A method for detecting defects in basketball floors

Abstract

The present invention discloses a basketball floor defect detection method, which belongs to the field of image analysis technology, and includes an image acquisition module, a comparison analysis module, a defect segmentation module, and an execution repair module. The present invention can analyze each defect area to be segmented separately through the defect segmentation module, and the defect segmentation module includes a damage analysis unit and a pattern analysis unit. The damage analysis unit can determine whether there is a damage gap on the floor surface through structured light technology, and the pattern analysis unit can analyze whether there is pattern wear and loss, thereby making the detection defect type range wider and improving the detection effect.

Description

Technical Field

[0001] The present invention relates to the technical field of image analysis, and in particular to a method for detecting defects in a basketball floor. Background Art

[0002] Basketball floor defect detection is a quality inspection and evaluation process specifically for basketball court floors. The purpose of this test is to discover various problems that may occur during the use of the floor, such as surface scratches, dents, bulges, degumming, cracking, etc., as well as hidden dangers such as hollowing and looseness in the floor structure. Through professional testing tools and methods, key indicators such as moisture content, hardness, sliding performance, and elasticity of the floor can be accurately evaluated to ensure that the floor meets the requirements for use in sports venues. Regular basketball floor defect detection is an important measure to maintain the quality of the venue, extend the service life of the floor, and prevent safety accidents.

[0003] Existing basketball floor defect detection methods generally detect floor surface defects through image analysis. For example, a basketball floor surface flatness visual detection method disclosed in Chinese Patent Publication No. CN117314902B can specifically analyze basketball floor surface flatness defects based on the grayscale histogram of the basketball floor image combined with a Gaussian mixture model. However, this method is difficult to detect wear and fading or stain adhesion caused by the use of basketball floors over time, which makes it difficult to apply this method to basketball floors in ordinary basketball venues that lack maintenance and are easily worn, and has application limitations. Summary of the invention

[0004] The purpose of the present invention is to propose a basketball floor defect detection method in order to solve the problems of limited usage scenarios and application limitations of basketball floor defect detection methods in the prior art.

[0005] The purpose of the present invention can be achieved by the following technical solution: A basketball floor defect detection method includes an image acquisition module, a comparison and analysis module, a defect segmentation module and an execution repair module. The specific detection steps are as follows:

[0006] Step 1: The image acquisition module is used to acquire a high-definition image of the basketball court, including a device analysis unit and an image acquisition unit. The device analysis unit divides the entire basketball court into multiple rectangular unit areas according to the variable focal length, resolution and installation height of the image acquisition device, and the image acquisition unit acquires high-definition images of all current unit areas;

[0007] Step 2: The comparison and analysis module compares and analyzes the high-definition images of the unit area, including an image comparison unit and an image analysis unit. The image comparison unit compares the currently acquired high-definition image with the first and last acquired high-definition images respectively. The image analysis unit calculates the similarity judgment value of each area according to the comparison result, and selects the area with defects as the defective area to be classified;

[0008] Step 3: The defect segmentation module analyzes each defect area to be segmented separately, including a damage analysis unit and a pattern analysis unit. The damage analysis unit analyzes the high-definition image of the defect area to be segmented and generates a surface defect signal. The pattern analysis unit analyzes the high-definition image of the defect area to be segmented and generates a pattern missing signal or a stain coverage signal.

[0009] Step 4: Execute the repair module to obtain the generated signal. When a surface defect signal is generated, an instruction to carry a repair tool is generated and sent to the staff's mobile terminal. When a pattern missing signal is generated, an instruction to carry a coating tool is generated and sent to the staff's mobile terminal. When a stain covering signal is generated, an instruction to carry a cleaning tool is generated and sent to the staff's mobile terminal.

