A prefabricated box beam template detection method and system

Abstract

The present invention relates to the field of image data processing, and more specifically, the present invention relates to a method and system for detecting a prefabricated box girder template, the method comprising: obtaining a surface image of a prefabricated box girder template, performing edge detection on the surface image, and marking it as a pixel point set to be analyzed; calculating the degree of defect expression of the pixel points in the pixel point set to be analyzed, and obtaining the possibility that each edge connected domain in the pixel point set to be analyzed is a bulge area according to the defect expression degree; in response to the possibility that each edge connected domain is a bulge area being greater than or equal to a defect threshold, the edge connected domain is a bulge area, and the pixel point with the lowest grayscale value in the corresponding bulge area is selected as a seed point; all selected seed points are used as input points of a watershed algorithm to identify bulge defects in the surface image. The present invention analyzes the change characteristics of the grayscale value of the bulge area, thereby adaptively determining the seed point of the watershed algorithm and improving the accuracy of detection.

Description

Technical Field

[0001] The present invention relates to the field of image data processing, and more specifically, to a method and system for detecting a prefabricated box beam template. Background Art

[0002] Box girder prefabrication is an important part of the construction of expressways and urban expressways. Its quality directly affects the quality of engineering construction, and the quality of prefabricated box girder templates directly affects the quality of box girders. Among them, the bulging defects of prefabricated box girder templates not only affect the appearance of the box girder, but also may affect its structural performance and durability. For example, the bulging area of ​​the prefabricated box girder template causes bulging of the box girder. The concrete in the bulging area may not be compact inside, and there may be holes or honeycombs, which reduces the bearing capacity and seismic performance of the structure.

[0003] The existing Chinese patent application document with publication number CN113935965A discloses a prefabricated box beam template detection system, including: a scanning module for scanning the prefabricated box beam template to be detected; a three-dimensional model construction module for constructing a three-dimensional model corresponding to the prefabricated box beam template based on the data scanned by the scanning module; a comparison module for comparing the three-dimensional model with the standard model to determine the comparison effect diagram; and an output module for outputting the comparison effect diagram. The prefabricated box beam template detection system of the present invention realizes automatic detection of the prefabricated box beam template.

[0004] In the above scheme, the box girder formwork is inspected by means of a scanning module, a comparison module and an output module, and it is mainly aimed at the inspection that has been used in construction. At the same time, when the left and right swing of the scanning equipment is too large or there is an object blocking it, it will have an adverse effect on the scanning accuracy. In addition, bulging defects will occur in the prefabricated box girder production process. At present, the existing technology is to identify the surface image after production, and usually uses the watershed algorithm. When selecting seed points, it often uses the image gradient method to determine, that is, simply setting the place with a larger gradient amplitude as the seed point. However, in the bulging defect detection process of the prefabricated box girder formwork, due to the possible existence of other normal grayscale value changes in the surface image, the method of determining the seed point by the gradient method will have errors, which will affect the segmentation result of the bulging defect, and further affect the detection result of the prefabricated box girder formwork. Summary of the invention

[0005] In order to solve the problem that other normal gray value changes may exist in the surface image, resulting in errors in the method of determining the seed point using the gradient method, which in turn affects the segmentation result of the bulge defect, the present invention provides solutions in the following aspects.

[0006] In a first aspect, a method for detecting a prefabricated box girder template includes: obtaining a surface image of a prefabricated box girder template, performing edge detection on the surface image, obtaining edge detection results, and marking them as a pixel set to be analyzed; calculating the degree of defect expression of pixels in the pixel set to be analyzed, and obtaining the possibility that each edge connected domain in the pixel set to be analyzed is a bulge area according to the defect expression degree; in response to the possibility that each edge connected domain is a bulge area being greater than or equal to a defect threshold, the edge connected domain is a bulge area, selecting the pixel with the lowest grayscale value in the corresponding bulge area as a seed point, and traversing all edge connected domains; using all selected seed points as input points of a watershed algorithm to identify bulge defects in the surface image, thereby performing positioning and segmentation, evaluating the severity of the bulge defects, and generating a visual defect detection result; wherein the possibility of the bulge area: , where Indicates the first pixel in the set to be analyzed The possibility that the edge connected domain is a bulge area, Indicates The number of pixels in the edge connected domain, Indicates The edge connected domain The degree of defect performance of each pixel, Indicates The length of the edge connected region, Indicates The side length of the minimum circumscribed rectangle of an edge-connected region.

