstWidth of dimension, and,,For the length of the image of the urban building,The width of the city building image, wherein the directions comprise a vertical direction, a horizontal direction and a diagonal direction, and the vertical direction, the horizontal direction and the diagonal direction are respectively marked as the firstDirection, the firstDirection and the firstDirection.
The invention obtains detail subgraphs of the urban building image in multiple scales and directions by utilizing the multi-resolution characteristic of wavelet transformation, thereby analyzing the intensity of detail characteristics of local areas of each pixel point.
Preferably, the degree of expression of the detail features by the different sizes satisfies the expression: In the formula (I), in the formula (II), Is the firstThe degree of performance of the size with respect to the detail features,Is the number of times of decomposition, and the number of kinds of all sizes is equal to,As a function of the natural index of refraction,Is the firstThe dimensions of the product are such that,Is the firstThe length of the dimension is such that,Is the firstWidth of dimension, and,,For the size of the image of the urban building,For the length of the image of the urban building,Is the width of the urban building image.
In the invention, the pixel values of the pixel points in the obtained detail subgraph can show the detail features in the city building image, and the pixel values of the pixel points in the detail subgraphs with different sizes are different in the expression degree of the detail features in the city building image, so the invention calculates the expression degree of the detail features with different sizes according to the ratio of the different sizes to the sizes of the city building image.
Preferably, the detail expression degree of the target pixel point in each direction satisfies the expression: In the formula (I), in the formula (II), As coordinates of the target pixel point in the urban building image,The abscissa and the ordinate of the target pixel point are respectively,Is at the first pixel pointThe degree of detail of the individual directions is expressed,Is the number of times of decomposition, and the number of kinds of all sizes is equal to,Is the firstThe degree of performance of the size with respect to the detail features,Is the firstFirst of all directionsThe coordinates in the detail drawing of the size areIs used for the pixel values of the pixel points of (a),As a round-up function.
According to the invention, the pixel values of the corresponding pixel points in the detail subgraphs with different sizes of the target pixel point are weighted by the expressive degree of the detail features with different sizes, so that the expressive degree of the detail of the target pixel point in different directions is obtained.
Preferably, the firstFirst of all directionsThe coordinates in the detail drawing of the size areIs that the target pixel is at the first pixelFirst of all directionsAnd the corresponding pixel points in the detail subgraph of the size.
Preferably, the two-dimensional gaussian modelHas two parameters, namely a mean value and a standard deviation, and a two-dimensional Gaussian modelIs of two dimensions the mean values were all 0.
Preferably, the Gaussian filter is essentially a matrix, wherein the number of rows of the Gaussian filter in the first dimension isThe number of rows in the second dimension is,Respectively two-dimensional Gaussian modelsIs defined as the standard deviation of the two dimensions of (1),In order to round up the function,Is the expansion coefficient.
In the invention, for the pixel points with larger detail expression level, a Gaussian filter with larger size is arranged, so that the blurring degree of the filtered pixel points is larger, the detail features contained in the local area of the pixel points in the finally obtained background image are fewer, thereby enhancing the definition of the obtained urban building image with uniform light receiving, and for the pixel points with smaller detail expression level, the Gaussian filter with smaller size is arranged, so that the effect of simulating brightness distribution in the local area of the pixel points in the finally obtained background image is better, thereby enhancing the uniform light effect of the obtained urban building image with uniform light receiving.
Preferably, the method for obtaining the urban building image with uniform light receiving comprises the steps of correspondingly stretching and displaying the image after subtraction treatment to finally obtain the urban building image with uniform light receiving, wherein the stretching and displaying method comprises 2% linear stretching, contrast parameter stretching and gradient stretching display.
Preferably, the method comprises the steps of taking the pixel point at the upper left corner in the urban building image as an origin, taking the origin level to the right as the positive direction of the transverse axis, and taking the origin vertically downward as the positive direction of the longitudinal axis, constructing a rectangular coordinate system, wherein the value range of the abscissa of the pixel point in the urban building image isThe range of the ordinate isAnd (2) andFor the length of the image of the urban building,Is the width of the urban building image.
In a second aspect, the invention provides an urban building image processing system based on unmanned aerial vehicle inspection, which comprises a processor and a memory, wherein the memory stores computer program instructions, and the urban building image processing method based on unmanned aerial vehicle inspection is realized when the computer program instructions are executed by the processor.
