CN115272174A - Municipal road detection method and system - Google Patents
Municipal road detection method and system Download PDFInfo
- Publication number
- CN115272174A CN115272174A CN202210675220.9A CN202210675220A CN115272174A CN 115272174 A CN115272174 A CN 115272174A CN 202210675220 A CN202210675220 A CN 202210675220A CN 115272174 A CN115272174 A CN 115272174A
- Authority
- CN
- China
- Prior art keywords
- image
- noise reduction
- pixel
- target image
- sub
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 42
- 238000012545 processing Methods 0.000 claims abstract description 75
- 230000009467 reduction Effects 0.000 claims abstract description 71
- 238000000034 method Methods 0.000 claims abstract description 22
- 238000012216 screening Methods 0.000 claims abstract description 22
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 21
- 230000001537 neural effect Effects 0.000 claims abstract description 7
- 230000012010 growth Effects 0.000 claims description 19
- 230000035040 seed growth Effects 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 claims description 6
- 239000006185 dispersion Substances 0.000 claims description 6
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 238000007689 inspection Methods 0.000 claims description 2
- 230000007547 defect Effects 0.000 abstract description 10
- 230000008569 process Effects 0.000 description 9
- 238000000354 decomposition reaction Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000004590 computer program Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000005764 inhibitory process Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
-
- G06T5/70—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Abstract
The invention discloses a municipal road detection method, which comprises the following steps: s1, acquiring a surface image of a municipal road; s2, carrying out graying processing on the surface image to obtain a grayed image; s3, carrying out self-adaptive region growing processing on the gray images to obtain a plurality of sub-regions; s4, screening the sub-regions according to a set screening rule to obtain a plurality of target images; s5, respectively carrying out noise reduction processing on each target image by using a self-adaptive noise reduction algorithm to obtain a noise-reduced target image; and S6, inputting the target image subjected to noise reduction into a convolution neural model trained in advance to carry out detection, and obtaining a detection result. The invention also discloses a system for realizing the method. The invention effectively reduces the number of pixel points entering the noise reduction processing step, thereby effectively improving the efficiency of using machine vision to detect the defects of the municipal roads.
Description
Technical Field
The invention relates to the field of road detection, in particular to a municipal road detection method and system.
Background
Road inspection is an important component of municipal road maintenance work. Traditional road detection generally detects through artificial mode, but the efficiency that human eyes visually inspected is than lower, is unfavorable for discovering defects such as crack, pot hole in the road in time. With the development of technology, ways of detecting roads by machine vision now appear. Compared with the traditional human eye visual detection, the machine visual detection mode enables the efficiency of detecting the municipal road to be obviously improved. The existing machine vision detection mode adopts a single noise reduction mode to reduce noise of the whole image in the noise reduction process of the road image, so that the number of pixels needing noise reduction is too large, and the efficiency of defect detection of the municipal road by using machine vision is reduced.
Disclosure of Invention
The invention aims to disclose a municipal road detection method, which solves the problems that in the prior art, in the process of denoising a road image, a single denoising mode is adopted to denoise the whole image, so that the number of pixel points needing denoising is too large, and the efficiency of detecting defects of the municipal road by using machine vision is reduced.
In order to achieve the purpose, the invention adopts the following technical scheme:
in one aspect, the invention provides a municipal road detection method, comprising:
s1, acquiring a surface image of a municipal road;
s2, carrying out graying processing on the surface image to obtain a grayed image;
s3, carrying out self-adaptive region growing processing on the gray images to obtain a plurality of sub-regions;
s4, screening the sub-regions according to a set screening rule to obtain a plurality of target images;
s5, performing noise reduction processing on each target image by using a self-adaptive noise reduction algorithm to obtain a noise-reduced target image;
and S6, inputting the target image subjected to noise reduction into a convolution neural model trained in advance for detection, and obtaining a detection result.
Preferably, the S2 includes:
performing graying processing on the surface image by adopting the following formula:
Gray(x,y)＝w1×R(x,y)+w2×G(x,y)+w3×B(x,y)
in the formula, gray represents a Gray image, (x, y) represents coordinates of a pixel point, gray (x, y) represents a pixel value of a pixel point with coordinates (x, y) in the Gray image, and w1、w2、w3And the calculation coefficients are preset, R (x, y), G (x, y) and B (x, y) respectively represent pixel values of pixel points with coordinates (x, y) in the images R, G and B, and R, G and B respectively represent images corresponding to red components, green components and blue components of the surface images in RGB color space.
