A water conservancy project leakage detection system

Abstract

The present application belongs to the technical field of water conservancy project leakage detection, and provides a water conservancy project leakage detection system. The present application firstly obtains a target image through an image acquisition module; then divides a window of the target image through an image enhancement module, and obtains a suspected tracer area in the window under contrast limitation, and analyzes the gray scale distribution characteristics and the gray scale change characteristics of the pixel points in the suspected tracer area to obtain the definition of the window, complete the enhancement of the target image, make the edge and characteristics of the tracer more prominent, so that it is easier to identify whether the tracer has been inhaled; finally, through a leakage detection module, whether the vertical seam metal cover sheet in the water conservancy project has failure or leakage phenomenon is judged according to the enhanced target image. The present application greatly improves the accuracy of leakage detection, thereby ensuring the safety and reliability of the water conservancy project.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of water conservancy engineering leakage detection, and in particular to a water conservancy engineering leakage detection system. BACKGROUND

[0002] The water conservancy engineering leakage detection system is an important technical means to ensure water resource safety and engineering structural integrity. With the continuous expansion of the scale and increasing complexity of water conservancy projects, leakage problems are becoming increasingly prominent, which may lead to water resource waste, environmental pollution, and engineering safety hazards. The leakage detection system can timely find the leakage position and degree by real-time monitoring and analysis of water flow dynamics, thereby providing a scientific basis for engineering maintenance and management. In addition, early leakage detection can effectively reduce maintenance costs, prolong the service life of the project, and improve overall management efficiency. Through real-time data feedback, relevant departments can quickly respond and take appropriate measures to ensure the sustainability and safety of water conservancy projects, thereby protecting water resources and maintaining the ecological environment. In summary, the water conservancy engineering leakage detection system plays a crucial role in ensuring the safety of water conservancy facilities, improving operational efficiency, and protecting the environment.

[0003] Safety inspection of reservoir dams is a key link to ensure the stability of water conservancy projects and prevent accidents. With the continuous increase in the size of dams and the extension of service life, traditional manual inspection methods face many challenges, such as limited inspection range, complex working environment, and high safety risks. Therefore, underwater robots (ROVs) have emerged as an important tool for modern reservoir dam safety inspection. Underwater robots (ROVs) can efficiently and accurately perform inspection tasks in complex underwater environments. They are equipped with high-definition cameras and sensors that can monitor the structural state of the dam, detect leaks, and assess sediment and corrosion. This technology not only improves the efficiency and accuracy of inspection, but also greatly reduces the risks and costs associated with manual inspection. At the same time, underwater robots (ROVs) can operate in low-visibility and deep-water areas, ensuring a comprehensive assessment of the dam.

[0004] In the prior art, when detecting the leakage of the vertical joint metal cover plate of the reservoir dam, underwater robots (ROVs) are often used for detection. However, due to the high content of underwater silt, the detected part is often blurred, even with the tracer mark, it is still difficult to determine whether the tracer is inhaled, which makes it difficult to determine whether the vertical joint metal cover plate has defects, thereby causing errors in the leakage detection of water conservancy projects. SUMMARY

[0005] To solve the above technical problems, the present application provides a water conservancy engineering leakage detection system.

[0006] According to the water conservancy engineering leakage detection system provided by the present application, the system comprises:

[0007] an image acquisition module, configured to spray a tracer on a target region of a reservoir dam and capture a target image;

[0008] an image enhancement module, configured to divide the target image into windows, acquire suspected tracer regions in the windows under contrast limitation, analyze gray scale distribution features and gray scale change features of pixel points in the suspected tracer regions, obtain a definition of the window, and complete enhancement of the target image;

[0009] a leakage detection module, configured to determine whether a vertical seam metal cover plate in a water conservancy project is invalid or has a leakage phenomenon according to the enhanced target image.