By adopting the technical scheme, the urban building image processing method based on unmanned aerial vehicle inspection generates a computer program, and the computer program is stored in the memory to be loaded and executed by the processor, so that terminal equipment is manufactured according to the memory and the processor, and the urban building image processing method based on unmanned aerial vehicle inspection is convenient to use.
The invention has the beneficial effects that:
According to the method, the detail expression degree of the target pixel point in each direction is obtained according to the expression degree of the detail features of different sizes and the pixel values of the corresponding pixel points in the detail subgraphs of different sizes of the target pixel point in the urban building image, the detail expression degree of the target pixel point in each direction is obtained, the comprehensive detail vector is obtained according to the detail expression degrees of a plurality of directions, the Gaussian filter of the target pixel point is obtained according to the comprehensive detail vector, for the pixel point with the larger detail expression degree, the standard deviation of the two-dimensional Gaussian model is larger, the detail features contained in the local area of the pixel point in the finally obtained background image are fewer, the definition of the obtained urban building image with uniform light is enhanced, for the pixel point with the smaller detail expression degree, the standard deviation of the two-dimensional Gaussian model is smaller, the effect of the local area of the pixel point in the finally obtained background image is good, and therefore the uniform light homogenizing effect of the obtained urban building image with uniform light is enhanced.
Drawings
The above, as well as additional purposes, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description when read in conjunction with the accompanying drawings. In the drawings, embodiments of the invention are illustrated by way of example and not by way of limitation, and like reference numerals refer to similar or corresponding parts and in which:
FIG. 1 is a flow chart schematically illustrating a method for processing urban building images based on unmanned aerial vehicle inspection in the present invention;
FIG. 2 is a schematic diagram schematically illustrating an acquired city building image;
fig. 3 is a flowchart schematically showing the gaussian filter for acquiring each pixel point in the urban building image in step S2;
FIG. 4 is a schematic diagram schematically illustrating a two-dimensional Gaussian model with standard deviation of 0.5 for both the first dimension and the second dimension;
FIG. 5 is a schematic diagram schematically illustrating a Gaussian filter obtained from the two-dimensional Gaussian model shown in FIG. 4;
FIG. 6 is a schematic diagram schematically illustrating a two-dimensional Gaussian model with standard deviation of 0.25 for both the first dimension and the second dimension;
FIG. 7 is a schematic diagram schematically illustrating a Gaussian filter obtained from the two-dimensional Gaussian model shown in FIG. 6;
fig. 8 is a schematic view schematically showing a background image obtained after filtering the urban building image shown in fig. 2;
fig. 9 is a schematic view schematically showing a city building image in which light reception is uniform obtained by subtracting the city building image shown in fig. 2 from the background image shown in fig. 9.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.
The embodiment of the invention discloses an urban building image processing method based on unmanned aerial vehicle inspection, which comprises the following steps of S1-S3 with reference to FIG. 1:
In the process of inspection and acquisition of urban building images based on unmanned aerial vehicles, uneven brightness distribution occurs in the obtained urban building images due to the influence of factors such as illumination conditions, shooting angles and atmospheric conditions, so that the difficulty in follow-up accurate analysis of urban buildings is increased.
In order to eliminate uneven brightness distribution in the urban building image which is inspected and collected by the unmanned aerial vehicle, the urban building image needs to be subjected to uniform light treatment, so that the urban building image with uniform light receiving is obtained.
The light homogenizing algorithm is a common method for homogenizing urban building images with uneven brightness distribution, and the method enhances the detail contrast of the images and inhibits the abnormal brightness change of the images by enhancing high-frequency information and inhibiting low-frequency information so as to achieve the aim of uniform brightness distribution,The dodging algorithm carries out low-pass filtering on the original image through a Gaussian filter, so that a background image simulating brightness distribution is obtained, the original image and the background image are subjected to difference, and an image with uniform light receiving is obtained.