[0010] As a preferred embodiment of the present invention, the analysis and division process of the device analysis unit is specifically as follows:

[0011] S1: Obtain the variable focal length, resolution and installation height of the image acquisition device, which are respectively recorded as variable focal length and resolution , Indicates that the focal length range of the image acquisition device is ;

[0012] S2: There are multiple unit rectangles preset, the areas of the multiple unit rectangles are different, the integer multiples of the areas of the multiple unit rectangles are equal to the area of ​​the standard basketball court, and there is a pixel unit value preset , let the area of ​​each unit rectangle be , extract its area to satisfy the relationship All unit rectangles of , and record them as target rectangles;

[0013] S3: Get the current installation height h of the image acquisition device and set the variable focal length Substitute into the formula The calculation is performed in is the preset coefficient, and the adaptation area range of the image acquisition device is obtained , the area of ​​all target rectangles and the range of the adaptation area Compare and filter out the area values ​​within the adaptation area range Among all target rectangles within the grid, extract the target rectangle with the largest area and take it as the unit area;

[0014] S4: Divide the basketball court into unit rectangles corresponding to the unit areas, and evenly divide the entire basketball court into a positive integer number of unit areas.

[0015] As a preferred embodiment of the present invention, the image acquisition unit labels each unit area, obtains high-definition images in all unit areas through the image acquisition device, and marks the acquired images, respectively. , where p represents the acquisition date serial number and q represents the unit area number.

[0016] As a preferred embodiment of the present invention, the image comparison unit performs the following high-definition image comparison process for each unit area:

[0017] The unit area to be compared is recorded as the target area. The current high-definition image, the most recent high-definition image, and the initial high-definition image of the target area are obtained, which are recorded as image 1, image 2, and image 3, respectively. The three images are grayed to obtain the grayscale histogram corresponding to each image. The color similarity between image 1 and image 2 and image 3 is obtained by the color histogram comparison method and recorded as , the texture similarity of image 1 compared with image 2 and image 3 is obtained by texture histogram comparison method and recorded as .

[0018] As a preferred embodiment of the present invention, the process of the image analysis unit screening out the defective area is specifically as follows:

[0019] Will Substitute into the formula Calculate the similarity judgment value of the target area ,in are all preset weight factors and satisfy A judgment threshold is preset, and the similarity judgment value is compared with the judgment threshold. When the similarity judgment value is less than or equal to the judgment threshold, the unit area corresponding to the target area is recorded as the defect area to be classified.

[0020] As a preferred embodiment of the present invention, the damage analysis unit scans the surface of the defective area to be divided by structured light to construct a three-dimensional shape model of the surface of the defective area to be divided. The three-dimensional shape model is composed of multiple three-dimensional coordinate points. The coordinates of the three-dimensional coordinate points are set to (x, y, z), where x and y are plane coordinates and z is a depth coordinate. The mode of z is obtained. , statistics satisfy The number of coordinates is recorded as the number of protrusions, and the selection meets The coordinates are marked as concave coordinates, and plane analysis is performed on the concave coordinates to extract the concave coordinates of x=0 or y=0 ;

[0021] Preset a horizontal quantity judgment value and obtain All the concave coordinates of the horizontal quantity judgment value are combined. When it is greater than or equal to the horizontal quantity judgment value, All the concave coordinates are marked as seam coordinates;

[0022] Preset a vertical quantity judgment value and obtain All the concave coordinates of the vertical quantity judgment value, when it is greater than or equal to the vertical quantity judgment value, All the concave coordinates are marked as seam coordinates;

[0023] The number of depressions is obtained by subtracting the number of seam coordinates from the number of depression coordinates. The sum of the number of depressions and the number of protrusions is calculated and compared with the preset loss number threshold. When it is greater than or equal to the loss number threshold, a surface defect signal is generated and sent to the repair execution module.