Preferably, the S3 includes:
respectively calculating the occurrence frequency of each pixel value in the gray-scale image;
for the ith pixel value, if it corresponds to the frequency freqiIf the pixel value is less than the preset frequency threshold value, the ith pixel value is stored into a set setwtbr to be grown, i belongs to [1, N ]]Where N denotes pixel values in a greyscale imageA total number of types;
and respectively carrying out seed growth treatment on the pixel points corresponding to each pixel value in the setwttb r to obtain a plurality of sub-regions.
Preferably, the performing the seed growth processing on the pixel point corresponding to each pixel value in the setwtbr to obtain a plurality of sub-regions includes:
sequencing pixel values in the setwtbr from small to large to obtain an ordered set setwtbr';
1, seed growth treatment:
all pixel points corresponding to the 1 st element in the ordered set setwtbr' are stored into a set wtgr1；
At wtgr, respectively1Taking each pixel point in the image as a seed, and performing image area growth processing to obtain a plurality of sub-areas;
set wtgr1The pixel points in the sub-area are subjected to area growth processing, and the pixel points of all the sub-areas are stored into a set setallu;
and k seed growth treatment:
all pixel points corresponding to the kth element in the ordered set setwtbr' are stored in a set wtgrk；
Set wtgrkThe pixel point corresponding to the intersection of the set setallu is selected from the set wtgrkDelete, get set wtgrk'；
Respectively in wtgrkEach pixel point in the' is used as a seed to carry out image area growth processing to obtain a plurality of subregions;
set wtgrkThe pixel points in the' are subjected to region growing processing to obtain pixel points of all subregions, and the pixel points are stored in a set setallu;
where k ∈ [2, M ], M represents the total number of elements contained in setwtbr.
Preferably, the S4 includes:
storing the subareas with the area ratio larger than a preset ratio threshold into a set H;
and respectively acquiring the minimum circumscribed rectangle corresponding to each subregion in the set H, and taking the pixel points within the range of the minimum circumscribed rectangle as the pixel points of the target image corresponding to the subregion.
Preferably, the area ratio is calculated by the following formula:
in the formula, areart represents the area ratio of the sub-region, numofblokc represents the number of pixels included in the sub-region, and numofGray represents the number of pixels included in the grayscale image.
Preferably, the S5 includes:
for the target image P, a pixel value dispersion coefficient dispcoef of the target image P is calculatedP；
Based on dispcoefPAnd selecting an algorithm for carrying out noise reduction processing on the target image P.
Preferably, the pixel value dispersion coefficient dispcoefPThe calculation is made by the following formula:
in the formula, alpha and beta represent preset proportionality coefficients, setP represents a set of pixel points in the target image P, grayuThe pixel value of the pixel point u in setP is represented, numfsetP represents the total number of elements included in setP, stdevi represents a preset pixel value variance standard value, and numfedg represents the total number of edge pixel points in the target image P.
Preferably, said dispcoef-basedPThe algorithm for selecting the denoising processing on the target image P comprises the following steps:
if dispcoefPIf the difference is larger than the preset discrete coefficient threshold, carrying out noise reduction processing on the target image P by adopting an improved wavelet noise reduction algorithm to obtain a noise-reduced target image;
if dispcoefPIf the difference is less than or equal to the preset discrete coefficient threshold value, performing noise reduction processing on the target image P by adopting a median filter to obtain a reductionAnd (5) a noise target image.
On the other hand, the invention also provides a municipal road detection system which comprises a shooting module, a graying module, a seed growing module, a screening module, a noise reduction module and a detection module;
the shooting module is used for acquiring a surface image of the municipal road;
the graying module is used for performing graying processing on the surface image to obtain a grayed image;
the seed growing module is used for carrying out self-adaptive region growing processing on the gray images to obtain a plurality of sub-regions;
the screening module is used for screening the sub-regions according to a set screening rule to obtain a plurality of target images;
the noise reduction module is used for respectively carrying out noise reduction processing on each target image by using a self-adaptive noise reduction algorithm to obtain a noise-reduced target image;
and the detection module is used for inputting the target image subjected to noise reduction into a convolution neural model which is trained in advance to carry out detection so as to obtain a detection result.
In the process of processing the surface image of the municipal road, graying is carried out firstly, then self-adaptive region growing processing is carried out on the image obtained by graying to obtain a plurality of subregions, a plurality of target images are obtained from the subregions, then noise reduction processing is carried out on the target images, and finally road defect detection is carried out based on the target images after the noise reduction processing. The invention effectively reduces the number of pixel points entering the noise reduction processing step, thereby effectively improving the efficiency of using machine vision to detect the defects of the municipal roads.