The method comprises the steps of determining the uniform light effect and definition of an image with uniform light receiving, wherein the selection of a Gaussian filter determines the uniform light effect and definition of the image with uniform light receiving, so that the accurate analysis of urban architecture is affected, the Gaussian filter is obtained by a two-dimensional Gaussian model, the larger the standard deviation of the two-dimensional Gaussian model is, the larger the size of the Gaussian filter is, the larger the number of pixel points participating in the simulation brightness distribution is, the more the distribution is scattered, the degree of blurring of the filtered pixel points is larger, the finally obtained background image contains fewer detail features, the simulation effect on the brightness distribution in the urban architecture image is poor, a large number of detail features are reserved in the urban architecture image with uniform light receiving, which is obtained by making the difference between an original image and the background image, but the uniform light receiving effect is poor, the smaller the standard deviation of the two-dimensional Gaussian filter is, the smaller the number of pixel points participating in the simulation brightness distribution is, the number of the pixel points participating in the simulation brightness distribution is smaller, the brightness distribution is better simulated, but the finally obtained background image can contain more detail features, and the obtained background image can better simulate the brightness distribution in the urban architecture image, the detail distribution is lost, namely the detail image with high uniform light receiving in the urban architecture image obtained by making the uniform light receiving, and the detail image with uniform light receiving.
The method comprises the steps of obtaining detail subgraphs of different sizes in multiple directions by decomposing the collected city building images, obtaining the detail expression level of the target pixel point in each direction according to the expression level of the detail characteristic of the different sizes and the pixel value of the corresponding pixel point in the detail subgraphs of the target pixel point in each direction, taking the detail expression level of each direction as the modular length of the detail vector in each direction, taking the sum of the detail vectors in all directions as the comprehensive detail vector, obtaining a Gaussian filter of the target pixel point according to the comprehensive detail vector, carrying out filtering treatment on each pixel point according to the Gaussian filter, subtracting the city building image from the obtained background image, and obtaining the city building image with uniform light, thereby ensuring that the obtained city building image with uniform light has better definition and light-homogenizing effect, and further ensuring the accuracy of analyzing the city building.
S1, inspection and acquisition of urban building images through an unmanned aerial vehicle.
Specifically, when patrolling and examining through unmanned aerial vehicle, unmanned aerial vehicle is through carrying on many cameras on a flight platform, gathers urban building image from the multi-angle, acquires the complete accurate information of urban building.
For the collected city building image, the pixel point at the upper left corner in the city building image is taken as an original point, the original point is taken as the positive direction of the horizontal axis to the right, and the original point is taken as the positive direction of the vertical axis to the down, so that a rectangular coordinate system is constructed, and the value range of the abscissa of the pixel point in the city building image isThe range of the ordinate isAnd (2) andFor the length of the image of the urban building,Is the width of the urban building image.
Illustratively, as shown in FIG. 2, a schematic view of the acquired city building image is shown.
S2, acquiring a Gaussian filter of each pixel point in the urban building image.
Specifically, by analyzing the urban building image to obtain the gaussian filter of each pixel point in the urban building image, in step S2, a flowchart of obtaining the gaussian filter of each pixel point in the urban building image is obtained, and referring to fig. 3, the flowchart includes steps S201 to S205, specifically:
s201, decomposing the urban building image through wavelet transformation to obtain detail subgraphs with different sizes in all directions.
The method and the device can obtain detail subgraphs of the urban building image in multiple scales and directions by utilizing the multi-resolution characteristic of wavelet transformation, so that the strength of detail features of local areas of each pixel point is analyzed.
Specifically, byThe wavelet transformation decomposes the urban building image, and performs the process togetherSub-decomposing to obtain detail subgraphs of different sizes in all directions, wherein the number of kinds of all sizes is equal toIn the present embodiment, the number of times of decompositionIn other embodiments, the number of decompositions can be set according to the actual application scenario and requirementsWavelet transformation is a well-known technique and will not be described in detail herein.
Wherein the directions include a vertical direction, a horizontal direction and a diagonal direction, and thus the obtained detail subgraphs of each direction include the detail subgraphs of the vertical direction, the detail subgraphs of the horizontal direction and the detail subgraphs of the diagonal direction, respectively, the vertical direction, the horizontal direction and the diagonal direction are marked as the firstDirection, the firstDirection and the firstDirection.
Wherein, the firstThe size of the detail subgraph of each direction obtained by sub-decomposition isWill beIs marked as the firstSize, thenIs the firstLength of dimension, and,Is the firstWidth of dimension, and,For the length of the image of the urban building,Is the width of the urban building image.
