Underwater concrete structure defect degradation image restoration and quality improvement method

Abstract

The invention discloses an underwater concrete structure defect degradation image restoration and quality improvement method, which comprises the following steps of: firstly, accurately estimating background light in an underwater complex environment by fusing a quadtree hierarchical search strategy and a multi-feature prior index; then, based on a transmission graph estimation method of an improved dark channel prior theory and a reverse saturation graph theory, the problem of transmission graph errors caused by uneven illumination and artificial light source interference is solved; furthermore, an image restoration and enhancement technology is combined, and an image degradation phenomenon caused by color distortion, contrast reduction and weak light turbidity is effectively restored. According to the method, restoration and quality improvement of the underwater concrete structure defect degradation image in the scenes of artificial light source irradiation, weak light turbidity and the like can be realized, the underwater image quality is effectively improved, and high-quality image support is provided for underwater structure defect identification and detection.

Description

Technical Field

[0001] The present invention relates to a method for restoring and improving the quality of images of underwater concrete structure defects, belonging to the technical field of underwater image restoration. Background Art

[0002] During the long-term service of water-related buildings, under the long-term coupling action of water pressure, temperature load, erosion and penetration, various defects such as cracks, holes, exposed aggregates and depressions will inevitably appear in the underwater concrete structure. Defects often form on the surface and are prone to extend inward under the action of factors such as hydraulic fracturing, weakening the overall stiffness and bearing capacity of the structure, and even causing structural instability. After diseases such as underwater structure cracking, leakage, damage and collapse occur, water-related buildings generally drain the water first and then conduct manual inspection and repair and reinforcement.

[0003] There are many drawbacks in the manual underwater inspection method, including shallow diving depth (usually less than 60m), long time consumption, difficult to achieve full coverage of deep water areas, and high operation risk and labor cost. The recognition results of manual inspection means are easily affected by the subjective judgment and practical operation ability of inspectors, and it is difficult to meet the defect detection requirements of deep water parts or large-area regions of hydraulic structures. The structural safety monitoring system can often sense the change of the effect quantity related to structural damage by burying sensors inside the structure. However, due to the covering and blocking effect of water, when the system senses abnormal signals such as a sudden increase in leakage volume or excessive uplift pressure, the defect often has penetrated deep into the structure, forming a leakage channel that may penetrate the dam body, and the difficulty of risk removal and reinforcement increases suddenly. However, due to the coupling action of multiple factors such as environment and load, the defects of underwater concrete structures are widely distributed and highly concealed, and different types of defects such as cracks, collapses, damages and leaks usually exist at the same time, and it is difficult for the structural safety monitoring system to identify and distinguish them.

[0004] An underwater robot (Remotely Operated Vehicle, ROV) is an unmanned operation device remotely controlled by onshore personnel and relying on an umbilical cable for power and signal transmission. Compared with traditional manual inspection means, the underwater robot detection technology has the advantages of wide operation surface, long effective working time and low risk. Among them, the optical detection technology has the advantages of high sampling accuracy and low cost, and is widely used in underwater close-range inspection. The optical imaging technology will generate a large amount of video and image data during a complete underwater operation task, and the video images contain rich information closely related to structural defects.

[0005] However, affected by the absorption, scattering and refraction effects of water on light, as well as the interference of waterborne suspended solids and bubbles on the light propagation path, underwater images are prone to problems such as reduced contrast, color distortion, uneven illumination and low visibility, affecting the defect recognition accuracy and size measurement accuracy. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a method for restoring and improving the quality of deteriorated images of underwater concrete structure defects, realizing the restoration and improvement of deteriorated images of underwater concrete structure defects in scenarios such as artificial light source illumination and weak light turbidity, and effectively improving the quality of underwater images.