Drawings
The invention is further illustrated by means of the attached drawings, but the embodiments in the drawings do not constitute any limitation to the invention, and for a person skilled in the art, other drawings can be obtained on the basis of the following drawings without inventive effort.
FIG. 1 is a diagram of an exemplary embodiment of a method of town road detection according to the present invention.
FIG. 2 is a diagram of an exemplary embodiment of a town road detection system according to the present invention.
Detailed Description
Reference will now be made in detail to embodiments of the present invention, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar function throughout. The embodiments described below with reference to the accompanying drawings are illustrative only for the purpose of explaining the present invention, and are not to be construed as limiting the present invention.
In one aspect, the present invention provides a method of town road detection, as illustrated in one embodiment in FIG. 1, comprising:
s1, acquiring a surface image of a municipal road;
s2, carrying out graying processing on the surface image to obtain a grayed image;
s3, carrying out self-adaptive region growing processing on the gray images to obtain a plurality of sub-regions;
s4, screening the sub-regions according to a set screening rule to obtain a plurality of target images;
s5, respectively carrying out noise reduction processing on each target image by using a self-adaptive noise reduction algorithm to obtain a noise-reduced target image;
and S6, inputting the target image subjected to noise reduction into a convolution neural model trained in advance for detection, and obtaining a detection result.
In the process of processing the surface image of the municipal road, graying is carried out firstly, then self-adaptive region growing processing is carried out on the image obtained by graying to obtain a plurality of subregions, a plurality of target images are obtained from the subregions, then noise reduction processing is carried out on the target images, and finally road defect detection is carried out based on the target images after the noise reduction processing. The invention effectively reduces the number of pixel points entering the noise reduction processing step, thereby effectively improving the efficiency of using machine vision to detect the defects of the municipal roads.
Preferably, the S2 includes:
performing graying processing on the surface image by adopting the following formula:
Gray(x,y)＝w1×R(x,y)+w2×G(x,y)+w3×B(x,y)
in the formula, gray represents a Gray image, (x, y) represents coordinates of a pixel point, gray (x, y) represents a pixel value of a pixel point with coordinates (x, y) in the Gray image, and w1、w2、w3And the calculation coefficients are preset, R (x, y), G (x, y) and B (x, y) respectively represent pixel values of pixel points with coordinates (x, y) in the images R, G and B, and R, G and B respectively represent images corresponding to red components, green components and blue components of the surface images in RGB color space.
Further, w1∈[0.29，031]，w2∈[0.58,0.61]，w3∈[0.09,0.12]。
Preferably, the S3 includes:
respectively calculating the occurrence frequency of each pixel value in the gray-scale image;
for the ith pixel value, if it corresponds to the frequency freqiIf the pixel value is less than the preset frequency threshold value, the ith pixel value is stored in a set setwtbr to be grown, i belongs to [1, N ]]N represents the total number of types of pixel values in the grayed image;
and respectively carrying out seed growth treatment on the pixel points corresponding to each pixel value in the setwttbr to obtain a plurality of sub-regions.
In the conventional region growing algorithm, the pixel points serving as seeds are generally considered to be selected, and obviously, the requirement is not met. Therefore, the region growing algorithm is improved, the pixel points serving as seeds are obtained by screening through the frequency of the pixel values, and automatic acquisition of the seed pixel points is achieved. In the process of obtaining, the invention discards the pixel value with overlarge frequency, thereby effectively compressing the number of seed pixel points and being beneficial to improving the operation speed. The larger the frequency is, the larger the probability that the pixel is taken as the pixel of the normal road surface area is, and the purpose of the invention is to detect the defect, so that the pixel of the defect area can be effectively obtained as the seed by setting the frequency threshold.
Further, the frequency of occurrence of the pixel values is equal to the total number of the pixel points corresponding to the pixel values divided by the total number of the pixel points in the grayed image.
Preferably, the performing the seed growth processing on the pixel point corresponding to each pixel value in the setwtbr respectively to obtain a plurality of sub-regions includes:
sequencing pixel values in setwtbr from small to large to obtain an ordered set setwtbr';
seed growth treatment for the 1 st time:
all pixel points corresponding to the 1 st element in the ordered set setwtbr' are stored into a set wtgr1；
At wtgr, respectively1Taking each pixel point in the image as a seed, and performing image area growth processing to obtain a plurality of sub-areas;
set wtgr1The pixel points in the sub-area are subjected to area growth processing, and the pixel points of all the sub-areas are stored into a set setallu;
and k seed growth treatment:
all pixel points corresponding to the kth element in the ordered set setwtbr' are stored in a set wtgrk；
Set wtgrkAnd the pixel point corresponding to the intersection of the set setallu is selected from the set wtgrkIs deleted to obtain a set wtgr'k；
Are respectively wtgr'kTaking each pixel point in the image as a seed, and performing image area growth processing to obtain a plurality of sub-areas;
set wtgr'kThe pixel points of all the sub-regions obtained by performing region growing processing on the pixel points in the set are stored in a set setallu;
where k ∈ [2, M ], M represents the total number of elements contained in setwtbr.