[0010] In some embodiments of the present application, the image enhancement module comprises a window division unit, a gray scale distribution feature analysis unit, a gray scale change feature analysis unit, a definition calculation unit and an image enhancement unit; wherein:

[0011] the window division unit is configured to divide the target image into windows and acquire suspected tracer regions in the windows under contrast limitation;

[0012] the gray scale distribution feature analysis unit is configured to analyze gray scale distribution features of pixel points in the suspected tracer regions;

[0013] the gray scale change feature analysis unit is configured to analyze gray scale change features of pixel points in the suspected tracer regions;

[0014] a definition analysis unit is configured to obtain a defect information performance degree of the suspected tracer regions according to the gray scale distribution features and the gray scale change features, and record a maximum value of the defect information performance degrees of all suspected tracer regions in the window as a definition of the window;

[0015] an image enhancement unit is configured to complete enhancement of the target image according to the definition.

[0016] In some embodiments of the present application, the window division unit is configured to:

[0017] divide the target image into grid-shaped windows, each window has a size of N*N, 2N*2N is a basic unit window for histogram equalization, and a moving distance is N each time;

[0018] for each 2N*2N basic unit window, Otsu threshold segmentation is performed under contrast limitation, a connected domain is extracted, and a suspected tracer region in the basic unit window is obtained.

[0019] In some embodiments of the present application, the gray scale distribution feature analysis unit is configured to:

[0020] cluster the suspected tracer region to obtain a far-from-fracture region and a near-from-fracture region;

[0021] analyze the width difference between the far-from-fracture region and the near-from-fracture region to obtain the gray scale distribution feature of the pixel points in the suspected tracer region.

[0022] In some embodiments of the present application, clustering the suspected tracer region to obtain a far-from-fracture region and a near-from-fracture region comprises:

[0023] perform K-means clustering on the suspected tracer region to divide the suspected tracer region into three regions;

[0024] extract the largest area region in each region as a reference region;

[0025] calculate the first variance of the gray scale values of the pixel points in each reference region;

[0026] calculate the average gray scale values of the pixel points in the two reference regions with smaller first variances, define the reference region corresponding to the larger average gray scale value as the far-from-fracture region, and define the reference region corresponding to the smaller average gray scale value as the near-from-fracture region.

[0027] In some embodiments of the present application, analyzing the width difference between the far-from-fracture region and the near-from-fracture region comprises:

[0028] determine the center points of the far-from-fracture region and the near-from-fracture region, connect the two center points to obtain a first line segment;

[0029] establish a rectangular coordinate system with the center point of the near-from-fracture region as the origin and the first line segment as the y-axis;

[0030] In the rectangular coordinate system, the difference between the maximum coordinate on the x-axis of the far-from-fracture region and the minimum coordinate on the x-axis of the far-from-fracture region is denoted as the width of the far-from-fracture region, and the difference between the maximum coordinate on the x-axis of the near-from-fracture region and the minimum coordinate on the x-axis of the near-from-fracture region is denoted as the width of the near-from-fracture region.

[0031] calculate the difference between the width of the far-from-fracture region and the width of the near-from-fracture region to obtain the width difference between the far-from-fracture region and the near-from-fracture region.

[0032] In some embodiments of the present application, the gray scale change feature analysis unit is configured to:

[0033] The degree of gradient decrease of the pixel gray value between the pixel points in the far-from-fissure region and the center point of the near-fissure region is analyzed to obtain the gray variation feature of the pixel points in the suspected tracer region.

[0034] In some embodiments of the present application, the analysis of the degree of gradient decrease of the pixel gray value between the pixel points in the far-from-fissure region and the center point of the near-fissure region comprises:

[0035] Each pixel point in the far-from-fissure region is connected with the center point of the near-fissure region to obtain a plurality of second line segments;

[0036] The pixel sequence of the second line segment is extracted;

[0037] The first derivative of the pixel gray value of the pixel sequence is calculated, and the number of negative values of the first derivative and the second variance of the first derivative are counted;

[0038] The degree of gradient decrease of the pixel gray value of the pixel sequence is obtained according to the number of negative values of the first derivative and the second variance.

[0039] In some embodiments of the present application, the image enhancement module further comprises an interference removal unit for analyzing the edge variation feature of the target image, identifying and removing the non-real information region of the target image.