[0024] As a preferred embodiment of the present invention, the pattern analysis unit analyzes the pattern defects of the defective area to be divided as follows:

[0025] P1: Obtain a high-definition image after grayscale processing corresponding to the defect area to be divided, record it as the target image, obtain the grayscale value corresponding to each pixel point of the target image, assign a coordinate to each pixel point, the coordinate corresponds to its row and column position in the high-definition image, record the surroundings of the rectangle as edge lines, and collectively refer to the pixel points on the edge lines of the rectangle as edge points;

[0026] P2: Get the coordinates of each edge point, record two edge points whose coordinate distance is less than or equal to 1 as adjacent points, each edge point has two adjacent points, get the grayscale values ​​of all edge points and count the proportion of each grayscale value to the total number of edge points, preset a ratio interval, extract all grayscale values ​​within the ratio interval, record them as target grayscale values, extract all grayscale values ​​less than the ratio interval, record them as problem grayscale values;

[0027] P3: Get the coordinates of all edge points corresponding to each target grayscale value, combine the adjacent points of each edge point on the same edge line to form a line segment composed of adjacent points, calculate the length of all line segments, preset a length threshold, and when the maximum length of all line segments is greater than the length threshold, the grayscale value corresponding to the line segment is recorded as the line grayscale value; when the length of all line segments is less than or equal to the length threshold, the grayscale value corresponding to the line segment is recorded as the problem grayscale value;

[0028] P4: A grayscale interval is established with the grayscale value of the line as the midpoint of the interval and the preset grayscale tolerance value as the width. All pixel points with grayscale values ​​within the grayscale interval are screened out and their coordinates are extracted to form a target contour set. Each defect area to be divided is preset with a corresponding standard contour set. The target contour set is compared with the standard contour set to calculate their similarity. A similarity threshold is preset. When the similarity between the target contour set and the standard contour set is less than the similarity threshold, a pattern missing signal is generated and sent to the execution repair module;

[0029] P5: Count the number of pixels whose grayscale value is equal to the problem grayscale value and remember the number of problem points. A stain number judgment value is preset. The number of problem points is compared with the stain number judgment value. When the number of problem points is greater than the stain number judgment value, a stain coverage signal is generated and sent to the repair execution module.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] 1. The image analysis unit of the present invention can reduce the error caused by oxidative fading or deepening of the floor surface color and improve the accuracy of the color comparison result by adjusting the weights of the two color similarities. By adjusting the weights of the two texture similarities, the influence of pattern wear caused by use on the texture comparison result can be reduced, thereby improving the accuracy of texture comparison, thereby more reasonably screening out floor areas that may have defects so as to more specifically find out the corresponding positions and defect types of the defects in the future, thereby improving the efficiency of defect detection.

[0032] 2. The present invention can analyze each defect area to be divided separately through the defect segmentation module. The defect segmentation module includes a damage analysis unit and a pattern analysis unit. The damage analysis unit can determine whether there is a damage gap on the floor surface through structured light technology. The pattern analysis unit can analyze whether there is pattern wear and missing, thereby making the detection defect type range wider and improving the detection effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0034] Figure 1 The principle framework of the present invention Figure 1 ;

[0035] Figure 2 The principle framework of the present invention Figure 2 ;

[0036] Figure 3 This is a schematic diagram of the basketball court after grayscale processing. DETAILED DESCRIPTION

[0037] The technical scheme of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0038] See also Figure 1 As shown, a basketball floor defect detection method includes an image acquisition module, a comparison and analysis module, a defect segmentation module and an execution repair module. The specific detection steps are as follows:

[0039] Step 1: The image acquisition module includes a device analysis unit and an image acquisition unit. The device analysis unit divides the entire basketball court into multiple rectangular unit areas according to the variable focal length, resolution and installation height of the image acquisition device. The image acquisition unit acquires high-definition images of all current unit areas.