In the process of growing the seeds, the invention firstly carries out regional growth from the pixel point corresponding to the minimum pixel value, and because the pixel value of the noise is generally larger, the invention can reduce the influence of the noise on the growth of the seeds
The increasing of the pixel values of the sub-pixel points increases the number of the pixel points which have been selected as the growth area, wtgr'kNumber of pixels in (2) and wtgrkThe difference of the number of the pixel points in the method is larger and larger, so that the number of the seed pixel points entering the growth treatment is effectively restrained, and the speed of the regional growth is effectively improved.
Preferably, in the process of performing the region growing, whether the growing condition is satisfied is determined as follows:
for a pixel point seddpixel serving as a seed, storing pixel points in the neighborhood into a set setZ;
for pixel point z in setZ, the difference coefficient between the seedpixel and z is calculated:
difidseedpixel,z＝Φ×|Grayseedpixel-Grayz|+(1-Φ)×|neiTseedpixel-neiTz|
in the formula, difidseedpixel,zRepresents the difference coefficient between seddpixel and z, phi represents the duty ratio control parameter, phi is epsilon (0, 1), grayseedpixelAnd GrayzThe pixel values, neiT, representing sedxpixel and z, respectivelyseedpixelMean value of gradient amplitude, neiT, representing the 8 neighborhood of seddipixelszThe mean value of the gradient magnitudes representing the 8 neighborhood of z;
if the difference coefficient is smaller than the set difference coefficient threshold, it indicates that the pixel point z satisfies the growth condition, and if the difference coefficient is greater than or equal to the set difference coefficient threshold, it indicates that the pixel point z does not satisfy the growth condition.
The existing growth condition judgment is only to judge the difference between pixel values generally, but the image obtained after the region growth has more cavities only depending on the pixel values, so the invention also adds the parameter of gradient amplitude, realizes the effective inhibition to the cavities, and improves the accuracy of the result of the region growth.
Preferably, the S4 includes:
storing the subareas with the area ratio larger than a preset ratio threshold into a set H;
and respectively acquiring the minimum circumscribed rectangle corresponding to each subregion in the set H, and taking the pixel points within the range of the minimum circumscribed rectangle as the pixel points of the target image corresponding to the subregion.
Specifically, since the sub-region after the region growing process may only include foreground pixels, reference pixels are absent when performing noise reduction processing on the pixels in the edge region, and noise reduction processing cannot be performed on the pixels in the edge region, the problem is effectively solved by obtaining the minimum circumscribed rectangle. Because the noise pixel points in the image always exist singly or in small areas, partial noise pixel points can be removed in advance through screening of the area ratio, and the number of the pixel points which subsequently participate in image noise reduction is reduced.
Preferably, the area ratio is calculated by the following formula:
in the formula, areart represents the area ratio of the sub-region, numofblokc represents the number of pixels included in the sub-region, and numofGray represents the number of pixels included in the grayscale image.
Preferably, the S5 includes:
for the target image P, a pixel value dispersion coefficient dispcoef of the target image P is calculatedP；
Based on dispcoefPAnd selecting an algorithm for carrying out noise reduction processing on the target image P.
The existing noise reduction processing mode generally uses the same noise reduction mode to perform noise reduction processing on all pixel points in a gray level image, but the distribution of pixel values is not considered in the noise reduction mode, the same noise reduction processing mode is used, if the noise reduction processing process is more complex, the time required by noise reduction is very long, and if the noise reduction processing process is simpler, the noise reduction effect is poorer. Therefore, after the target image is obtained, the balance between the noise reduction effect and the noise reduction speed is realized by the method of obtaining the noise reduction processing through the pixel value discrete coefficient.
Preferably, the pixel value dispersion coefficient dispcoefPThe calculation is made by the following formula:
in the formula, alpha and beta represent preset proportionality coefficients, setP represents a set of pixel points in the target image P, grayuThe pixel value of the pixel point u in setP is represented, numfsetP represents the total number of elements included in setP, stdevi represents a preset pixel value variance standard value, and numfedg represents the total number of edge pixel points in the target image P.