[0007] The effect is: through the comprehensive application of edge detection, local grayscale analysis and watershed algorithm, the automatic, high-precision identification and evaluation of bulge defects are realized; it not only improves the accuracy and efficiency of detection, but also enhances the convenience of manual review by generating intuitive visual results, thereby ensuring the quality control of prefabricated box girder formwork.

[0008] Preferably, edge detection is performed on the surface image to obtain edge detection results, which are marked as a pixel point set to be analyzed, including:

[0009] The edge pixels in the edge detection result are used as a pixel set to be analyzed, and the pixel set to be analyzed includes edges caused by bulging areas, dust, oil stains or other pollutants.

[0010] Preferably, the degree of defect manifestation includes:

[0011] Any pixel in the pixel set to be analyzed is taken as the target pixel, and a preset window size with the target point as the center is taken as the local range. A surface is fitted to the pixels in the local range, and the mean difference between the grayscale value of the pixel in the local range on the fitting surface and the actual grayscale value of the pixel is calculated. The mean difference is normalized as the degree of defect manifestation.

[0012] The effect is that the fitting and analysis method of local area grayscale difference can more accurately identify tiny defects in the image, reduce the possibility of misjudgment and missed judgment, and significantly improve the performance and efficiency of defect detection in prefabricated box girder formwork re-inspection.

[0013] Preferably, the defect manifestation degree also includes:

[0014] Any pixel in the pixel set to be analyzed is taken as the target pixel, a preset window size with the target point as the center is taken as the local range, a surface is fitted to the pixels in the local range, and the mean difference between the grayscale value of the pixel in the local range on the fitting surface and the actual grayscale value of the pixel is calculated;

[0015] Obtain the gradient direction of each pixel in the local range, and calculate the standard deviation of all gradient directions in the local range as the credibility;

[0016] The ratio of the difference mean to the credibility is normalized and used as the defect manifestation degree.

[0017] The effect is: by further calculating the consistency of the gradient direction and comprehensively considering the credibility of the grayscale difference and the gradient direction consistency, a more comprehensive and reliable defect detection method is provided, which effectively improves the quality control and detection efficiency of the prefabricated box girder formwork.

[0018] Preferably, fitting a curved surface to the pixel points within the local range includes:

[0019] Use polynomial fitting to perform surface fitting on the pixels in the local range, where the index of the pixel in the local range is axis, Axis, where The axis represents the local range. OK, The axis represents the local range. Column, with the gray value corresponding to the pixel point axis, and the fitting result is .

[0020] The effect is that the grayscale change analysis of the local area of ​​the image is enhanced through polynomial fitting, the accuracy and reliability of defect detection are improved, and thus the detection efficiency of the prefabricated box girder formwork is effectively improved.

[0021] Preferably, the method of using all selected seed points as input points of a watershed algorithm to identify bulge defects in a surface image, thereby performing positioning and segmentation, includes:

[0022] Among them, the watershed algorithm starts from the seed point and simulates the process of water flowing to the low-lying area. The water will flow along the path with a smaller gradient until it encounters the water flow of other seed points. When the water flows meet, a watershed is formed, which divides the image into different areas and identifies the bulge defects.

[0023] The effect is: by improving the way of obtaining seed points in the watershed algorithm, the accuracy of the watershed algorithm in detecting prefabricated box girders is improved. The principle of the watershed algorithm is to simulate the water flow process and provide an effective image segmentation and defect recognition method, which helps to improve the quality control and detection efficiency of prefabricated box girder templates.

[0024] Preferably, evaluating the severity of the bulge defect and generating a visual defect detection result includes:

[0025] The defect information is sorted out according to the bulge defect, wherein the defect information includes: position, size and shape, etc. The severity of the bulge defect is evaluated according to the characteristics of the bulge defect, wherein the characteristics include: area, shape and irregularity, etc.

[0026] In a second aspect, a prefabricated box girder formwork detection system includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned prefabricated box girder formwork detection method is implemented.

[0027] The present invention has the following effects:

[0028] 1. The present invention analyzes the changing characteristics of the grayscale value of the bulge area, thereby adaptively determining the seed point of the watershed algorithm, improving the accuracy of the segmentation of the bulge defect of the prefabricated box girder template, thereby improving the automation and robustness of the algorithm and improving the detection efficiency.