[0007] The present invention adopts the following technical solutions to solve the above technical problems:

[0008] A method for restoring and improving the quality of deteriorated images of underwater concrete structure defects includes the following steps:

[0009] Step 1, obtain the deteriorated image of the underwater concrete structure of the water-related building, that is, the original underwater optical image, and preprocess the original underwater optical image to obtain the preprocessed underwater optical image;

[0010] Step 2, for the preprocessed underwater optical image, based on the quadtree hierarchical search method, combined with three multi-feature prior indexes of smoothness, maximum color difference and maximum brightness, estimate the background light candidate regions corresponding to each feature prior index, and adaptively fuse all the background light candidate regions to obtain the background light region image;

[0011] Step 3, perform reverse saturation map estimation on the original underwater optical image to separate the artificial light source illumination influence region and the non-artificial light source illumination influence region; for the non-artificial light source illumination influence region, generate the first red channel transmission map based on the improved dark channel prior theory; for the artificial light source illumination influence region, generate the second red channel transmission map based on the improved dark channel prior theory, and correct the second red channel transmission map according to the transmission information of the artificial light source illumination region, and fuse the corrected second red channel transmission map with the first red channel transmission map to obtain the accurate red channel transmission map;

[0012] Step 4, use the accurate red channel transmission map obtained in Step 3 to estimate the scene depth, obtain the scene depth map, perform reverse solution of the transmission map and guided filtering calculation on the scene depth map, and estimate the green and blue channel transmission maps;

[0013] Step 5, fuse the accurate red channel transmission map with the green and blue channel transmission maps to obtain the accurate transmission map, and substitute the background light region and the accurate transmission map parameters obtained in Step 2 into the underwater optical imaging model to obtain the restored underwater optical image.

[0014] Compared with the prior art, the present invention adopting the above technical solutions has the following technical effects:

[0015] 1. The present invention realizes the accurate estimation of the underwater image background light through the quadtree hierarchical search and the multi-feature prior indexes that fuse smoothness, maximum color difference and maximum brightness, and improves the influence of uneven background light on image restoration.

[0016] 2. In view of the problem of inaccurate estimation of the transmission map in the area affected by artificial light sources, the present invention adopts a transmission map estimation method that combines the improved dark channel prior theory and the reverse saturation map theory to solve the problem of inaccurate transmission map estimation.

[0017] 3. The method proposed by the present invention fully considers the complex underwater lighting conditions and the characteristics of the image itself. The restored underwater image has been improved in terms of visual quality, detail performance, and color restoration. On the basis of restoration, the image details and clarity are further enhanced to ensure the visualization effect of underwater concrete defect features in complex environments, providing high-quality image support for defect detection and precise measurement.

[0018] 4. The method proposed by the present invention has good adaptability and robustness in processing the restoration of underwater concrete structure deterioration images affected by artificial light sources, and can be further popularized and applied to the restoration and quality improvement of underwater structure defect deterioration images of various water-related buildings, improving the clarity of underwater imaging and the accuracy of defect detection, and having broad practical application value. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 is a flowchart of the method for restoring and improving the quality of underwater concrete structure defect deterioration images of the present invention;

[0020] Figure 2 is a schematic diagram of an underwater optical imaging model;

[0021] Figure 3 is an underwater image acquisition device;

[0022] Figure 4 is a visual comparison of the image restoration effect of the indoor test artificial light source scene;

[0023] Figure 5 is a comparison of the image restoration effect of the underwater defect affected by the artificial light source in the actual project. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0024] The following details the embodiments of the present invention, and the examples of the embodiments are shown in the drawings. The embodiments described below with reference to the drawings are exemplary and are only used to explain the present invention and should not be construed as limiting the present invention.

[0025] Such as Figure 1As shown in the figure, the present invention proposes a method for restoring and improving the quality of underwater concrete structure defect degradation images. First, by fusing the quadtree hierarchical search strategy and multi-feature prior indexes, the background light in the complex underwater environment is accurately estimated. Then, based on the transmission map estimation method of the improved dark channel prior theory and the reverse saturation map theory, the problem of transmission map error caused by uneven illumination and artificial light source interference is solved. Further, by combining image restoration and enhancement technologies, the image degradation phenomenon caused by color distortion, reduced contrast, and weak light turbidity is effectively repaired. The specific steps are as follows:

[0026] Step 1, as Figure 3 shown, an underwater robot is equipped with an optical sensor to collect data information on the underwater concrete structure defects of the wading building, and the original underwater optical image is obtained. Figure 2 is the underwater optical imaging model. The collected image data is preliminarily processed, including denoising, equalizing the illumination distribution, and enhancing details, so as to improve the image quality and provide reliable basic data for the subsequent restoration steps.