[0040] In some embodiments of the present application, the analysis of the edge variation feature of the target image and the identification and removal of the non-real information region of the target image comprise:

[0041] The target image is subjected to K-Means clustering to obtain a plurality of suspected regions;

[0042] The edge chain code of the suspected region is obtained, and the first difference chain code of the edge chain code is calculated;

[0043] The third variance of the pixel gray value in the suspected region is calculated;

[0044] The probability that the suspected region is a non-real information region is obtained according to the first difference chain code and the third variance;

[0045] A preset probability threshold is set, and the suspected region with a probability greater than the probability threshold is recorded as a non-real information region, and the non-real information region is excluded.

[0046] As can be seen from the above embodiments, the water conservancy project leakage detection system provided by the embodiments of the present application has the following beneficial effects:

[0047] The application firstly obtains a target image through an image acquisition module; then divides a window of the target image through an image enhancement module, and obtains a suspected tracer area in the window under contrast limitation, and analyzes the gray scale distribution characteristics and the gray scale change characteristics of the pixel points in the suspected tracer area to obtain the definition of the window, completes the enhancement of the target image, makes the edge and characteristics of the tracer more prominent, so that it is easier to identify whether the tracer appears inhaled condition; finally, a leakage detection module judges whether the vertical joint metal cover plate in the water conservancy project exists failure or leakage phenomenon according to the enhanced target image. The system greatly improves the accuracy of leakage detection, thereby ensuring the safety and reliability of the water conservancy project. In addition, the system can effectively reduce the amount of calculation and improve the detection efficiency by setting an interference removal unit to exclude the interference of non-real information areas such as non-illumination areas and light column areas.

[0048] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, and cannot limit the application. BRIEF DESCRIPTION OF DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the application or the prior art and the advantages thereof, the drawings needed to be used in the embodiments or the prior art description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can be obtained without creative labor based on these drawings.

[0050] Figure 1 A basic component schematic diagram of a water conservancy project leakage detection system provided by the embodiment of the application;

[0051] Figure 2 A dam vertical joint leakage schematic diagram provided by the embodiment of the application;

[0052] Figure 3 An enhanced target image schematic diagram provided by the embodiment of the application.

[0053] Serial number Description: 10-Image acquisition module, 20-Image enhancement module, 30-Leakage detection module, 21-Interference removal unit, 22-Window division unit, 23-Gray scale distribution characteristic analysis unit, 24-Gray scale change characteristic analysis unit, 25-Definition calculation unit, 26-Image enhancement unit. DETAILED DESCRIPTION

[0054] In order to further illustrate the technical means and effects taken by the present application to achieve the predetermined inventive objectives, the specific implementation, structure, features and effects of a water conservancy project leakage detection system according to the present application are described in detail below in combination with the drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments can be combined in any suitable form.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The use of terms such as "comprising", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that the circuit structure, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed or inherent to such article or device. Without more limitations, the element defined by the phrase "including one" does not exclude the presence of additional identical elements in the article or device including the element. The relational terms "first" and "second" and the like are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between the entities or operations.

[0056] The scenario targeted by the present application is mainly applied to leakage detection of water conservancy projects such as reservoir dams, especially in environments with high silt content. Due to the high silt content under water, the detected part often appears blurred, even under the mark of the tracer, it is still difficult to determine whether the tracer is inhaled, and further to determine whether the metal cover plate of the vertical joint has defects, and further to cause errors in the leakage detection of water conservancy projects. Therefore, the present application aims to solve the problems existing in the prior art of water conservancy project leakage detection, especially the problem of image blurring and inaccurate leakage judgment caused by underwater silt. Through image enhancement and leakage feature judgment, the inhalation feature of the tracer in the image is more obvious, and the accuracy of defect judgment is enhanced, so as to better serve the maintenance and management of water conservancy projects, and ensure the safety and stability of key facilities such as reservoir dams.

[0057] The water conservancy project leakage detection system provided by the present embodiment will be described in detail below in combination with the drawings.

[0058] Please refer to Figure 1 which shows the basic composition of a water conservancy project leakage detection system provided by one embodiment of the present application.

[0059] As Figure 1As shown, one embodiment of the present application provides a water conservancy project leakage detection system, which specifically comprises: an image acquisition module 10, an image enhancement module 20, and a leakage detection module 30. Specifically,

[0060] The image acquisition module 10 is configured to spray tracer on the target area of the reservoir dam and take target images.