When the pixel value discrete coefficient is calculated, the pixel value discrete coefficient can more accurately represent the distribution condition of the pixel values among the pixels in the target image by calculating from two aspects of the difference of the pixel values and the difference of the edge pixels.
Preferably, said dispcoef-basedPThe algorithm for selecting the noise reduction processing of the target image P comprises the following steps:
if dispcoefPIf the difference is larger than the preset discrete coefficient threshold value, carrying out noise reduction processing on the target image P by adopting an improved wavelet noise reduction algorithm to obtain a noise-reduced target image;
if dispcoefPAnd if the difference is less than or equal to the preset discrete coefficient threshold value, performing noise reduction processing on the target image P by adopting a median filter to obtain a noise-reduced target image.
The larger the difference between the pixel points is, the more the number of the edge pixel points is, the more unbalanced the distribution of the expressed pixel values is, the larger the discrete coefficient of the pixel values is, the more complex denoising algorithm needs to be used for denoising so as to ensure the denoising effect, otherwise, the more balanced the distribution of the expressed pixel values is, and the simpler algorithm with the higher denoising speed is used for denoising so as to improve the denoising speed.
Preferably, the performing denoising processing on the target image P by using the improved wavelet denoising algorithm to obtain a denoised target image includes:
performing D-layer wavelet decomposition on the target image to obtain a plurality of wavelet high-frequency coefficients and 1 wavelet low-frequency coefficient;
and respectively carrying out the following processing on each wavelet high-frequency coefficient to obtain a denoised wavelet high-frequency coefficient:
if | smwafrd,c|≥thrsdd,cThen, the wavelet high-frequency coefficient is calculated by adopting the following formula:
if | smwafrd,c|＜thrsdd,cThen, the wavelet high-frequency coefficient is calculated by adopting the following formula:
afsmwafrd,c＝0
in the formula, smwafrd,cRepresents the c-th wavelet high frequency coefficient obtained by the d-th wavelet decomposition, c is equal to [1,3 ]]，afsmwafrd,cRepresents pair smwafrd,cThe wavelet high-frequency coefficient obtained after denoising is obtained, delta represents a preset first control coefficient, delta is larger than 0,
which represents a preset second control coefficient,
thrsdd,cthe self-adaptive threshold value of the high-frequency coefficient of the c-th wavelet obtained by decomposing the d-th wavelet is represented, hfn represents a judgment function, and if smwafrd,cIf it is greater than 0, hfn (smwafr)d,c) Has a value of 1, if smwafrd,cIf it is less than 0, hfn (smwafr)d,c) Has a value of-1, if smwafrd,cEqual to 0, then hfn (smwafr)d,c) Is 0; d is an element of [1, D ]]；
And performing wavelet reconstruction on the wavelet low-frequency coefficient and the denoised wavelet high-frequency coefficient to obtain a denoised target image.
When the noise reduction processing is carried out, the noise reduction processing is carried out on the wavelet high-frequency coefficients under different conditions through the self-adaptive threshold value, so that the pertinence of a noise reduction processing function is improved, and the accuracy of a noise reduction effect is further improved. The setting mode of the noise reduction processing function can avoid the problem that the traditional threshold has breakpoints, so that the noise reduction effect is deviated. Compared with the hard threshold, when the wavelet high-frequency coefficient is smaller, the self-adaptive threshold obtained by the invention is continuous, and the occurrence of breakpoints is avoided.
Preferably, the adaptive threshold is calculated by using the following formula:
in the formula, thrsdd,cAn adaptive threshold value zm representing the c-th wavelet high frequency coefficient obtained by the d-th wavelet decompositiondRepresents the maximum value am of wavelet high-frequency coefficient obtained by wavelet classification of the d-th layerdDenotes the minimum value, lm, of the wavelet high frequency coefficient obtained by the wavelet hierarchy of the d-th layerdRepresents the intermediate value of the wavelet high-frequency coefficient obtained by the wavelet grading of the d-th layer, and omega represents a preset constant coefficient.
The threshold value of the invention is related to each wavelet high-frequency coefficient obtained by wavelet decomposition of the same layer, and the adaptability of the obtained threshold value is improved by setting the maximum value, the intermediate value and the minimum value to be equal to parameters directly related to the wavelet decomposition result.