[0027] Step 2, adopt the quadtree hierarchical search and fuse three multi-feature prior indexes of smoothness, maximum color difference, and maximum brightness to calculate the background light candidate area of the underwater concrete structure defect image. Specifically as follows:

[0028] The background light candidate area of the underwater image estimated by smoothness can be calculated by the following formula:

[0029]

[0030] In the formula, B1 is the intensity of the image block v in the background light candidate area based on the smoothness index, Ω(v) represents the pixel set within the image block v obtained by the quadtree hierarchical search, I c (x) represents the intensity value of pixel x on the original underwater optical image, and m1 and n1 respectively represent the width and height of the image block v.

[0031] The background light candidate area of the underwater image estimated based on the maximum color difference feature can be expressed as follows:

[0032]

[0033] In the formula, B2 is the intensity of the image block w in the background light candidate area based on the maximum color difference index, Ω(w) represents the pixel set within the image block w selected by the quadtree hierarchical search for background light estimation, and m2 and n2 respectively represent the width and height of w.

[0034] The background light candidate area B3 of the underwater image can be calculated by the average value of the top 0.1% brightest pixel points in the depth map, and can be expressed as follows:

[0035]

[0036] In the formula, J depth (x) represents the depth value of pixel x on the original underwater optical image, and P 0.1％ is composed of the set of 0.1% pixels with the minimum depth, corresponding to the area closest to the camera in the image, usually the brighter area in the underwater image, and is used to estimate the background light; B3 is the average depth value of the background light candidate area based on the maximum brightness.

[0037] Based on the estimation of the three background light candidate areas respectively, the present invention proposes a background light adaptive weight fusion estimation method. Its basic idea is to search for the weight coefficients of the background light candidate areas by adjusting different indicators to obtain a more robust background light estimation result, which can be expressed as:

[0038] B = (1 - α)[(1 - β)B1 + βB2] + αB3 (4)

[0039]

[0040] In the formula, B represents the estimated value of the final background light area image, grI represents the grayscale image of the original underwater optical image, and A grI represents the average value of the grayscale image, and A EI represents the average value after edge segmentation of the original underwater optical image, exp(·) represents the exponential function with the natural constant e as the base, and σ m represents the offset parameter, which is used to adjust the center position of the function, and σ n = 0.1, s is an empirical constant, which is set to 32. α and β are weight parameters, which are used to control the contribution of each estimated value to the final background light estimated value. α adjusts the background light estimation result through the average value of the grayscale image. If the grayscale value is higher, it indicates that the image brightness is higher, and the α value is larger, then the weight of B3 is increased. β adjusts the background light estimation result through the image edge information. If the red channel value is larger, it means that the image color deviation is larger, and the β value is larger, then the weight of B2 is increased.

[0041] Step 3, the area affected by the artificial light source irradiation and the area not affected by the artificial light source irradiation can be separated through the reverse saturation map estimation. For the area not affected by the artificial light source irradiation, the improved dark channel theory is used to directly generate the red channel transmission map. For the area affected by the artificial light source irradiation, the red channel transmission map is corrected by using the transmission information of the artificial light source illumination area to obtain the accurate red channel transmission map.

[0042] After a large number of statistical analyses of the dark channel pixel value histogram and the cumulative distribution law of underwater images, J dark = 0.1 is more in line with the actual situation of underwater images. The present invention sets J darkReplace the numerical value with 0.1, and accordingly improve the dark channel theory, and propose an improved dark channel prior method applicable to underwater concrete structure deterioration images. At both ends, a minimization operation based on a local block Ω = 9×9 is adopted, specifically as follows:

[0043]

[0044] In the formula, the background light is known. Divide both sides of the above formula by respectively, and the formula can be rewritten as:

[0045]

[0046] In the formula, in a very small local block, the transmittance t c (x) is considered to be a constant. Therefore, the above formula can be further expressed as:

[0047]

[0048] Apply the minimum filter algorithm to the red, green, and blue channels of the image, which can be expressed as:

[0049]

[0050] In the formula, can be further expressed as Replace with V, and there is the following inequality relationship, as follows:

[0051]

[0052] In the formula, Then the above formula can be rewritten in the following inequality form:

[0053]

[0054] According to the Lambert-Beer law, the transmittance t c (x) can also be expressed as an exponential combination form of the scene depth and the spectral attenuation coefficient, that is, as shown in the following formula:

[0055] t c (x) = Nrer(c) d(x) (13)

[0056] In the formula, Nrer(c) represents the spectral attenuation coefficient. The radiation transfer rule is used to reflect the scattering and absorption behavior of light during propagation in the medium. The value of Nrer(c) can be expressed as follows:

[0057]

[0058] As can be seen from Equation (14), Nrer(c) has the smallest value in the red channel. That is, when d(x) is fixed, the transmittance of the red channel is the smallest, and it can be obtained that in the local block Therefore, Equation (12) is rewritten as the following equation:

[0059]

[0060] In the formula, t r (x) ranges from 0 to 1. When V is larger, t r (x) is more likely to be greater than 1. In order to reduce the loss of transmittance information, V can be set to Using B max to represent the maximum value of the background light in the red, green, and blue channels, and finally the mathematical expression of the red channel transmittance is derived as follows:

[0061]

[0062] In the formula, B r,∞ and B c′,∞ respectively represent the red channel background light and the blue-green channel background light obtained from the farthest point in the image. The wavelengths of the standard red, green, and blue channels are λ r = 620nm, λ g = 540nm, and λ b = 450nm. Then, the transmittances t g (x) and t b (x) of the green and blue channels in the underwater image can be expressed in the following form:

[0063]

[0064] In the actual underwater detection scenario, there is a lack of sufficient natural light, and artificial light sources are mostly used to provide auxiliary lighting. The pixel value intensity of the area of the underwater structure irradiated by the artificial light source will increase on the image, which easily leads to the inaccurate estimation of the transmission map by the dark channel prior theory. In view of this, the present invention uses the reverse saturation map to separate the areas affected by the artificial light source and the areas not affected by the artificial light source. For the areas not affected by the artificial light source, the improved dark channel prior theory is used to calculate the transmission map. For the areas affected by the artificial light source irradiation, the reverse saturation map is used to calculate the transmission map. The above-obtained transmission maps are fused to obtain the red channel transmission corrected transmission map, and then the transmission maps of other channels are obtained.

[0065] In conventional underwater optical imaging, there is a general rule that "red light attenuates, and the intensities of the blue-green color channels are stronger than those of the red channel", but the illumination of artificial light sources will change this rule. Under the illumination of artificial light sources, light in different bands will be compensated, the radiance of the red channel in the illuminated area will increase, and the pixel color difference between lights in different bands will decrease. At this time, the color components of the three channels tend to be evenly distributed, resulting in a saturation close to zero. When there is no illumination of artificial light sources in the image acquisition scene, the saturation of the underwater image is not zero. At this time, the lack of saturation can be used to reflect the number of white lights existing within the color of the image, that is, the influence degree of the artificial light source.

[0066] Therefore, the present invention uses reverse saturation as an index for determining the influence of artificial light sources to separate the area affected by artificial light sources from other areas of the image. The saturation index can be expressed as:

[0067]

[0068] In the formula, Sat(I) represents the image saturation, and I R 、I G and I B represent the illumination intensities of the red channel, green channel, and blue channel respectively.

[0069] When a color is in a pure spectrum, it is completely saturated and does not contain any white light. When white light is added, a color will lose saturation because white light includes the energy of all wavelengths. Therefore, the lack of saturation can be used to reveal the number of white lights appearing within the image color. The unsaturated areas in the image can be interpreted as areas where there is illumination of artificial light sources. Therefore, the Reversed Saturation Map (RSM) can be used to describe the influence degree of artificial light source illumination. Areas with a relatively high reversed saturation map are usually regarded as areas affected by artificial light sources, and can be expressed as:

[0070] Sat rev =1 - Sat(I) (21)

[0071] In the formula, Sat rev represents the reversed saturation map. For areas with low reversed saturation, that is, areas not affected by the illumination of artificial light sources, the transmission map can be estimated according to the improved dark channel prior theory formula proposed by the present invention. For areas with high reversed saturation, due to the influence of being close to the artificial light source, the pixel values increase rapidly. At this time, using formula (16) to estimate the transmission map will be too small, resulting in an unsatisfactory restoration effect. Therefore, the transmission information of the artificial illumination area can be used to correct the dark channel transmission map to reduce the influence of the artificial light source, which is expressed as:

[0072]

[0073] Step 4: Use the calculated accurate transmission map of the red channel to perform scene depth estimation to obtain a depth map, and estimate the transmission maps of the blue and green channels by combining inverse solution of the transmission map and calculations such as guided filtering.

[0074] Step 5: Fuse the transmission maps of the red, green, and blue channels to obtain an accurate transmission map:

[0075]

[0076] In the formula, λ ∈ [0, 1] represents a scalar multiplier to adapt to the number of artificial light sources.