[0061] The dam vertical joint is a vertical construction joint set to adapt the face slab of the face slab rock-fill dam to the deformation of the rock-fill body and prevent cracks or damage due to temperature changes (thermal expansion and cold contraction). The dam vertical joint metal cover is usually a metal sheet used to seal the vertical joint to prevent water infiltration. When the metal cover fails, it may cause vertical joint leakage, such as Figure 2 As shown. For detection of vertical joint leakage, a mounted inkjet system is often used for verification. When the tracer is sucked along the direction of water flow, it proves that the vertical joint metal cover has failed.

[0062] Therefore, first, the image acquisition module 10 is configured to spray tracer on the target area of the reservoir dam and take target images. Specifically, the image acquisition module 10 is configured to arrange an underwater robot (ROV) carrying an inkjet system, a high-definition camera, and a sensor, spray tracer on the target area of the reservoir dam, take target images, and perform preprocessing operations such as cropping on the target images. Then the image acquisition module 10 is further configured to: the image acquisition module 10 is electrically connected with the image enhancement module 20, the image acquisition module 10 transmits the preprocessed target images to the image enhancement module 20, and the image enhancement module 20 is convenient for further analyzing the target images.

[0063] The image enhancement module 20 is configured to divide the target image into windows, obtain the suspected tracer area within the window under the contrast limit, and then analyze the gray scale distribution characteristics and gray scale change characteristics of the pixel points in the suspected tracer area to obtain the definition of the window, and complete the enhancement of the target image. Further, the image enhancement module 20 includes an interference removal unit 21, a window division unit 22, a gray scale distribution characteristic analysis unit 23, a gray scale change characteristic analysis unit 24, a definition calculation unit 25, and an image enhancement unit 26.

[0064] In the target image, the non-real information region is characterized by regular edge change, and the gray value of the pixel points in the region changes little. For example, the non-illumination region has a low overall gray value and the gray value changes little, and the overall region edge presents an arc shape; the light column region is caused by the scattering of light due to the ROV lighting device encountering suspended particles (such as silt, algae, etc.) in water, and is characterized by a high overall gray value and little change in the gray value, and the region edge is long strip-shaped. According to the above characteristics, the non-real information region can be excluded, and the calculation amount can be greatly reduced. Therefore, the interference removal unit 21 is arranged to analyze the edge change characteristics of the target image, and identify and remove the non-real information region of the target image. The specific implementation is as follows:

[0065] Firstly, the target image is subjected to K-Means clustering to obtain a plurality of suspected regions. In the embodiment of the present application, K=4, that is, four suspected regions can be obtained.

[0066] Then, the edge chain code of the suspected region is obtained, and the edge chain code of any suspected region j is denoted as a1, a2, …, a m m represents the total number of edge chain codes of the suspected region j, and the first-order difference chain code of the edge chain code is calculated. The first-order difference chain code corresponding to any suspected region j is |a1-a2|, |a2-a3|, …, |a m-1 -a m |; the third variance of the pixel gray value in the suspected region is calculated, and the third variance of the pixel gray value in any suspected region j is denoted as S j .

[0067] Then, according to the first-order difference chain code and the third variance, the probability that the suspected region is a non-real information region is obtained. The probability calculation formula of the suspected region j being a non-real information region is constructed as follows:

[0068]

[0069] In the formula, ZSXX j represents the probability that the suspected region j belongs to the non-real information region; a u represents the u-th edge chain code of the suspected region j; a u+1 represents the u+1-th edge chain code of the suspected region j; S j represents the third variance of the pixel gray value in the suspected region j; m represents the total number of edge chain codes of the suspected region j; and exp represents the exponential function with the natural constant e as the base.

[0070] represents the cumulative first-order difference chain code of the edge chain code of the suspected region j. The smaller the cumulative value is, the more regular the edge of the suspected region j is, and the less likely it contains real defect information; S jThe third variance represents the pixel gray value in the suspected area j. The smaller the variance, the smaller the change of the pixel gray value in the suspected area j, the less the real defect information contained, and the greater the probability that the suspected area j belongs to the non-real information area.