[0029] 2. The present invention determines the seed points by adaptively determining them. Compared with the gradient method for selecting seed points, the present invention is easily affected by the local gradient changes in the image, causing the algorithm to generate too many small areas, thereby causing the problem of over-segmentation. In addition, the gradient method also needs to determine the appropriate gradient parameters. The present invention avoids the determination of gradient parameters, thereby further improving the automation and robustness of the algorithm. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0031] Figure 1 It is a method flow chart of steps S1 to S4 in a prefabricated box girder formwork detection method according to an embodiment of the present invention.

[0032] Figure 2 It is a structural block diagram of a prefabricated box girder formwork detection system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0034] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0035] Reference Figure 1 A prefabricated box beam template detection method includes steps S1 to S4, which are specifically as follows:

[0036] S1: Obtain a surface image of a prefabricated box girder template, perform edge detection on the surface image, obtain edge detection results, and mark them as a pixel point set to be analyzed.

[0037] An industrial camera is used to vertically capture the surface image of the prefabricated box template.

[0038] It should be noted that in the surface image of the prefabricated box girder formwork, the presence of dust, oil or other pollutants on the surface may reduce the surface reflectivity, thereby reducing the grayscale value in the collected image. If the gradient method is used to determine the seed point of the watershed algorithm, such an area will also generate seed points. In the final segmentation result of the watershed algorithm, the seed points may be in other areas of the bulge area, resulting in inaccurate detection results.

[0039] The bulging defect of the prefabricated box girder formwork is a defect that expands or bulges outward in the defective area of ​​the formwork. In its surface image, due to the effect of light, the grayscale value of the bulging part will be significantly higher than that of the normal area, and a faint shadow will be produced around the bulge, resulting in a lower grayscale value than the normal area. This will produce an obvious gradient. Therefore, the edge detection algorithm can be used to identify the edge pixels in the image, thereby screening the pixels.

[0040] The edge pixels in the edge detection result are used as a pixel set to be analyzed, and the pixel set to be analyzed includes edges caused by bulging areas, dust, oil or other pollutants.

[0041] In this embodiment, the canny edge detection algorithm is used. This technology is a common technology for those skilled in the art and will not be described in detail.

[0042] Further analysis shows that because the grayscale value in the local range of the bulge area will gradually change in a certain direction under the influence of light, while the grayscale value changes in the local range of the edge pixels of dust, oil or other pollutants are more chaotic. Therefore, the regularity of the grayscale value change in the local range of an edge pixel can be used to calculate the defect performance degree of each pixel in the pixel set to be analyzed.

[0043] S2: Calculate the defect manifestation degree of the pixel points in the pixel point set to be analyzed, and obtain the possibility that each edge connected domain in the pixel point set to be analyzed is a bulge area according to the defect manifestation degree.

[0044] It should be noted that the regularity of the pixels in the pixel set to be analyzed within the local range can be represented by the fitting error in the surface fitting process of the pixels within the local range. This is because if the gray value changes regularly within the local range, each pixel can be well fitted on the surface during surface fitting. On the contrary, if the gray value changes chaotically within the local range, the error of the pixel during surface fitting is large. Therefore, the larger the fitting error, the worse the regularity, which is negatively correlated.

[0045] Any pixel in the pixel set to be analyzed is taken as the target pixel, and a preset window size with the target point as the center is used as the local range. A surface is fitted to the pixels in the local range, and the mean difference between the grayscale value of the pixel in the local range on the fitting surface and the actual grayscale value of the pixel is calculated. The mean difference is normalized as the degree of defect manifestation.

[0046] In this embodiment, the preset window is , , can be adjusted according to actual conditions, or set to radius The circular local area.

[0047] Specifically, the curve performance degree satisfies the following relationship:

[0048] ,

[0049] In the formula, Indicates the pixel points in the pixel set to be analyzed The degree of defect manifestation, Represents pixel The local range window side length, Indicates the number of pixels in the local area. Represents pixel In local area Line Column pixels, Represents pixel The local range, represents the maximum and minimum normalized function, Represented by natural numbers An exponential function with base .

[0050] At the same time, in order to improve the credibility of the regularity calculation of the bulge defect area, because the grayscale value of a certain pixel in the local range of the bulge defect area will gradually change along a certain direction under the influence of light, which means that the direction of the gradient value of each pixel is relatively consistent. Therefore, the consistency of the gradient direction of the pixel in the local range is used as the credibility of the regularity. The more consistent the gradient direction, the higher the credibility of the regularity. In summary, the degree of defect expression can be obtained through the regularity of the pixel and its credibility. The higher the regularity and the higher the credibility, the higher the degree of defect expression.