[0077] Substitute the obtained background light parameters and the calculated accurate transmission map into the underwater optical imaging model to obtain the restored underwater defect image. The specific formula is as follows:

[0078]

[0079] In the formula, c = r, g, b, J c represents the illumination intensity of the restored underwater image, represents the illumination intensity of the pixel x on the background light of the calculated underwater image, t c (x) represents the transmittance of the pixel x on the estimated underwater image transmission map, and its upper and lower limits are set to 0.9 and 0.2 respectively. Figure 4 is the visual comparison of the image restoration effect of the indoor test artificial light source scene; Figure 5 is the comparison of the underwater defect image restoration effect affected by the actual engineering artificial light source. From Figure 4 and Figure 5 it can be seen that the restored image obtained by the method of the present invention has higher details and clarity, and better visibility effect.

[0080] Based on the same inventive concept, an embodiment of the present application provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the foregoing method for restoring and improving the quality of the underwater concrete structure defect degradation image are implemented.

[0081] Based on the same inventive concept, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the steps of the foregoing method for restoring and improving the quality of the underwater concrete structure defect degradation image are implemented.

[0082] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0083] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for realizing the functions specified in Figure 1 one or more of the processes Figure 1 or blocks.

[0084] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that realize the functions specified in Figure 1 one or more of the processes Figure 1 or blocks.

[0085] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for realizing the functions specified in Figure 1 one or more of the processes Figure 1 or blocks.

[0086] The above embodiments are only for illustrating the technical idea of the present invention and cannot be used to limit the protection scope of the present invention. Any modifications made on the basis of the technical solution according to the technical idea proposed by the present invention fall within the protection scope of the present invention.

Claims

1. A method for restoring and improving the quality of an underwater concrete structure defect deterioration image, characterized in that: The steps include: Step 1, obtaining an underwater concrete structure defect degradation image of a water-related building, that is, an original underwater optical image, and preprocessing the original underwater optical image to obtain a preprocessed underwater optical image; Step 2: for the preprocessed underwater optical image, based on the quadtree hierarchical search method, combined with three types of multi-feature prior indicators of smoothness, maximum color difference and maximum brightness, estimate the background light candidate area corresponding to each feature prior indicator, and adaptively fuse all the background light candidate areas to obtain the background light area image; Step 3, performing reverse saturation map estimation on the original underwater optical image, separating the area affected by artificial light source illumination and the area affected by non-artificial light source illumination; for the area affected by non-artificial light source illumination, generating a first red channel transmission map based on the improved dark channel prior theory; for the area affected by artificial light source illumination, generating a second red channel transmission map based on the improved dark channel prior theory, and correcting the second red channel transmission map according to the transmission information of the artificial light source illumination area, and fusing the corrected second red channel transmission map with the first red channel transmission map to obtain a red channel accurate transmission map; Step 4, using the accurate red channel transmission map obtained in step 3 to estimate the scene depth, obtain a scene depth map, perform a transmission map inverse solution and guided filter calculation on the scene depth map, and estimate the green and blue channel transmission maps; Step 5: Fuse the red channel precise transmission map with the green and blue channel transmission maps to obtain a precise transmission map. Substitute the background light area and precise transmission map parameters obtained in step 2 into the underwater optical imaging model to obtain a restored underwater optical image.

2. The method for restoring and improving the quality of an underwater concrete structure defect deterioration image according to claim 1, characterized in that: In the step 1, an underwater robot equipped with an optical sensor is used to collect images of underwater concrete structure defects and degradation of water-related buildings, and the collected images are preprocessed, including denoising, light distribution balancing, and detail enhancement operations, to obtain preprocessed underwater optical images.