Wherein the pixel points in the detail subgraph have pixel values, a higher pixel value generally indicates the presence of an edge or texture, and a lower pixel value indicates the presence or absence of an edge or texture, so the pixel values of the pixel points reflect the detail features of the local area of the pixel points in the detail subgraph, the detail features comprise edge features and texture features, and the larger the pixel value is, the greater the intensity of the detail features of the local area of the pixel points is.
The detail subgraph in the vertical direction comprises a low-frequency component of the urban building image in the horizontal direction and a high-frequency component of the urban building image in the vertical direction, so that the detail subgraph in the vertical direction comprises a detail characteristic of the urban building image in the vertical direction, the detail subgraph in the horizontal direction comprises a high-frequency component of the urban building image in the horizontal direction and a low-frequency component of the urban building image in the vertical direction, and the detail subgraph in the diagonal direction comprises a detail characteristic of the urban building image in the horizontal direction and a high-frequency component of the urban building image in the vertical direction, so that the detail subgraph in the diagonal direction comprises a detail characteristic of the urban building image in the diagonal direction.
S202, calculating the expressive degree of different sizes on the detail features.
The pixel values of the pixel points in the detail subgraph can show detail features in the city building image, and the pixel values of the pixel points in the detail subgraphs with different sizes are different in the degree of showing the detail features in the city building image, namely, the smaller the degree of showing the detail features in the city building image is by the pixel values of the pixel points in the detail subgraphs with larger sizes, the larger the degree of showing the detail features in the city building image is by the pixel values of the pixel points in the detail subgraphs with smaller sizes, so that the degree of showing the detail features with different sizes can be calculated according to the ratio of the different sizes to the sizes of the city building image.
Specifically, according to the ratio of different sizes to the sizes of the urban building images, calculating the expressive degree of the different sizes on the detailed characteristics, and then the firstThe degree of performance of the size on the detail features satisfies the expression:
;
In the formula, Is the firstThe degree of performance of the size with respect to the detail features,Is the number of times of decomposition, and the number of kinds of all sizes is equal to,As a function of the natural index of refraction,Is the firstThe dimensions of the product are such that,Is the firstLength of dimension, and,Is the firstWidth of dimension, and,For the size of the image of the urban building,For the length of the image of the urban building,Is the width of the urban building image.
S203, calculating the detail expression degree of the pixel points in the urban building image in different directions.
Specifically, for the coordinates in the urban building imageIs used for the display of the display panel, the first pixel point in each directionThe coordinates of the corresponding pixel points in the detail subgraph of the size areWherein, the method comprises the steps of, wherein,Is a function of rounding up, wherein,The abscissa and the ordinate of the pixel point, respectively.
Further, according to the expressive degree of detail features of different sizes, the coordinates in the urban building image are as followsThe pixel values of the corresponding pixel points in the detail subgraphs with different sizes in each direction are weighted to obtain the detail expression degree of the pixel points in each direction, and the coordinates in the urban building image are as followsIs at the first pixel pointThe detail expression degree of each direction satisfies the expression:
;
In the formula, For coordinates in city building imagesIs at the first pixel pointThe degree of detail of the individual directions is expressed,Is the number of times of decomposition, and the number of kinds of all sizes is equal to,Is the firstThe degree of performance of the size with respect to the detail features,Is the firstFirst of all directionsThe coordinates in the detail drawing of the size areIs used for the pixel values of the pixel points of (a),For coordinates in city building imagesThe pixel point of (2) is in the first directionCoordinates of corresponding pixel points in the detail drawing of the size,As a round-up function.
The degree of representation of the detail features by different sizesFor the pixel value of the corresponding pixel point in the detail subgraphs with different sizesWeighting to obtain coordinates of city building imageIs at the first pixel pointDetail level of each directionWherein the pixel value of the pixel reflects the detail feature of the local area of the pixel in the detail drawing, and the larger the pixel value is, the larger the intensity of the detail feature of the local area of the pixel is, and the larger the intensity of the detail feature of the local area of the pixel is, the coordinates in the urban building image areIs at the first pixel pointDetail level of each directionThe larger the (first)Degree of performance of size versus detail featuresThe larger the city building image is, the coordinates areIs at the first pixel pointDetail level of each directionThe larger.