[0071] Finally, a preset probability threshold is set, and the suspected area corresponding to the probability greater than the probability threshold is recorded as a non-real information area, and the non-real information area is excluded. In the embodiment of the application, the probability threshold KU=0.95, that is, the suspected area ZSXX j The suspected area greater than 0.95 is marked as a non-real information area, and is excluded from the subsequent calculation process.

[0072] Since the target image is taken underwater, the tracer absorption around the vertical seam metal cover plate in the target image may be blurred, which affects the leakage quality detection. Therefore, it is necessary to enhance the tracer absorption feature around the vertical seam metal cover plate. In order to enhance the target area, the window needs to be divided first, the image in the window is histogramed, different contrast parameters are set, the clarity of the target area is calculated, and the enhancement result with the highest clarity is taken as the image enhancement result of the window.

[0073] For the calculation of the clarity in the window, when the image in the window is under the support of the tracer (such as ink), the higher the clarity in the window, the more the image in the window reflects the defect information of the metal cover plate, that is, the image in the window shows the absorption of the tracer, and the higher the clarity in the window. When the vertical seam metal cover plate fails, the tracer may be absorbed into the interior along the water flow direction. At this time, the tracer does not diffuse to the surrounding, and the closer to the defect absorption port of the metal cover plate, the greater the concentration of the tracer, and the smaller the gray value of the pixel in the image, and the shape of the tracer is more consistent with the defect area (such as crack, hole, etc.) of the metal cover plate. As for the tracer away from the defect area of the metal cover plate, and the tracer without defect area in the image, the whole presents a state of diffusion to the surrounding (i.e. diffusion state), which is reflected in the gray value that the gradient change direction of the gray value is more chaotic, and the shape of the tracer does not fit the defect area of the metal cover plate. Accordingly, the clarity of the window under the parameter of a certain contrast limit is calculated.

[0074] First, a window partitioning unit 22 is set up to partition the target image into windows and obtain the suspected tracer region within the window under contrast constraints. Specifically, the window partitioning unit 22 is configured to: divide the target image into grid-like windows, each window being N×N in size, with 2N×2N as the basic unit window for histogram equalization, and moving a distance of N each time; for each 2N×2N basic unit window, under contrast constraints (in this embodiment, the contrast constraint is [2,20]), when the contrast constraint parameter is n, n∈[2,20], the image in the basic unit window t is subjected to Otsu threshold segmentation to extract the connected components, denoted as T={t1,t2,…,t d ,…,t r The resulting connected components are labeled as suspected tracer (ink) regions, meaning multiple suspected tracer regions are obtained within a basic unit window, where T represents the set of connected components (suspected tracer regions) in the basic unit window t, and t d denoted as d-th connected component (suspected tracer region) in basic unit window t, and r represents the total number of connected components (suspected tracer regions) in basic unit window t.

[0075] The grayscale distribution feature analysis unit 23 is used to analyze the grayscale distribution features of pixels within the suspected tracer region. Further, the grayscale distribution feature analysis unit 23 is configured as follows:

[0076] The suspected tracer regions are clustered to identify regions far from and near the cracks. Specifically, firstly, for any suspected tracer region t within the basic unit window t... d K-means clustering is performed, with K=3 in this embodiment of the invention, to cluster the suspected tracer regions t d The regions are divided into three categories. Since the suspected tracer regions are obtained through Otsu thresholding, this segmentation method only divides the image into foreground and background (relatively coarse). When further performing K-means clustering on the suspected tracer regions, more refined connected components are extracted, and then clustering is completed. Therefore, the three categories of regions obtained by K-means clustering are essentially regions containing multiple small connected components. Therefore, the region with the largest area in each category is further extracted as the reference region, denoted as β1, β2, and β3 respectively. Then, the first variance of the pixel grayscale values ​​within these three reference regions is calculated, and the regions are sorted from smallest to largest according to the first variance, denoted as S. β,1 ,S β,2 ,S β,3 S β,1 ,S β,2 ,S β,3respectively represent the minimum value, the second minimum value and the maximum value of the first variance of the pixel gray value in the reference region; finally, since when the tracer shows the inhalation feature, the entire suspected tracer region is divided into three types of regions, the two reference regions with smaller variance values may be the two ends of the tracer, and the middle reference region is the inhalation stage from light to deep, so the variance value S β,1 , β,2 The corresponding two reference regions are further distinguished, and in the embodiment of the present application, the average gray value of the pixels in the two reference regions with smaller first variance (S β,1 , β,2 ) is calculated, the reference region corresponding to the larger average gray value is defined as the region far from the crack, denoted as β yllf , and the reference region corresponding to the smaller average gray value is defined as the region close to the crack, denoted as β kjlf .