[0040] The analysis and division process of the equipment analysis unit is as follows:

[0041] S1: Obtain the variable focal length, resolution and installation height (installation height is a variable value determined by the installation bracket of the image acquisition device) of the image acquisition device (the image acquisition device includes but is not limited to industrial cameras, high-definition fixed-focus cameras, etc.), and record them as variable focal lengths. ( Indicates that the focal length range of the image acquisition device is ) and resolution ;

[0042] It should be noted that the maximum installation height should be less than the height of the bottom edge of the backboard from the ground to avoid interference between the device and the backboard during image acquisition.

[0043] S2: There are multiple unit rectangles preset, and the areas of the multiple unit rectangles are different. The integer multiples of the areas of the multiple unit rectangles are equal to the area of ​​a standard basketball court (the standard basketball court is 28 meters long and 15 meters wide). There is a pixel unit value preset. (the area corresponding to each pixel), let the area of ​​each unit rectangle be , extract its area to satisfy the relationship All unit rectangles of , and record them as target rectangles;

[0044] It should be noted that the pixel unit value refers to the area value corresponding to each pixel point under the premise of ensuring image clarity, that is, when the area of ​​the unit rectangle meets Under the premise of high-definition images captured by the image acquisition device, the image is clear enough to have enough pixels for displaying the real situation of every square millimeter of the floor surface;

[0045] S3: Get the current installation height h of the image acquisition device and set the variable focal length Substitute into the formula The calculation is performed in is the preset coefficient, and the adaptation area range of the image acquisition device is obtained , the area of ​​all target rectangles and the range of the adaptation area Compare and filter out the area values ​​within the adaptation area range Among all target rectangles within the , extract the target rectangle with the largest area and take it as the unit area.

[0046] S4: Divide the basketball court into unit rectangles corresponding to the unit areas, and evenly divide the entire basketball court into a positive integer number of unit areas.

[0047] It should be noted that the equipment analysis unit can determine a unit area of ​​a suitable size to divide the entire basketball court, so that high-definition images of each unit area can be captured by the image acquisition device to facilitate subsequent image analysis, and the area of ​​the unit area is determined by the variable focal length, resolution and installation height of the image acquisition device, so that the high-definition image corresponding to the area size of each unit area has higher clarity, thereby improving the accuracy of the image analysis results.

[0048] The image acquisition unit labels each unit area, obtains high-definition images of all unit areas through the image acquisition device, and marks the acquired images, which are recorded as , where p represents the acquisition date serial number, and q represents the unit area number, for example, Indicates the high-definition image within the unit area numbered 1 obtained for the first time.

[0049] It should be noted that the date serial number and unit area number are both represented by consecutive positive integers, and each date serial number corresponds to a collection date.

[0050] Step 2: The comparison and analysis module performs comparison and analysis on the high-definition images of the unit area, including an image comparison unit and an image analysis unit. The image comparison unit compares the currently acquired high-definition image with the first and last acquired high-definition images respectively. The image analysis unit calculates the similarity judgment value of each area according to the comparison result and screens out the areas with defects.

[0051] The image comparison unit compares the high-definition images of each unit area in the following way:

[0052] The unit area to be compared is recorded as the target area, and the current high-definition image, the most recent high-definition image (i.e., the high-definition image acquired last time), and the initial high-definition image (i.e., the image acquired when the basketball court floor is laid) of the target area are obtained, which are recorded as image 1, image 2, and image 3, respectively. The three images are gray-scaled to obtain the grayscale histogram corresponding to each image. The color similarity between image 1 and image 2 and image 3 is obtained by the color histogram comparison method and recorded as , the texture similarity of image 1 compared with image 2 and image 3 is obtained by texture histogram comparison method and recorded as .

[0053] It should be noted that the color histogram comparison method refers to judging the similarity of two images by comparing the quantitative distribution of different grayscale values ​​of two grayscale histograms, while the texture histogram comparison method uses texture analysis techniques, such as local binary pattern (LBP), gray-level co-occurrence matrix (GLCM), Gabor filtering, etc., to extract the texture features of the image and perform similarity measurement on the texture features of the two grayscale histograms (commonly used methods include Euclidean distance, Manhattan distance, cosine similarity, etc.) to obtain the texture similarity of the two images. Both the color histogram comparison method and the texture histogram comparison method are existing technologies, and their specific comparison process will not be described in detail here.