On the other hand, as shown in an embodiment of fig. 2, the invention also provides a municipal road detection system, which comprises a shooting module, a graying module, a seed growing module, a screening module, a noise reduction module and a detection module;
the shooting module is used for acquiring a surface image of the municipal road;
the graying module is used for performing graying processing on the surface image to obtain a grayed image;
the seed growing module is used for carrying out self-adaptive region growing treatment on the gray images to obtain a plurality of sub-regions;
the screening module is used for screening the sub-regions according to a set screening rule to obtain a plurality of target images;
the noise reduction module is used for respectively carrying out noise reduction processing on each target image by using a self-adaptive noise reduction algorithm to obtain a noise-reduced target image;
and the detection module is used for inputting the target image subjected to noise reduction into a convolution neural model which is trained in advance to carry out detection so as to obtain a detection result.
While embodiments of the invention have been shown and described, it will be understood by those skilled in the art that: various changes, modifications, substitutions and alterations can be made to the embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.
It should be noted that, functional units/modules in the embodiments of the present invention may be integrated into one processing unit/module, or each unit/module may exist alone physically, or two or more units/modules are integrated into one unit/module. The integrated units/modules may be implemented in the form of hardware, or may be implemented in the form of software functional units/modules.
From the above description of embodiments, it is clear for a person skilled in the art that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code or any appropriate combination thereof. For a hardware implementation, a processor may be implemented in one or more of the following units: an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a processor, a controller, a microcontroller, a microprocessor, other electronic units designed to perform the functions described herein, or a combination thereof. For a software implementation, some or all of the flow of the embodiments may be accomplished by a computer program instructing the associated hardware.
In practice, the program may be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage media may be any available media that can be accessed by a computer. Computer-readable media can include, but is not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer.
Claims (10)
1. A method of town road inspection, comprising:
s1, acquiring a surface image of a municipal road;
s2, performing graying processing on the surface image to obtain a grayed image;
s3, carrying out self-adaptive region growing processing on the gray images to obtain a plurality of sub-regions;
s4, screening the sub-regions according to a set screening rule to obtain a plurality of target images;
s5, respectively carrying out noise reduction processing on each target image by using a self-adaptive noise reduction algorithm to obtain a noise-reduced target image;
and S6, inputting the target image subjected to noise reduction into a convolution neural model trained in advance to carry out detection, and obtaining a detection result.
2. The method of claim 1, wherein S2 comprises:
performing graying processing on the surface image by adopting the following formula:
Gray(x,y)＝w1×R(x,y)+w2×G(x,y)+w3×B(x,y)
in the formula, gray represents a Gray scale mapImage, (x, y) represents the coordinates of the pixel, gray (x, y) represents the pixel value of the pixel with coordinates (x, y) in the Gray-scale image, w1、w2、w3And the calculation coefficients are preset, R (x, y), G (x, y) and B (x, y) respectively represent pixel values of pixel points with coordinates (x, y) in the images R, G and B, and R, G and B respectively represent images corresponding to red components, green components and blue components of the surface images in RGB color space.
3. The method of claim 1, wherein said S3 comprises:
respectively calculating the occurrence frequency of each pixel value in the gray-scale image;
for the ith pixel value, if it corresponds to the frequency freqiIf the pixel value is less than the preset frequency threshold value, the ith pixel value is stored in a set setwtbr to be grown, i belongs to [1, N ]]N represents the total number of types of pixel values in the grayed-out image;
and respectively carrying out seed growth treatment on the pixel points corresponding to each pixel value in the setwttb r to obtain a plurality of sub-regions.
4. The method of claim 3, wherein the step of performing seed growth on the pixel points corresponding to each pixel value in setwtbr to obtain a plurality of sub-regions comprises:
sequencing pixel values in the setwtbr from small to large to obtain an ordered set setwtbr';
seed growth treatment for the 1 st time:
all pixel points corresponding to the 1 st element in the ordered set setwtbr' are stored into a set wtgr1；
At wtgr, respectively1Taking each pixel point in the image as a seed, and performing image area growth processing to obtain a plurality of sub-areas;
set wtgr1The pixel points of all the sub-regions obtained by performing region growing processing on the pixel points in the set are stored in a set setallu;
and k seed growth treatment:
all pixel points corresponding to the kth element in the ordered set setwtbr' are stored into a set wtgrk；
Set wtgrkThe pixel point corresponding to the intersection of the set setallu is selected from the set wtgrkGet the set wtgr'k；
Are respectively wtgr'kTaking each pixel point in the image as a seed, and performing image area growth processing to obtain a plurality of sub-areas;
set wtgr'kThe pixel points in the sub-area are subjected to area growth processing, and the pixel points of all the sub-areas are stored into a set setallu;
where k ∈ [2, M ], M represents the total number of elements contained in setwtbr.