[0051] In addition, in another embodiment, any pixel in the pixel set to be analyzed is taken as a target pixel, a preset window size with the target point as the center is taken as a local range, a surface is fitted to the pixels in the local range, and the mean difference between the grayscale value of the pixel in the local range on the fitting surface and the actual grayscale value of the pixel is calculated;

[0052] Obtain the gradient direction of each pixel in the local range, and calculate the standard deviation of all gradient directions in the local range as the credibility;

[0053] To further explain, in this embodiment, the Sobel operator is used to calculate the gradient direction of each pixel point, and the consistency of the gradient direction can be characterized according to the standard deviation of the gradient directions of all pixel points, that is, the credibility of the consistency rule, wherein the larger the standard deviation, the worse the consistency of the gradient direction.

[0054] The ratio of the difference mean to the credibility is normalized and used as the defect performance degree.

[0055] Specifically, the curve performance degree satisfies the following relationship:

[0056] ,

[0057] In the formula, Indicates the pixel points in the set of pixels to be analyzed The degree of defect manifestation, Represents pixel The local range window side length, Indicates the number of pixels in the local area. Represents pixel In local area Line Column pixels, Represents pixel The local range, Represents pixel In local area Line Column pixels, Represents pixel In local area Line The gradient direction of the column pixel, represents the maximum and minimum normalized function, Represented by natural numbers The exponential function with base , Represents the standard deviation function.

[0058] It should be noted that Indicates Line The values ​​of the pixels in the column on the fitting surface, It represents the difference between the value of the pixel point on the fitting surface and its true gray value, that is, the pixel point In local area Line The true gray value of the pixel in the column, Indicates Line The fitting error of the pixel points in the column during the surface fitting process; Represents pixel The average fitting error of the pixels in the local range during the surface fitting process, then the negative exponential power Represents pixel The regularity of gray value changes within a local range.

[0059] Represents pixel In local area Line The gradient direction of the pixel points in the column, Represents pixel The standard deviation of the gradient direction of all pixels in the local range is Indicates the consistency of the gradient direction and also indicates the pixel point The credibility of the regularity of gray value changes in a local area.

[0060] The pixel points in the local range are fitted with a surface, including the following steps:

[0061] Use polynomial fitting to perform surface fitting on the pixels in the local range, where the index of the pixel in the local range is axis, Axis, where The axis represents the local range. OK, The axis represents the local range. Column, with the gray value corresponding to the pixel point axis, and the fitting result is .

[0062] It should be noted that the fitting error is the average of the difference between the value of the pixel point on the fitting surface in the local range and its true gray value. The larger the value, the greater the fitting error.

[0063] Further explanation: since bulge defects often appear as circles or ellipses, the greater the average defect expression level of each edge connected domain in the pixel point set to be analyzed, and the larger the ratio of the side length of the edge connected domain to its minimum circumscribed rectangle, the more circular or elliptical the edge connected domain appears, and the greater the possibility that the area is the edge of a bulge defect.

[0064] Specifically, the possibility of the bulge area satisfies the following relationship:

[0065] ,

[0066] In the formula, Indicates the first pixel in the set to be analyzed The possibility that the edge connected domain is a bulge area, Indicates The number of pixels in the edge connected domain, Indicates The edge connected domain The degree of defect performance of each pixel, Indicates The length of the edge connected region, Indicates The side length of the minimum circumscribed rectangle of an edge-connected region.

[0067] Further explanation: represents the average defect level of the edge connected domain; Represents the ratio of the side length of the edge connected domain to its minimum enclosing rectangle.

[0068] S3: In response to the possibility that each edge connected domain is a bulge area being greater than or equal to the defect threshold, the edge connected domain is a bulge area, and the pixel point with the lowest grayscale value in the corresponding bulge area is selected as a seed point, and all edge connected domains are traversed.

[0069] In this embodiment, the defect threshold is set to 0.67, that is, when the probability is greater than or equal to the defect threshold 0.67, it is a bulge area, otherwise, it is an edge caused by dust, oil or other pollutants. Traverse each edge connected domain, and when the edge connected domain is a bulge area, determine a seed point until all edge connected domains are traversed. In particular, if no edge connected domain meets the above conditions, it means that the prefabricated box beam template does not have a bulge defect.

[0070] S4: All selected seed points are used as input points of the watershed algorithm to identify bulge defects in the surface image, thereby locating and segmenting them, evaluating the severity of the bulge defects, and generating visual defect detection results.