[0054] The process of the image analysis unit screening out defective areas is as follows:

[0055] Will Substitute into the formula Calculate the similarity judgment value of the target area ,in are all preset weight factors and satisfy A judgment threshold is preset, and the similarity judgment value is compared with the judgment threshold. When the similarity judgment value is less than or equal to the judgment threshold, the unit area corresponding to the target area is recorded as the defect area to be classified.

[0056] It should be noted that the similarity judgment value reflects the degree of change in the color and texture of the corresponding unit area. The larger the similarity judgment value, the greater the degree of change in the unit area and the more defects there are. Due to long-term use, the color of the floor surface will be oxidized, faded or deepened, which will cause the grayscale value of each pixel in the target area after grayscale processing to be higher or lower than the grayscale value of the corresponding pixel in the initial high-definition image, thereby causing errors in the color histogram comparison results. This phenomenon has a smaller impact when compared with the most recently acquired high-definition image (because the time interval between the two sets of images is close, the effect of oxidation is not obvious). Therefore, by adjusting the weights of the two color similarities, the error caused by this phenomenon can be reduced and the accuracy of the color comparison results can be improved. For the judgment of the current texture of the unit area, it is obvious that the comparison result with the initial high-definition image (the initial high-definition image meets the standard site marking) is more referenceable, and when the most recently acquired high-definition image has irreversible texture changes (such as floor scratches but does not affect use), it also has a certain reference. Therefore, by adjusting the weights of the two texture similarities, the accuracy of texture comparison can be improved.

[0057] Step 3: The defect segmentation module analyzes each defect area to be segmented separately. The defect segmentation module includes a damage analysis unit and a pattern analysis unit. The damage analysis unit is used to analyze whether there is a damage gap in the high-definition image of the defect area to be segmented. The pattern analysis unit is used to analyze whether there is pattern wear and missing in the high-definition image of the defect area to be segmented.

[0058] The damage analysis unit scans the surface of the defective area to be divided by structured light to construct a three-dimensional shape model of the surface of the defective area to be divided. The three-dimensional shape model is composed of multiple three-dimensional coordinate points. Let the coordinates of the three-dimensional coordinate points be (x, y, z), where x and y are plane coordinates and z is the depth coordinate. Get the mode of z , statistics satisfy The number of coordinates is recorded as the number of protrusions, and the selection meets The coordinates are marked as concave coordinates, and plane analysis is performed on the concave coordinates to extract the concave coordinates of x=0 or y=0 ;

[0059] Structured light projection is an active optical measurement technology. Its basic principle is to project controllable light spots, light strips or light surface structures onto the surface of the object to be measured through a structured light projector to obtain an image. Then, the three-dimensional coordinates of the object are calculated through the system geometric relationship and the triangulation principle. It is a relatively mature technology and will not be described in detail here.

[0060] Preset a horizontal quantity judgment value and obtain All the concave coordinates of the horizontal quantity judgment value are combined. When it is greater than or equal to the horizontal quantity judgment value, All the concave coordinates are marked as seam coordinates;

[0061] Preset a vertical quantity judgment value and obtain All the concave coordinates of the vertical quantity judgment value, when it is greater than or equal to the vertical quantity judgment value, All the concave coordinates are marked as seam coordinates;

[0062] The number of depressions is obtained by subtracting the number of seam coordinates from the number of depression coordinates. The sum of the number of depressions and the number of protrusions is calculated and compared with the preset loss number threshold. When it is greater than or equal to the loss number threshold, a surface defect signal is generated and sent to the repair execution module.