5. The method of claim 1, wherein said S4 comprises:
storing the subareas with the area ratio larger than a preset ratio threshold into a set H;
and respectively acquiring the minimum circumscribed rectangle corresponding to each subregion in the set H, and taking the pixel points within the range of the minimum circumscribed rectangle as the pixel points of the target image corresponding to the subregion.
6. The method of claim 5, wherein the area fraction is calculated by the formula:
in the formula, areart represents the area ratio of the sub-region, numofblokc represents the number of pixels included in the sub-region, and numofGray represents the number of pixels included in the grayscale image.
7. The method of claim 1, wherein S5 comprises:
for the target image P, a pixel value dispersion coefficient dispcoef of the target image P is calculatedP；
Based on dispcoefPAnd selecting an algorithm for carrying out noise reduction processing on the target image P.
8. The method of claim 7, wherein the pixel value dispersion coefficient dispcoefPThe calculation is performed by the following formula:
in the formula, α and β represent preset proportionality coefficients, setP represents a set of pixel points in the target image P, grayuThe pixel value of the pixel point u in setP is represented, numfsetP represents the total number of elements included in setP, stdevi represents a preset pixel value variance standard value, and numfedg represents the total number of edge pixel points in the target image P.
9. The town road detection method according to claim 7, wherein the detection is based on dispcoefPThe algorithm for selecting the noise reduction processing of the target image P comprises the following steps:
if dispcoefPIf the difference is larger than the preset discrete coefficient threshold, carrying out noise reduction processing on the target image P by adopting an improved wavelet noise reduction algorithm to obtain a noise-reduced target image;
if dispcoefPAnd if the difference is less than or equal to the preset discrete coefficient threshold, carrying out noise reduction treatment on the target image P by adopting a median filter to obtain a noise-reduced target image.
10. A municipal road detection system is characterized by comprising a shooting module, a graying module, a seed growing module, a screening module, a noise reduction module and a detection module;
the shooting module is used for acquiring a surface image of the municipal road;
the graying module is used for performing graying processing on the surface image to obtain a grayed image;
the seed growing module is used for carrying out self-adaptive region growing processing on the gray images to obtain a plurality of sub-regions;
the screening module is used for screening the sub-regions according to a set screening rule to obtain a plurality of target images;
the noise reduction module is used for respectively carrying out noise reduction processing on each target image by using a self-adaptive noise reduction algorithm to obtain a noise-reduced target image;
and the detection module is used for inputting the target image subjected to noise reduction into a convolution neural model which is trained in advance to carry out detection so as to obtain a detection result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210675220.9A CN115272174B (en) | 2022-06-15 | 2022-06-15 | Municipal road detection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210675220.9A CN115272174B (en) | 2022-06-15 | 2022-06-15 | Municipal road detection method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115272174A true CN115272174A (en) | 2022-11-01 |
| CN115272174B CN115272174B (en) | 2023-05-19 |
Family
ID=83762003
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210675220.9A Active CN115272174B (en) | 2022-06-15 | 2022-06-15 | Municipal road detection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115272174B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115797473A (en) * | 2023-02-09 | 2023-03-14 | 国网山东省电力公司东营供电公司 | Concrete forming evaluation method for civil engineering |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105404847A (en) * | 2014-09-16 | 2016-03-16 | 北京计算机技术及应用研究所 | Real-time detection method for object left behind |
| CN106446914A (en) * | 2016-09-28 | 2017-02-22 | 天津工业大学 | Road detection based on superpixels and convolution neural network |
| CN111223074A (en) * | 2019-12-27 | 2020-06-02 | 宁夏建筑科学研究院股份有限公司 | Method for detecting concrete defects at top of concrete column member |
| CN112907460A (en) * | 2021-01-25 | 2021-06-04 | 宁波市鄞州区测绘院 | Remote sensing image enhancement method |
| CN113077486A (en) * | 2021-04-30 | 2021-07-06 | 深圳世源工程技术有限公司 | Method and system for monitoring vegetation coverage rate in mountainous area |