[0077] After obtaining the region far from the crack and the region close to the crack, the gray distribution characteristics of the pixels in the suspected tracer region are obtained by analyzing the width difference between the region far from the crack and the region close to the crack. Specifically, first, the center point of the region far from the crack β yllf and the center point of the region close to the crack β kjlf , i.e. the center point of the circumscribed rectangle of the region, are determined, and the first line segment is obtained by connecting the two center points; then, the center point of the region close to the crack β kjlf is taken as the origin, and the first line segment is taken as the y-axis to establish a rectangular coordinate system; then, in the rectangular coordinate system, the widths of the region far from the crack β yllf and the region close to the crack β kjlf , i.e. the difference between the maximum coordinate on the x-axis and the minimum coordinate on the x-axis of the region far from the crack β yllf and the region close to the crack β kjlf , are calculated, and are respectively denoted as the width KD yllf of the region far from the crack β kjlf and the width KD yllf of the region close to the crack β kjlf ; the greater the difference between KD yllf and KD kjlf , the greater the possibility that the tracer is being inhaled into the crack, the more likely the suspected tracer region contains tracer inhalation information, and the clearer the image, so finally, the difference between the width of the region far from the crack and the width of the region close to the crack is calculated to obtain the width difference between the region far from the crack and the region close to the crack.

[0078] The gray scale change feature analysis unit 24 is configured to analyze the gray scale change feature of the pixels in the suspected tracer region. Further, the gray scale change feature analysis unit 24 is configured to analyze the degree of gradient decline of the gray scale values of the pixels between the pixels away from the crack region and the center point close to the crack region, to obtain the gray scale change feature of the pixels in the suspected tracer region. Specifically, first, connect each pixel point away from the crack region β yllf to the center point close to the crack region β kjlf to obtain a plurality of second line segments; then, extract the pixel sequence of the second line segment, denoted as xsxl1,xsxl2,…,xsxl a ,…,xsxl b , a<b, b represents the total number of second line segments, that is, the total number of pixel sequences in the suspected tracer region t d . The more the gray scale values of these pixel sequences satisfy the gradient decline, the more likely the suspected tracer region contains tracer inhalation information, and the clearer the image is. Therefore, finally, the gray scale values of any pixel sequence xsxl a are denoted as i a,1 ,i a,2 ,…,i a,e , e represents the total number of pixel points of the pixel sequence, the first derivative of the gray scale values of any pixel sequence xsxl a is calculated and denoted as i'1,i'2,…,i' e-1 , the number of negative values of the first derivative is counted as num, and the second variance of the first derivative is calculated as S i ; then, according to the number of negative values of the first derivative and the second variance, the degree of gradient decline of the gray scale values of the pixel sequence xsxl a is obtained; the degree of gradient decline calculation formula of the gray scale values of the pixel sequence xsxl a is constructed as follows:

[0079]

[0080] In the formula, TDXJ a represents the degree of gradient decline of the gray scale values of the pixel sequence xsxl a ; num represents the number of negative values of the first derivative of the gray scale values of the pixel sequence xsxl a ; e-1 represents the total number of first derivatives of the gray scale values of the pixel sequence xsxl a ; S i represents the second variance of the first derivative of the gray scale values of the pixel sequence xsxl a ; norm represents a linear normalization function; and exp represents an exponential function with natural constant e as the base.

[0081] represents the pixel sequence xsxla The greater the proportion, the greater the decline in the gray value of the pixel sequence xsxl a S i represents the second variance of the first derivative of the gray value of the pixel sequence xsxl a The smaller the variance, the more uniform the change in the gray value of the pixel sequence xsxl a The greater the decline in the gray value of the pixel sequence xsxl a .