[0063] It should be noted that when there is a floor joint in the defective area to be divided, the floor joint must be a straight line perpendicular to the edge of the defective area to be divided, so The number of concave coordinates within the straight line range of the vertical point is analyzed, and the concave coordinates corresponding to the floor joints are eliminated to reduce the influence of the calculation of the number of concave gaps on the floor surface. The greater the sum of the number of concave and convex numbers, the more convex or concave there are on the floor surface, and the more obvious the defects on the floor surface.

[0064] The pattern analysis unit analyzes the pattern defects in the defect area to be divided as follows:

[0065] P1: Obtain a high-definition image after grayscale processing corresponding to the defect area to be divided, record it as the target image, obtain the grayscale value corresponding to each pixel point of the target image, assign a coordinate to each pixel point, the coordinate corresponds to its row and column position in the high-definition image, record the surroundings of the rectangle as edge lines, and collectively refer to the pixel points on the edge lines of the rectangle as edge points;

[0066] P2: Get the coordinates of each edge point, record two edge points whose coordinate distance is less than or equal to 1 as adjacent points, each edge point has two adjacent points, get the grayscale values ​​of all edge points and count the proportion of each grayscale value to the total number of edge points, preset a ratio interval, extract all grayscale values ​​within the ratio interval, record them as target grayscale values, extract all grayscale values ​​less than the ratio interval, record them as problem grayscale values;

[0067] P3: Get the coordinates of all edge points corresponding to each target grayscale value, combine the adjacent points of each edge point on the same edge line to form a line segment composed of adjacent points (if two points are adjacent points, they are assigned to the same line segment), calculate the length of all line segments (the distance between the two farthest edge points in the line segment), preset a length threshold, when the maximum length of all line segments is greater than the length threshold, the grayscale value corresponding to the line segment is recorded as the line grayscale value, when the length of all line segments is less than or equal to the length threshold, the grayscale value corresponding to the line segment is recorded as the problem grayscale value;

[0068] It should be noted that when a basketball court is divided by multiple preset unit rectangles, when there are field markings (such as the center line, center circle, three-point line, free throw line, etc. of the basketball court) in the unit rectangle, the field markings must intersect with one or more side lines of the unit rectangle, such as Figure 3 As shown in the figure (black is the unit rectangle), the field marking marks appear as gray line segments of a certain grayscale on one or more side lines of the unit rectangle, and because the standard markings of the basketball court have a width limit, the length of the gray line segments corresponding to the field marking marks must be greater than a certain value, and then by analyzing these line segments, the field marking marks within the unit rectangle can be effectively analyzed.

[0069] P4: A grayscale interval is established with the grayscale value of the line as the midpoint of the interval and the preset grayscale tolerance value as the width. All pixel points with grayscale values ​​within the grayscale interval are screened out and their coordinates are extracted to form a target contour set. Each defect area to be divided is preset with a corresponding standard contour set (i.e., the target contour set obtained by the initial high-definition image through P1-P3). The target contour set is compared with the standard contour set to calculate their similarity (the degree of overlap of all coordinates in the two sets). A similarity threshold is preset. When the similarity between the target contour set and the standard contour set is less than the similarity threshold, a pattern missing signal is generated and sent to the execution repair module;

[0070] It should be noted that since the basketball court lines corresponding to the initial high-definition image are the clearest and most standard, they are most suitable as standard contours. By comparing the target contour with the standard contour, it can be determined whether the pattern is worn or missing. The smaller the similarity, the more serious the pattern defect.

[0071] P5: Count the number of pixels whose grayscale value is equal to the problem grayscale value and remember the number of problem points. A stain number judgment value is preset. The number of problem points is compared with the stain number judgment value. When the number of problem points is greater than the stain number judgment value, a stain coverage signal is generated and sent to the repair execution module.

[0072] It should be noted that the number of problem points can indirectly reflect the amount of stains attached. The larger the number of problem points, the more stains there are on the floor surface in the defective area to be divided.