| CN113191985A (en) * | 2021-05-24 | 2021-07-30 | 北京美医医学技术研究院有限公司 | Intelligent freckle removing system based on infrared laser |
| CN113361532A (en) * | 2021-03-10 | 2021-09-07 | 江西理工大学 | Image identification method, system, storage medium, equipment, terminal and application |
| CN113450340A (en) * | 2021-07-13 | 2021-09-28 | 北京美医医学技术研究院有限公司 | Skin texture detecting system |
| CN114170674A (en) * | 2021-11-02 | 2022-03-11 | 天津工业大学 | Method for distinguishing personnel state in man-machine cooperation scene |
| CN114581374A (en) * | 2022-01-25 | 2022-06-03 | 深圳飞亮智能科技有限公司 | Electric vehicle driving assisting system based on artificial intelligence technology |
-
2022
- 2022-06-15 CN CN202210675220.9A patent/CN115272174B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105404847A (en) * | 2014-09-16 | 2016-03-16 | 北京计算机技术及应用研究所 | Real-time detection method for object left behind |
| CN106446914A (en) * | 2016-09-28 | 2017-02-22 | 天津工业大学 | Road detection based on superpixels and convolution neural network |
| CN111223074A (en) * | 2019-12-27 | 2020-06-02 | 宁夏建筑科学研究院股份有限公司 | Method for detecting concrete defects at top of concrete column member |
| CN112907460A (en) * | 2021-01-25 | 2021-06-04 | 宁波市鄞州区测绘院 | Remote sensing image enhancement method |
| CN113361532A (en) * | 2021-03-10 | 2021-09-07 | 江西理工大学 | Image identification method, system, storage medium, equipment, terminal and application |
| CN113077486A (en) * | 2021-04-30 | 2021-07-06 | 深圳世源工程技术有限公司 | Method and system for monitoring vegetation coverage rate in mountainous area |
| CN113191985A (en) * | 2021-05-24 | 2021-07-30 | 北京美医医学技术研究院有限公司 | Intelligent freckle removing system based on infrared laser |
| CN113450340A (en) * | 2021-07-13 | 2021-09-28 | 北京美医医学技术研究院有限公司 | Skin texture detecting system |
| CN114170674A (en) * | 2021-11-02 | 2022-03-11 | 天津工业大学 | Method for distinguishing personnel state in man-machine cooperation scene |
| CN114581374A (en) * | 2022-01-25 | 2022-06-03 | 深圳飞亮智能科技有限公司 | Electric vehicle driving assisting system based on artificial intelligence technology |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115797473A (en) * | 2023-02-09 | 2023-03-14 | 国网山东省电力公司东营供电公司 | Concrete forming evaluation method for civil engineering |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115272174B (en) | 2023-05-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105719259B (en) | Pavement crack image detecting method | |
| CN110889812B (en) | Underwater image enhancement method for multi-scale fusion of image characteristic information | |
| CN116703910B (en) | Intelligent detection method for quality of concrete prefabricated bottom plate | |
| CN112561804A (en) | Low-illumination underwater image enhancement method based on multi-scale detail enhancement | |
| WO2021114564A1 (en) | Enhancement method for low-contrast infrared image | |
| CN111210395B (en) | Retinex underwater image enhancement method based on gray value mapping | |
| JPH05225332A (en) | Method and apparatus for variable-space wave filtering | |
| CN115272174A (en) | Municipal road detection method and system | |
| CN110675317B (en) | Super-resolution reconstruction method based on learning and adaptive trilateral filtering regularization | |
| CN107038444A (en) | A kind of image-recognizing method of pointer dial plate | |
| CN100367770C (en) | Method for removing isolated noise point in video | |
| CN114898210B (en) | Neural network-based remote sensing image target identification method | |
| CN116703784B (en) | Heart ultrasonic image vision enhancement method | |
| KR20140109801A (en) | Method and apparatus for enhancing quality of 3D image | |
| CN108898561A (en) | A kind of defogging method, server and the system of the Misty Image containing sky areas | |
| CN110930330B (en) | Image segmentation and region growth based salt and pepper noise reduction algorithm | |
| CN113269698B (en) | Low-exposure vein image enhancement method based on Actor-Critic model | |
| CN116823686A (en) | Night infrared and visible light image fusion method based on image enhancement | |
| Peris-Fajarnés et al. | Rank-ordered differences statistic based switching vector filter | |
| CN110852977B (en) | Image enhancement method for fusing edge gray level histogram and human eye visual perception characteristics | |
| CN111882529A (en) | Display screen Mura defect detection method and device based on Gaussian-Surface edge preserving filtering | |
| CN110647843B (en) | Face image processing method | |
| CN109712099B (en) | Sonar image equalization method based on SLIC and adaptive filtering | |
| Tangsakul et al. | Improving single image haze removal based on cellular automata model | |
| Brad | Satellite image enhancement by controlled statistical differentiation |
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 |