S204, synthesizing the detail expression degree of the pixel points in the urban building image in all directions to obtain the comprehensive detail vector of the pixel points.
Specifically, coordinates in the urban building image are as followsIs at the first pixel pointDetail level of each directionAs the pixel point at the firstDetail vector of each directionIs to be at the same timeIn the first direction as the pixel pointDetail vector of each directionTo obtain the coordinates in the city building image asIs at the first pixel pointDetail vector of each direction
Further, coordinates in the city building image are calculated asThe sum of the detail vectors of the pixel points in all directions is taken as the comprehensive detail vector of the pixel pointsThe result of summing the plurality of vectors is simply one vector, and summing the plurality of vectors is a known technique, and will not be described herein.
It should be noted that, by integrating the detail vectors of the pixel points in all directions, an integrated detail vector of the pixel point is obtained, and the integrated detail vector comprehensively reflects the distribution condition of the detail features of the local area of the pixel point.
S205, determining a Gaussian filter of the pixel point according to the comprehensive detail vector of the pixel point in the urban building image.
Specifically, for pixel points in an urban building image, a Gaussian filter of the pixel points is obtained according to a two-dimensional Gaussian model, the two dimensions are respectively a first dimension and a second dimension, the directions of the two dimensions are perpendicular to each other, each dimension is provided with two parameters, and the two parameters are respectively a mean value and a standard deviation.
Further, for the coordinates in the city building image asIs used for integrating the detailed vectors of the pixel pointsThe direction of (2) is recorded asObtaining a two-dimensional Gaussian model according to the comprehensive detail vector of the pixel point
Wherein the two-dimensional Gaussian modelThe directions of the two dimensions of (a) are respectivelyAndAnd the direction isAnd direction ofAre mutually perpendicular.
In the gaussian filter, since the larger the weight value along the two dimensions is, the larger the influence of the pixel value of the other pixel at the position where the weight value is larger is on the pixel value of the filtered pixel, the smaller the influence of the pixel value of the other pixel at the position where the weight value is smaller is on the pixel value of the other pixel, and the comprehensive detail vector of the pixel comprehensively reflects the distribution of the detail features of the local area of the pixel, so if the direction of the comprehensive detail vector of the pixel is to be determinedAndAs the vertical direction of two dimensions of the Gaussian filter, the pixel value of the filtered pixel point keeps more detail features, so that a great amount of detail features are lost in the urban building image with uniform light receiving obtained by differencing the original image and the background image, namely the definition of the obtained urban building image with uniform light receiving is not highAdding toAnd the vertical direction of the pixel point is taken as the direction of two dimensions of the Gaussian filter, so that the greater the blurring degree of the filtered pixel point is, the fewer detail features are contained in the finally obtained background image, and the definition of the obtained urban building image with uniform light receiving is enhanced.
Wherein the two-dimensional Gaussian modelEach dimension of (2) has two parameters, namely a mean value and a standard deviation, and a two-dimensional Gaussian modelThe mean value of the two dimensions of (a) is 0, and the standard deviation isAndThe expression is satisfied:
;
;
Wherein, in the formula, wherein, Comprehensive detail vectors of the pixel points respectivelyThe length of the projection in the first dimension and the second dimension.
The pixel points with greater detail expression level are integrated with detail vectorsThe longer the modulo length of (2), the more detail vectors are synthesizedThe longer the projection in the first and second dimensions, the longer the length of the projection in the first dimension, and in this case the two-dimensional gaussian modelThe larger the standard deviation of each dimension of the pixel points, the larger the size of the Gaussian filter, the larger the number of pixel points participating in simulating brightness distribution, but the more discrete the distribution, the larger the blurring degree of the filtered pixel points, and the fewer detail features contained in the local area of the pixel points in the finally obtained background image, thereby enhancing the definition of the obtained urban building image with uniform light receiving, and for the pixel points with smaller detail expression degree, the comprehensive detail vectorThe shorter the modular length of (2), the more detail vectors are synthesizedThe shorter the length of the projection in the first and second dimensions, the two-dimensional gaussian model at this timeThe smaller the standard deviation of each dimension of the background image is, the smaller the size of the Gaussian filter is, the smaller the number of pixel points participating in the simulation brightness distribution is, but the more concentrated the distribution is, and the better the effect of simulating the brightness distribution in the local area of the pixel points in the finally obtained background image is, so that the light homogenizing effect of the obtained urban building image with uniform light receiving is enhanced.