[0082] Similarly, the decline in the gray value of all pixel sequences xsxl1,xsxl2,…,xsxl a ,…,xsxl b is obtained, and is denoted as TDXJ1,TDXJ2,…,TDXJ a ,…,TDXJ b , respectively.

[0083] The sharpness analysis unit is configured to obtain the defect information performance degree of the suspected tracer region according to the gray distribution feature and the gray change feature, and the maximum value of the defect information performance degree corresponding to all suspected tracer regions in the window is denoted as the sharpness of the window. Specifically, the sharpness analysis unit is configured to:

[0084] obtain the defect information performance degree of the suspected tracer region according to the gray distribution feature and the gray change feature based on the preset defect information performance degree calculation formula, wherein the preset defect information performance degree calculation formula is:

[0085]

[0086] In the formula, QX d represents the defect information performance degree of the suspected tracer region t d ; TDXJ a represents the decline in the gray value of the pixel sequence xsxl a ; b represents the total number of pixel sequences in the suspected tracer region t d ; KD yllf ,KD kjlf respectively represent the width of the region β d far from the crack and the region β yllf close to the crack in the suspected tracer region t kjlf ; and norm represents a linear normalization function.

[0087] represents the distance between each pixel point in the region β yllf far from the crack and the region β kjlfThe average of the gradient descent degree of the gray value of all pixel sequences of the center point connecting line of the suspected tracer region t d The greater the average, the more the suspected tracer region t d satisfies the feature of the tracer in the inhalation process that the gray value gradient decreases. yllf The greater the average, the more the suspected tracer region t kjlf represents the distance from the crack region β yllf and the distance from the crack region β kjlf The greater the difference, the more the suspected tracer region t d satisfies the feature of the tracer in the inhalation process that the tracer width decreases.

[0088] The defect information performance of each suspected tracer region in the basic unit window t is calculated, and is denoted as QX1, QX2, …, QX d , …, QX r The maximum value among them is taken as the definition of the basic unit window t in the basic unit window, and the definition of the image is denoted as QXD t,n .

[0089] 2N×2N is taken as the basic unit window of histogram equalization, and the moving distance is N each time. All windows in the target image are traversed to obtain the definition of the image corresponding to each basic unit window. For the definition of the overlapping region, the average of the definitions of multiple basic unit windows is taken as the final definition.

[0090] The image enhancement unit 26 is configured to complete the enhancement of the target image according to the definition. Specifically, the image enhancement unit 26 is configured to perform image enhancement on the image in each window according to the definition obtained by each window, so as to realize the enhancement of the entire target image, that is, to take the enhancement result with the highest definition as the image enhancement result of the window, to obtain the enhancement result of the entire target image, and the enhanced target image is as shown in Figure 3 .

[0091] The image enhancement module 20 is electrically connected with the leakage detection module 30, and the image enhancement module 20 can transmit the enhanced target image to the leakage detection module 30. The leakage detection module 30 is used for judging whether the vertical joint metal cover plate in the water conservancy project exists failure or leakage phenomenon according to the enhanced target image, and can send the judgment result to the user end for early warning. Through the image enhancement module 20, the enhancement of the tracer form in the leakage detection of the water conservancy project is realized, the edges and characteristics of the tracer are more prominent in the enhanced target image, so that whether the tracer appears the inhalation condition is more easily identified. Through this clear visual presentation, whether the vertical joint metal cover plate exists failure or leakage phenomenon can be accurately judged, the accuracy of the judgment is greatly improved, engineers can identify potential problems in time, and take necessary maintenance measures, so as to guarantee the safety and reliability of the water conservancy project.

[0092] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description, and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0093] Each of the embodiments in the specification is described in a progressive manner, and the same and similar parts between the embodiments can be referred to each other. Each embodiment mainly describes the difference from other embodiments.

