On-line detection method, system and equipment for bevel edge gypsum board and medium

Abstract

The invention discloses an on-line detection method, system and device for a bevel edge gypsum board and a medium, and relates to the technical field of image data processing, and the method comprises the following steps: obtaining a detection image based on the overlook angle of a target gypsum board; according to the detection image, identifying pit characteristics of the target gypsum board to obtain a pit defect value; according to the detection image, identifying crack characteristics of the target gypsum board to obtain a crack defect value; selecting a plurality of target measurement points on the bevel edge of the target gypsum board based on a preset selection rule; performing position deviation detection on each target measurement point to obtain a bevel edge inclination defect value; inputting the pit defect value, the crack defect value and the bevel edge inclination defect value into a preset defect evaluation model to obtain a defect evaluation value; whether the defect evaluation value is larger than a preset standard threshold value or not is judged, if yes, the target gypsum board is identified as a defective product, and if not, the target gypsum board is identified as a qualified product.

Description

Technical Field

[0001] The present invention relates to the technical field of image data processing, and particularly to an on-line detection method, system, device and medium for beveled gypsum boards. Background Art

[0002] To meet the filling of the gaps between boards, beveled gypsum boards have emerged on the market. The bevel angles of their bevels are designed according to process requirements. Therefore, when processing beveled gypsum boards, it is necessary to detect whether the bevel angles meet the error requirements, and at the same time, it is also necessary to detect whether there are surface defects (such as pits, cracks, etc.) on the gypsum boards.

[0003] The existing detection methods for beveled gypsum boards mainly rely on manual visual inspection or the use of simple detection tools for detection. Due to the large number of detection items, not only the detection accuracy is low, but also for the gypsum boards transported in batches on the production line, the detection efficiency is low, ultimately affecting the transfer and transportation efficiency. Summary of the Invention

[0004] The main purpose of the present invention is to provide an on-line detection method, system, device and medium for beveled gypsum boards, aiming to solve the technical problem of low detection accuracy of the existing defect detection methods for beveled gypsum boards.

[0005] To achieve the above purpose, the present invention provides an on-line detection method for beveled gypsum boards, including the following steps: Obtain a detection image based on the downward view angle of the target gypsum board; wherein, bevels are provided on all four side walls of the target gypsum board; According to the detection image, identify the pit features of the target gypsum board to obtain the pit defect value; According to the detection image, identify the crack features of the target gypsum board to obtain the crack defect value; Select a plurality of target measurement points on the bevel of the target gypsum board based on a preset selection rule; Perform position deviation detection on each target measurement point to obtain the bevel slope defect value; Input the pit defect value, the crack defect value and the bevel slope defect value into a preset defect evaluation model to obtain the defect evaluation value; Judge whether the defect evaluation value is greater than a preset standard threshold. If so, identify the target gypsum board as a defective product; if not, identify the target gypsum board as a qualified product.

[0006] Optionally, performing position deviation detection on each target measurement point to obtain the bevel slope defect value includes: Obtain the first detection distance from the first detection point to the corresponding target measurement point, and at the same time obtain the second detection distance from the second detection point to the same target measurement point; wherein, the detection direction of the first detection point is the top view direction of the target gypsum board, and the detection direction of the second detection point is the side view direction of the target gypsum board; Obtain the first deviation value between the first detection distance and the preset first theoretical distance; Obtain the second deviation value between the second detection distance and the preset second theoretical distance; Output the average value of the first deviation value and the second deviation value as the defect characteristic value corresponding to the target measurement point; Output the maximum value of the defect characteristic values corresponding to multiple target measurement points as the bevel slope defect value.

[0007] Optionally, obtaining the first detection distance from the first detection point to the corresponding target measurement point and at the same time obtaining the second detection distance from the second detection point to the same target measurement point includes: Based on laser ranging, obtain the first measurement distance from the first detection point to the corresponding target measurement point, and at the same time, based on machine vision recognition, obtain the second measurement distance from the first detection point to the corresponding target measurement point; Assign weights to the first measurement distance and the second measurement distance through confidence to obtain the first detection distance after data fusion; Based on laser ranging, obtain the third measurement distance from the second detection point to the same target measurement point, and at the same time, based on machine vision recognition, obtain the fourth measurement distance from the second detection point to the same target measurement point; Assign weights to the third measurement distance and the fourth measurement distance through confidence to obtain the second detection distance after data fusion.

[0008] Optionally, the selection rules are: Avoid the area with pit features on the bevel of the target gypsum board; Avoid the area with crack features on the bevel of the target gypsum board; Select multiple target measurement points along the length direction and the slope direction of the bevel of the target gypsum board.

[0009] Optionally, according to the detection image, identify the pit features of the target gypsum board to obtain the pit defect value, including: According to the detection image, identify and extract the pit features of the target gypsum board; wherein, the pit features include the pit area S, the pit depth H, and the number of pits N1; According to the pit features, obtain the pit defect value F, and the expression of F is: F = K1*S + K2*H + K3*N1; In the formula, K1 is the first adjustment coefficient, K2 is the second adjustment coefficient, and K3 is the third adjustment coefficient.