Further, according to a two-dimensional Gaussian modelA Gaussian filter for obtaining a target pixel point, the Gaussian filter having a matrix, wherein the number of lines of the Gaussian filter in a first dimension isThe number of rows in the second dimension is,Respectively two-dimensional Gaussian modelsIs defined as the standard deviation of the two dimensions of (1),In order to round up the function,Is an expansion coefficient, in the present embodiment, an expansion coefficientIn other embodiments, the number of decompositions may be set according to the actual application scenario and requirements.
Exemplary are a schematic diagram of a two-dimensional gaussian model having a standard deviation of 0.5 in both the first and second dimensions, a schematic diagram of a gaussian filter obtained from the two-dimensional gaussian model shown in fig. 4, a schematic diagram of a two-dimensional gaussian model having a standard deviation of 0.25 in both the first and second dimensions, as shown in fig. 6, and a schematic diagram of a gaussian filter obtained from the two-dimensional gaussian model shown in fig. 6, as shown in fig. 7.
S3, according to the Gaussian filter of each pixel point in the urban building image, passingThe light homogenizing algorithm obtains the urban building image with uniform light receiving.
Specifically, the Gaussian filter according to each pixel point in the urban building image passesThe method comprises the steps of filtering each pixel point according to a Gaussian filter of each pixel point in the urban building image to obtain a background image formed by the filtered pixel points, simulating brightness distribution in the background image, subtracting the urban building image from the obtained background image, and correspondingly stretching and displaying the subtracted image to finally obtain the urban building image with uniform light receiving.
The method and the device have the advantages that the detail expression degree of the target pixel point in each direction is obtained according to the expression degree of the detail features of different sizes and the pixel values of the corresponding pixel points in the detail subgraphs of the target pixel point in different sizes in each direction, the detail expression degree of the target pixel point in each direction is obtained, the comprehensive detail vectors are obtained according to the detail expression degrees of multiple directions, the Gaussian filter of the target pixel point is obtained according to the comprehensive detail vectors, for the pixel point with the larger detail expression degree, the standard deviation of the two-dimensional Gaussian model is larger, the detail features contained in the local area of the pixel point in the finally obtained background image are fewer, so that the definition of the obtained urban building image with uniform light is enhanced, for the pixel point with the smaller detail expression degree, the standard deviation of the two-dimensional Gaussian model is smaller, the effect of the local area of the pixel point in the finally obtained background image is better, the obtained urban building image with uniform light receiving effect is enhanced, and the obtained urban building image with uniform light receiving effect is guaranteed, and the following analysis is accurate.
Exemplary is a schematic diagram of a background image obtained by filtering the urban building image shown in fig. 2, as shown in fig. 8, and a schematic diagram of a uniformly light-receiving urban building image obtained by subtracting the urban building image shown in fig. 2 from the background image shown in fig. 9, as shown in fig. 9.
Among them, common methods of stretching display include, but are not limited to, 2% linear stretching, contrast parameter stretching, and gradient stretching displayThe light homogenizing algorithm, 2% linear stretching, contrast parameter stretching and gradient stretching are all known techniques, and will not be described here.
The whole contrast of the obtained subtracted image is reduced, especially in the darker area, in order to increase the whole contrast, and to maintain the consistency of the whole contrast of the subtracted image, to highlight the details in the image, to maintain the sharpness of the image, and to stretch the subtracted image.
The embodiment of the invention also discloses a city building image processing system based on unmanned aerial vehicle inspection, which comprises a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the city building image processing method based on unmanned aerial vehicle inspection is realized.
The above system further comprises other components well known to those skilled in the art, such as a communication bus and a communication interface, the arrangement and function of which are known in the art and therefore are not described in detail herein.
In the description of the present specification, the meaning of "a plurality", "a number" or "a plurality" is at least two, for example, two, three or more, etc., unless specifically defined otherwise.
While various embodiments of the present invention have been shown and described herein, it will be obvious to those skilled in the art that such embodiments are provided by way of example only. Many modifications, changes, and substitutions will now occur to those skilled in the art without departing from the spirit and scope of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in practicing the invention.