3. The method for restoring and improving the quality of an underwater concrete structure defect deterioration image according to claim 1, characterized in that: The specific process of step 2 is as follows: Step 21, for the pre-processed underwater optical image, a quadtree hierarchical search is performed based on the smoothness feature index, and the background light candidate area is estimated according to the search results. The formula is as follows: In the formula, I c (x) represents the intensity value of pixel x on the original underwater optical image, Ω(v) represents the pixel set in the image block v obtained based on the quadtree hierarchical search, B1 represents the intensity of the image block v in the background light candidate area based on smoothness, and m1 and n1 represent the width and height of the image block v, respectively; Step 22, for the pre-processed underwater optical image, a quadtree hierarchical search is performed based on the maximum color difference feature index, and the background light candidate area is estimated according to the search results. The formula is as follows: Where Ω(w) represents the pixel set in the image region w obtained based on the quadtree hierarchical search, B2 represents the intensity of the image block w in the background light candidate region based on the maximum color difference, and m2 and n2 represent the width and height of the image region w, respectively; Step 23, obtain the depth map of the original underwater optical image, and estimate the background light candidate area based on the first 0.1% of the brightest pixels in the depth map, using the following formula: In the formula, J depth (x) represents the depth value of pixel x in the original underwater optical image; P 0.1％ Indicates that the depth values ​​of each pixel in the depth map are sorted from small to large, and the pixel set with the top 0.1% of the depth value is sorted; |P 0.1％ | represents the number of pixels in the top 0.1% of the depth value ranking, and B3 represents the average depth value of the background light candidate area based on the maximum brightness; Step 24, adaptively weight fusion is performed on the background light candidate regions obtained in steps 21-23 to obtain an underwater background light region image, and the formula is as follows: B=(1-α)[(1-β)B1+βB2]+αB3 In the formula, B represents the fused background light area image, α and β are weight parameters, s is an empirical constant, and A grI Represents the average value of the grayscale image of the original underwater optical image, A EI represents the average value of the original underwater optical image after edge segmentation, σ m is the offset parameter, σ n =0.

1.

4. The method for restoring and improving the quality of an underwater concrete structure defect deterioration image according to claim 1, characterized in that: In step 3, the reverse saturation is used as an indicator to determine the influence of the artificial light source, and the area affected by the artificial light source and the area not affected by the artificial light source are separated. The reverse saturation is expressed as: Sat rev =1-Sat(I) In the formula, Sat rev Indicates the inverse saturation of the image, Sat(I) indicates the image saturation, and Sat(I) is expressed as: In the formula, I R ,I G and I B Represent the light intensity of the red channel, green channel and blue channel respectively, and I represents the light intensity; the pixels whose reverse saturation exceeds the preset threshold are divided into the area affected by artificial light source, and the remaining pixels are divided into the area not affected by artificial light source; Let the dark channel prior knowledge in the dark channel prior theory be dark = 0.1, the dark channel prior theory is improved; the first and second red channel transmission images generated by the improved dark channel prior theory are expressed as follows: Where, t r (x) represents the transmittance of pixel x on the first or second red channel transmission image, I c (x) represents the intensity value of pixel x in the original underwater optical image, B max represents the maximum value of the background light in the red, green and blue channels, and Ω(x) represents the area range of a preset size centered on pixel x; The second red channel transmission map is corrected according to the transmission information of the artificial light source irradiation area. The corrected second red channel transmission map is expressed as: In the formula, Represents the transmittance of pixel x on the corrected second red channel transmission map.

5. The method for restoring and improving the quality of an underwater concrete structure defect deterioration image according to claim 4, characterized in that: In step 4, the green and blue channel transmission images are shown as follows: Where, t g (x), t b (x) represents the transmittance of pixel x on the green and blue channel transmission images, respectively, g , b Represent the wavelengths of the green and blue channels respectively.

6. The method for restoring and improving the quality of an underwater concrete structure defect deterioration image according to claim 5, characterized in that: In step 5, the precise transmission map is represented as follows: Where, t c (x) represents the transmittance of pixel x on the precise transmission map, c = r, g, b, ω represents the weight factor, λ∈[0,1] represents the scalar multiplier, I r (x), I g (x), I b (x) represents the light intensity of pixel x in the red, green, and blue channels, respectively. represents the illumination intensity of pixel x on the fused background light area image, and Sat(x) represents the saturation at pixel x; The restored underwater optical image is shown as follows: In the formula, J c represents the illumination intensity of the restored underwater optical image, I c (x) represents the intensity value of pixel x in the original underwater optical image, t c (x) represents the transmittance of pixel x on the precise transmission map.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and capable of running on the processor, characterized in that: When the processor executes the computer program, the steps of the method for restoring and improving the quality of an underwater concrete structure defect deterioration image as described in any one of claims 1 to 6 are implemented.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method for restoring and improving the quality of an underwater concrete structure defect deterioration image are implemented as described in any one of claims 1 to 6.

Images