Claims

1. A hydraulic works leakage detection system, characterised in that, The system comprises: An image acquisition module for spraying a tracer on a target area of a reservoir dam and taking a target image; An image enhancement module for dividing a window on the target image, obtaining a suspected tracer area in the window under contrast limitation, analyzing gray scale distribution characteristics and gray scale change characteristics of pixel points in the suspected tracer area, obtaining a definition of the window, and completing enhancement of the target image; A leakage detection module for judging whether a vertical seam metal cover plate in a water conservancy project has failure or leakage phenomenon according to the enhanced target image; The image enhancement module comprises a window division unit, a gray scale distribution characteristic analysis unit, a gray scale change characteristic analysis unit, a definition calculation unit, and an image enhancement unit; wherein: The window division unit is configured to divide a window on the target image and obtain a suspected tracer area in the window under contrast limitation; The gray scale distribution characteristic analysis unit is configured to analyze gray scale distribution characteristics of pixel points in the suspected tracer area; The gray scale change characteristic analysis unit is configured to analyze gray scale change characteristics of pixel points in the suspected tracer area; The definition calculation unit is configured to obtain a defect information performance degree of the suspected tracer area according to the gray scale distribution characteristics and the gray scale change characteristics, and record a maximum value of the defect information performance degree corresponding to all suspected tracer areas in the window as a definition of the window; The image enhancement unit is configured to complete enhancement of the target image according to the definition; The gray scale distribution characteristic analysis unit is configured to: cluster the suspected tracer area to obtain a far-from-crack area and a near-crack area, including: perform K-means clustering on the suspected tracer area to divide the suspected tracer area into three areas; extract a largest-area area in each area as a reference area; calculate a first variance of pixel gray scale values in each reference area; calculate average gray scale values of pixel points in two reference areas with smaller first variances, define the reference area corresponding to a larger average gray scale value as the far-from-crack area, and define the reference area corresponding to a smaller average gray scale value as the near-crack area; analyze a width difference between the far-from-crack area and the near-crack area to obtain gray scale distribution characteristics of pixel points in the suspected tracer area, including: determine a center point of the far-from-crack area and a center point of the near-crack area, connect the two center points to obtain a first line segment; The center point of the region close to the crack is taken as the origin, and the first line segment is taken as the X-axis An orthogonal coordinate system is established with the shaft as the Z-axis In the Cartesian coordinate system, the region far from the crack and the region near the crack are... The maximum coordinate of the axis minus The difference between the minimum coordinates of the axes is denoted as the width of the region far from the crack and the width of the region near the crack, respectively. calculate a difference between a width of the far-from-crack area and a width of the near-crack area to obtain the width difference between the far-from-crack area and the near-crack area; The gray scale change characteristic analysis unit is configured to: analyze a degree of gradient decline of pixel gray scale values between pixel points in the far-from-crack area and a center point of the near-crack area to obtain gray scale change characteristics of pixel points in the suspected tracer area, including: connect each pixel point in the far-from-crack area to the center point of the near-crack area to obtain a plurality of second line segments; extracting a pixel sequence of the second line segment; calculating a first derivative of the gray value of the pixel sequence, and counting a number of negative values of the first derivative and calculating a second variance of the first derivative; obtaining a gradient descent degree of the gray value of the pixel sequence according to the number of negative values of the first derivative and the second variance.

2. The hydraulic works leak detection system of claim 1, wherein, The window division unit is configured to: for dividing the target image into grid-shaped windows, each of which has a size of to a basic unit window for histogram equalization, each time moving a distance of ; For each In the basic unit window, under the contrast restriction, Otsu threshold segmentation is performed, and a connected domain is extracted to obtain a suspected tracer region in the basic unit window.

3. The hydraulic works leak detection system of claim 1, wherein, The image enhancement module further comprises an interference removal unit configured to analyze edge change characteristics of the target image, and identify and remove a non-real information region of the target image.

4. The hydraulic works leak detection system of claim 3, wherein, The analysis of the edge change characteristics of the target image, and the identification and removal of the non-real information region of the target image comprise: performing K-Means clustering on the target image to obtain a plurality of suspected regions; obtaining an edge chain code of the suspected region, and calculating a first difference chain code of the edge chain code; calculating a third variance of the gray value of the pixel points in the suspected region; obtaining a probability that the suspected region is a non-real information region according to the first difference chain code and the third variance; a preset probability threshold, and recording the suspected region corresponding to the probability greater than the probability threshold as a non-real information region, and excluding the non-real information region.

Images