Claims (8)

1. The city building image processing method based on unmanned aerial vehicle inspection is characterized by comprising the following steps of:
decomposing the urban building image through wavelet transformation to obtain detail subgraphs with different sizes in multiple directions, and calculating the expressive degree of the detail features of the different sizes according to the ratio of the different sizes to the sizes of the urban building image;
According to the expressive degree of the detail characteristics of different sizes, the pixel values of the corresponding pixel points in the detail subgraphs of different sizes of the target pixel points in each direction are weighted and summed to obtain the detailed expressive degree of the target pixel points in each direction, and the expression is satisfied:
;
In the formula, As coordinates of the target pixel point in the urban building image,The abscissa and the ordinate of the target pixel point are respectively,Is at the first pixel pointThe degree of detail of the individual directions is expressed,Is the number of times of decomposition, and the number of kinds of all sizes is equal to,Is the firstThe degree of performance of the size with respect to the detail features,Is the firstFirst of all directionsThe coordinates in the detail drawing of the size areIs used for the pixel values of the pixel points of (a),Is an upward rounding function;
The detail expression degree of the target pixel point in each direction is used as the modular length of the detail vector of the target pixel point in each direction, the sum of the detail vectors of the target pixel point in all directions is calculated and used as the comprehensive detail vector of the target pixel point, and a two-dimensional Gaussian model is obtained according to the comprehensive detail vector ,The directions of the two dimensions of (a) are respectivelyAndStandard deviation of two dimensions is equal toAndWherein, the method comprises the steps of, wherein,In order to integrate the direction of the detail vector,The length of the projection of the comprehensive detail vector in the first dimension and the second dimension respectively according to a two-dimensional Gaussian modelObtaining a Gaussian filter of a target pixel point;
the method comprises the steps of carrying out low-pass filtering treatment on each pixel point according to a Gaussian filter of each pixel point in the urban building image to obtain a background image, subtracting the urban building image from the background image to obtain the urban building image with uniform light receiving, and the method comprises the following steps:
And correspondingly stretching and displaying the subtracted image to finally obtain the urban building image with uniform light receiving, wherein the stretching and displaying method comprises 2% linear stretching, contrast parameter stretching and gradient stretching display.
2. The method for processing the urban building image based on unmanned aerial vehicle inspection according to claim 1, wherein the decomposing the urban building image by wavelet transformation to obtain detail subgraphs with different sizes in a plurality of directions comprises:
When decomposing the city building image by wavelet transformation, the method is carried out together Sub-decomposition, the number of kinds of all sizes is equal toWherein, the method comprises the steps of,First, theThe size of the detail subgraph of each direction obtained by sub-decomposition isWill beIs marked as the firstSize, thenIs the firstThe length of the dimension is such that,Is the firstWidth of dimension, and,,For the length of the image of the urban building,The width of the urban building image;
Wherein the plurality of directions include a vertical direction, a horizontal direction, and a diagonal direction, and the vertical direction, the horizontal direction, and the diagonal direction are respectively referred to as a first Direction, the firstDirection and the firstDirection.
3. The unmanned aerial vehicle inspection-based urban building image processing method according to claim 1, wherein the expressive degree of detail features of the different sizes satisfies the expression:
;
In the formula, Is the firstThe degree of performance of the size with respect to the detail features,Is the number of times of decomposition, and the number of kinds of all sizes is equal to,As a function of the natural index of refraction,Is the firstThe dimensions of the product are such that,Is the firstThe length of the dimension is such that,Is the firstWidth of dimension, and,,For the size of the image of the urban building,For the length of the image of the urban building,Is the width of the urban building image.
4. The method for processing urban building images based on unmanned aerial vehicle inspection according to claim 1, wherein the first step isFirst of all directionsThe coordinates in the detail drawing of the size areIs that the target pixel is at the first pixelFirst of all directionsAnd the corresponding pixel points in the detail subgraph of the size.
5. The method for processing urban building images based on unmanned aerial vehicle inspection according to claim 1, wherein the two-dimensional gaussian modelHas two parameters, namely a mean value and a standard deviation, and a two-dimensional Gaussian modelIs of two dimensions the mean values were all 0.
6. The unmanned aerial vehicle inspection-based city building image processing method of claim 1, wherein the gaussian filter is essentially a matrix, wherein the number of lines of the gaussian filter in the first dimension isThe number of rows in the second dimension is,Respectively two-dimensional Gaussian modelsIs defined as the standard deviation of the two dimensions of (1),In order to round up the function,Is the expansion coefficient.
7. The unmanned aerial vehicle inspection-based urban building image processing method according to claim 1, wherein the method comprises:
the pixel point at the upper left corner in the urban building image is taken as an origin, the origin is taken as the positive direction of the horizontal axis to the right, and the origin is taken as the positive direction of the vertical axis to the down, so that a rectangular coordinate system is constructed, and the value range of the abscissa of the pixel point in the urban building image is The range of the ordinate isAnd (2) andFor the length of the image of the urban building,Is the width of the urban building image.
8. An unmanned aerial vehicle inspection-based urban building image processing system, comprising a processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement an unmanned aerial vehicle inspection-based urban building image processing method according to any one of claims 1-7.
CN202411884141.4A 2024-12-20 2024-12-20 Urban building image processing method and system based on unmanned aerial vehicle inspection Active CN119338695B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202411884141.4A CN119338695B (en) 2024-12-20 2024-12-20 Urban building image processing method and system based on unmanned aerial vehicle inspection