[0073] Step 4: Execute the repair module to obtain the generated signal. When a surface defect signal is generated, an instruction to carry a repair tool is generated and sent to the staff's mobile terminal. When a pattern missing signal is generated, an instruction to carry a coating tool is generated and sent to the staff's mobile terminal. When a stain covering signal is generated, an instruction to carry a cleaning tool is generated and sent to the staff's mobile terminal.

[0074] The preferred embodiments of the present invention disclosed above are only used to help explain the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to only specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present invention, so that those skilled in the art can understand and use the present invention well. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A basketball floor defect detection method, characterized in that: The specific detection steps are as follows: Step 1: The image acquisition module is used to acquire a high-definition image of the basketball court, including a device analysis unit and an image acquisition unit. The device analysis unit divides the entire basketball court into a plurality of rectangular unit areas according to the variable focal length, resolution and installation height of the image acquisition device. The specific analysis and division process of the device analysis unit is as follows: S1: Obtain the variable focal length, resolution and installation height of the image acquisition device, which are respectively recorded as variable focal length and resolution , Indicates that the focal length range of the image acquisition device is ; S2: There are multiple unit rectangles preset, the areas of the multiple unit rectangles are different, the integer multiples of the areas of the multiple unit rectangles are equal to the area of ​​the standard basketball court, and there is a pixel unit value preset , let the area of ​​each unit rectangle be , extract its area to satisfy the relationship All unit rectangles of , and record them as target rectangles; S3: Get the current installation height h of the image acquisition device and set the variable focal length Substitute into the formula The calculation is performed in is the preset coefficient, and the adaptation area range of the image acquisition device is obtained , the area of ​​all target rectangles and the range of the adaptation area Compare and filter out the area values ​​within the adaptation area range Among all target rectangles within the grid, extract the target rectangle with the largest area and take it as the unit area; S4: Divide the basketball court into unit rectangles corresponding to the unit areas, and evenly divide the entire basketball court into a positive integer number of unit areas. The image acquisition unit obtains high-definition images of all current unit areas. Step 2: The comparison and analysis module compares and analyzes the high-definition images of the unit area, including an image comparison unit and an image analysis unit. The image comparison unit compares the currently acquired high-definition image with the first and last acquired high-definition images respectively. The image analysis unit calculates the similarity judgment value of each area according to the comparison result, and selects the area with defects as the defective area to be classified; Step 3: The defect segmentation module analyzes each defect area to be segmented separately, including a damage analysis unit and a pattern analysis unit. The damage analysis unit analyzes the high-definition image of the defect area to be segmented and generates a surface defect signal; The damage analysis unit scans the surface of the defective area to be divided by structured light to construct a three-dimensional shape model of the surface of the defective area to be divided. The three-dimensional shape model is composed of multiple three-dimensional coordinate points. The coordinates of the three-dimensional coordinate points are set to (x, y, z), where x and y are plane coordinates and z is a depth coordinate. The mode of z is obtained. , statistics satisfy The number of coordinates is recorded as the number of protrusions, and the selection meets The coordinates are marked as concave coordinates, and plane analysis is performed on the concave coordinates to extract the concave coordinates of x=0 or y=0 ; Preset a horizontal quantity judgment value and obtain All the concave coordinates of the horizontal quantity judgment value are combined. When it is greater than or equal to the horizontal quantity judgment value, All the concave coordinates are marked as seam coordinates; Preset a vertical quantity judgment value and obtain All the concave coordinates of the vertical quantity judgment value, when it is greater than or equal to the vertical quantity judgment value, All the concave coordinates are marked as seam coordinates; Calculate the number of concave coordinates minus the number of seam coordinates to obtain the number of concave, calculate the sum of the number of concave and convex and compare it with a preset loss number threshold, and when it is greater than or equal to the loss number threshold, generate a surface defect signal and send it to the execution repair module; The pattern analysis unit analyzes the high-definition image of the defect area to be divided and generates a pattern missing signal or a stain coverage signal; The pattern analysis unit analyzes the pattern defects of the defective area to be divided as follows: P1: Obtain a high-definition image after grayscale processing corresponding to the defect area to be divided, record it as the target image, obtain the grayscale value corresponding to each pixel point of the target image, assign a coordinate to each pixel point, the coordinate corresponds to its row and column position in the high-definition image, record the surroundings of the rectangle as edge lines, and collectively refer to the pixel points on the edge lines of the rectangle as edge points; P2: Get the coordinates of each edge point, record two edge points whose coordinate distance is less than or equal to 1 as adjacent points, each edge point has two adjacent points, get the grayscale values ​​of all edge points and count the proportion of each grayscale value to the total number of edge points, preset a ratio interval, extract all grayscale values ​​within the ratio interval, record them as target grayscale values, extract all grayscale values ​​less than the ratio interval, record them as problem grayscale values; P3: Get the coordinates of all edge points corresponding to each target grayscale value, combine the adjacent points of each edge point on the same edge line to form a line segment composed of adjacent points, calculate the length of all line segments, preset a length threshold, and when the maximum length of all line segments is greater than the length threshold, the grayscale value corresponding to the line segment is recorded as the line grayscale value; when the length of all line segments is less than or equal to the length threshold, the grayscale value corresponding to the line segment is recorded as the problem grayscale value; P4: A grayscale interval is established with the grayscale value of the line as the midpoint of the interval and the preset grayscale tolerance value as the width. All pixel points with grayscale values ​​within the grayscale interval are screened out and their coordinates are extracted to form a target contour set. Each defect area to be divided is preset with a corresponding standard contour set. The target contour set is compared with the standard contour set to calculate their similarity. A similarity threshold is preset. When the similarity between the target contour set and the standard contour set is less than the similarity threshold, a pattern missing signal is generated and sent to the execution repair module; P5: Count the number of pixels whose grayscale value is equal to the problem grayscale value and remember the number of problem points. A stain number judgment value is preset. The number of problem points is compared with the stain number judgment value. When the number of problem points is greater than the stain number judgment value, a stain coverage signal is generated and sent to the execution repair module. Step 4: Execute the repair module to obtain the generated signal. When a surface defect signal is generated, an instruction to carry a repair tool is generated and sent to the staff's mobile terminal. When a pattern missing signal is generated, an instruction to carry a coating tool is generated and sent to the staff's mobile terminal. When a stain covering signal is generated, an instruction to carry a cleaning tool is generated and sent to the staff's mobile terminal.