[0071] Among them, the watershed algorithm starts from the seed point and simulates the process of water flowing to the low-lying area. The water will flow along the path with a smaller gradient until it encounters the water flow of other seed points. When the water flows meet, a watershed is formed, which divides the image into different areas and identifies the defects containing bulges.

[0072] The defect information is sorted out according to the bulge defect, wherein the defect information includes: location, size and shape, etc. The severity of the bulge defect is evaluated according to its characteristics, wherein the characteristics include: area, shape and irregularity, etc.

[0073] To further illustrate, after the segmentation is completed, the size, shape and other characteristics of the bulge can be measured, the severity of the defect can be evaluated, a defect detection report can be generated and visual results can be obtained, such as marking the defect location in the original image to facilitate manual review and processing.

[0074] The method of the present invention ends here.

[0075] The present invention also provides a prefabricated box beam template detection system. Figure 2 As shown, the system includes a processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a prefabricated box beam template detection method according to the first aspect of the present invention is implemented.

[0076] The system also includes other components familiar to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and will not be described in detail here.

[0077] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of a device or accessible or connectable to a device. Any application or module described in the present invention may be implemented using computer-readable / executable instructions that may be stored or otherwise maintained by such a computer-readable medium.

[0078] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0079] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A method for detecting a prefabricated box beam formwork, characterized in that: include: Acquire a surface image of a prefabricated box beam template, perform edge detection on the surface image, obtain edge detection results, and mark them as a pixel point set to be analyzed; Any pixel in the pixel set to be analyzed is taken as the target pixel, and a preset window size with the target point as the center is taken as the local range. A surface is fitted to the pixels in the local range, and the mean difference between the grayscale value of the pixel in the local range on the fitting surface and the actual grayscale value of the pixel is calculated; the gradient direction of each pixel in the local range is obtained, and the standard deviation of all gradient directions in the local range is calculated as the credibility; the ratio of the mean difference to the credibility is normalized and used as the degree of defect manifestation, and the possibility of each edge connected domain in the pixel set to be analyzed being a bulge area is obtained according to the degree of defect manifestation; In response to the possibility that each edge connected domain is a bulge area being greater than or equal to the defect threshold, the edge connected domain is a bulge area, and the pixel point with the lowest gray value in the corresponding bulge area is selected as a seed point, and all edge connected domains are traversed; All selected seed points are used as input points of the watershed algorithm to identify bulge defects in the surface image, thereby locating and segmenting them, evaluating the severity of the bulge defects, and generating visual defect detection results; Among them, the possibility of the bulge area is: , where Indicates the first pixel in the set to be analyzed The possibility that the edge connected domain is a bulge area, Indicates The number of pixels in the edge connected domain, Indicates The edge connected domain The degree of defect performance of each pixel, Indicates The length of the edge connected region, Indicates The side length of the minimum circumscribed rectangle of an edge-connected region.

2. A prefabricated box beam template detection method according to claim 1, characterized in that: Perform edge detection on the surface image to obtain edge detection results, which are marked as a pixel point set to be analyzed, including: The edge pixels in the edge detection result are used as a pixel set to be analyzed, and the pixel set to be analyzed includes edges caused by bulging areas, dust, oil stains or other pollutants.

3. A method for detecting a prefabricated box beam formwork according to claim 1, characterized in that: Fitting a surface to the pixel points within the local range includes: Use polynomial fitting to perform surface fitting on the pixels in the local range, where the index of the pixel in the local range is axis, Axis, where The axis represents the local range. OK, The axis represents the local range. Column, with the gray value corresponding to the pixel point axis, and the fitting result is .

4. A method for detecting a prefabricated box beam formwork according to claim 1, characterized in that: All selected seed points are used as input points of the watershed algorithm to identify bulge defects in the surface image, thereby performing positioning and segmentation. include: Among them, the watershed algorithm starts from the seed point and simulates the process of water flowing to the low-lying area. The water will flow along the path with a smaller gradient until it encounters the water flow of other seed points. When the water flows meet, a watershed is formed, which divides the image into different areas and identifies the bulge defects.

5. A method for detecting prefabricated box beam formwork according to claim 1, characterized in that: Evaluate the severity of bulge defects and generate visual defect detection results, including: Defect information is sorted out according to the bulge defect, wherein the defect information includes: position, size and shape, and the severity of the bulge defect is evaluated according to its characteristics, wherein the characteristics include: area, shape and irregularity.

6. A prefabricated box beam template detection system, characterized in that: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the prefabricated box girder formwork detection method according to any one of claims 1 to 5 is implemented.

Images