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202411884141.4A CN119338695B (en) 2024-12-20 2024-12-20 Urban building image processing method and system based on unmanned aerial vehicle inspection

Publications (2)

Publication Number Publication Date
CN119338695A CN119338695A (en) 2025-01-21
CN119338695B true CN119338695B (en) 2025-03-07

Family

ID=94271509

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202411884141.4A Active CN119338695B (en) 2024-12-20 2024-12-20 Urban building image processing method and system based on unmanned aerial vehicle inspection

Country Status (1)

Country Link
CN (1) CN119338695B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101777179A (en) * 2010-02-05 2010-07-14 电子科技大学 Method for de-noising dual-tree complex wavelet image on basis of partial differential equation
CN104574308A (en) * 2014-12-30 2015-04-29 北京工业大学 SAR image denoising method based on sampling matrix direction optimization

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8280185B2 (en) * 2008-06-27 2012-10-02 Microsoft Corporation Image denoising techniques
CN115526799A (en) * 2022-09-30 2022-12-27 重庆邮电大学 Endoscopic Image Enhancement Method Based on Improved Retinex and Weighted Guided Filter

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101777179A (en) * 2010-02-05 2010-07-14 电子科技大学 Method for de-noising dual-tree complex wavelet image on basis of partial differential equation
CN104574308A (en) * 2014-12-30 2015-04-29 北京工业大学 SAR image denoising method based on sampling matrix direction optimization

Also Published As

Publication number Publication date
CN119338695A (en) 2025-01-21

Similar Documents

Publication Publication Date Title
US8179393B2 (en) Fusion of a 2D electro-optical image and 3D point cloud data for scene interpretation and registration performance assessment
Chen et al. Automatic building information model reconstruction in high-density urban areas: Augmenting multi-source data with architectural knowledge
Gruszczyński et al. Comparison of low-altitude UAV photogrammetry with terrestrial laser scanning as data-source methods for terrain covered in low vegetation
Müller-Linow et al. The leaf angle distribution of natural plant populations: assessing the canopy with a novel software tool
JP2023534175A (en) Neural network analysis of LFA specimens
CN116228780B (en) Silicon wafer defect detection method and system based on computer vision
CN113902812A (en) Laser radar and camera external parameter automatic calibration method based on multiple calibration plates
CN108447058A (en) A kind of image quality evaluating method and system
CN119477770B (en) Three-dimensional Gaussian splatter-based orthographic image generation method, device and storage medium
CN109766896B (en) Similarity measurement method, device, equipment and storage medium
CN113706400B (en) Image correction method, device, microscope image correction method and electronic equipment
CN119338695B (en) Urban building image processing method and system based on unmanned aerial vehicle inspection
Budianti et al. Background blurring and removal for 3d modelling of cultural heritage objects
CN117911512A (en) Camera pose relation determining method, point cloud fusion method and system thereof
CN117115669A (en) Object-level ground object sample self-adaptive generation method and system with double-condition quality constraint
CN117495958A (en) Warehouse box image recognition positioning method and system based on target detection network
CN116596820A (en) Remote sensing image mosaic method based on image quality
CN116612097A (en) Method and system for predicting internal section morphology of wood based on surface defect image
CN116092035A (en) Lane line detection method, lane line detection device, computer equipment and storage medium
Pargieła et al. The Utilization of Synthetic and Semisynthetic Point Clouds and Images for Testing Novel Approaches for Correcting LIDAR Data
CN119091076B (en) Three-dimensional building model generation method, system, equipment and storage medium
CN118965658B (en) Arrangement method and system for transmission line crossing measurement
CN119579805B (en) A method and system for three-dimensional reconstruction of urban buildings based on virtual modeling
CN119757793B (en) A glacier monitoring method and system based on binocular 3D reconstruction
Jarman et al. Hyperspectral neural radiance fields for 3D scene understanding of gas plumes

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