2. A basketball floor defect detection method according to claim 1, characterized in that: The image acquisition unit labels each unit area, obtains high-definition images in all unit areas through the image acquisition device, and marks the acquired images, respectively , where p represents the acquisition date serial number and q represents the unit area number.

3. A basketball floor defect detection method according to claim 1, characterized in that: The image comparison unit performs the following high-definition image comparison on each unit area: The unit area to be compared is recorded as the target area. The current high-definition image, the most recent high-definition image, and the initial high-definition image of the target area are obtained, which are recorded as image 1, image 2, and image 3, respectively. The three images are grayed to obtain the grayscale histogram corresponding to each image. The color similarity between image 1 and image 2 and image 3 is obtained by the color histogram comparison method and recorded as , the texture similarity of image 1 compared with image 2 and image 3 is obtained by texture histogram comparison method and recorded as .

4. A basketball floor defect detection method according to claim 3, characterized in that: The process of the image analysis unit screening out defective areas is specifically as follows: Will Substitute into the formula Calculate the similarity judgment value of the target area ,in are all preset weight factors and satisfy A judgment threshold is preset, and the similarity judgment value is compared with the judgment threshold. When the similarity judgment value is less than or equal to the judgment threshold, the unit area corresponding to the target area is recorded as the defect area to be classified.

Images