[0010] Optionally, based on the detected image, identify the crack characteristics of the target gypsum board to obtain a crack defect value, including: Based on the detected image, identify and extract the crack characteristics of the target gypsum board; wherein, the crack characteristics include the crack length L and the crack number N2; Based on the crack characteristics, obtain a crack defect value R, and the expression of R is: R = K4 * L + K5 * N2; In the formula, K4 is the fourth adjustment coefficient, and K5 is the fifth adjustment coefficient.

[0011] Optionally, the expression of the defect evaluation model is: Q = W1 * E + W2 * F + W3 * R; In the formula, Q is the defect evaluation value, E is the bevel slope defect value, F is the pit defect value, R is the crack defect value, W1 is the first weight value, W2 is the second weight value, and W3 is the third weight value.

[0012] To achieve the above object, the present invention also provides an on-line detection system for beveled gypsum boards, including: An image acquisition module, configured to acquire a detection image based on the top view angle of the target gypsum board; wherein, bevels are provided on all four side walls of the target gypsum board; A pit detection module, configured to identify the pit characteristics of the target gypsum board according to the detection image to obtain a pit defect value; A crack detection module, configured to identify the crack characteristics of the target gypsum board according to the detection image to obtain a crack defect value; A measurement point selection module, configured to select a plurality of target measurement points on the bevel of the target gypsum board based on a preset selection rule; A bevel detection module, configured to perform position deviation detection on each target measurement point to obtain a bevel slope defect value; A defect evaluation module, configured to input the pit defect value, the crack defect value, and the bevel slope defect value into a preset defect evaluation model to obtain a defect evaluation value; A data processing module, configured to determine whether the defect evaluation value is greater than a preset standard threshold. If so, identify the target gypsum board as a defective product; if not, identify the target gypsum board as a qualified product.

[0013] To achieve the above object, the present invention also provides a computer device, which includes a memory and a processor. A computer program is stored in the memory, and the processor executes the computer program to implement the above method.

[0014] To achieve the above object, the present invention further provides a computer-readable storage medium, on which a computer program is stored, and a processor executes the computer program to implement the above method.

[0015] The beneficial effects that can be achieved by the present invention are as follows: The detection image obtained based on the top-down angle of the target gypsum board of the present invention can display the characteristic information of the main detection areas such as the upper surface and the bevel of the gypsum board, including the pit characteristics and the crack characteristics, so as to calculate the corresponding pit defect value and the crack defect value respectively. When the bevel slope detection is performed, if the slope detection is directly performed on the entire surface of the bevel, not only is the detection difficult, but it is also difficult to accurately grasp the local deformation information by the whole surface detection, resulting in insufficient detection accuracy. Therefore, the present invention first selects a plurality of target measurement points on the bevel of the target gypsum board based on a preset selection rule, and then performs position deviation detection on each target measurement point. If there is a slope error, the position of the corresponding target measurement point will be offset relative to the theoretical position, so as to calculate the deviation according to the offset. The bevel slope defect value is calculated, and the bevel slope error can be characterized based on the defect value. The present invention uses a plurality of measuring points for position detection, which can effectively detect the local deformation of the bevel accurately, and more accurately grasp the local deformation of the bevel, and the detection difficulty is also relatively low. Finally, the pit defect value, the crack defect value and the bevel slope defect value are input into a preset defect assessment model to obtain a defect assessment value, and it is determined whether the defect assessment value is greater than a preset standard threshold. If so, the target gypsum board is identified as a defective product, and if not, the target gypsum board is identified as a qualified product, thereby realizing online defect detection of the gypsum board, and performing comprehensive detection of various defects, overcoming the problems of low accuracy and low efficiency of manual detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following is a brief introduction to the drawings required for the specific embodiments or the description of the prior art. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn according to the actual scale.

[0017] Figure 1 It is a schematic flow chart of an online detection method for beveled edge gypsum board in an embodiment of the present invention; Figure 2 It is a schematic diagram of identifying pit features and crack features based on a detection image in an embodiment of the present invention; Figure 3 It is a schematic diagram of the principle of performing position deviation detection on a target measurement point in an embodiment of the present invention.

[0018] Reference numerals: 110 - target gypsum board, 111 - bevel edge, 120 - pit feature, 130 - crack feature, 140 - target measurement point, 150 - first detection point, 160 - second detection point.

[0019] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0020] 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 only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] It should be noted that if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0022] Example 1 Reference Figures 1-3 This embodiment provides an online detection method for beveled edge gypsum board, comprising the following steps: A detection image is acquired based on a top-down perspective of the target gypsum board 110; wherein the four side walls of the target gypsum board 110 are all provided with bevel edges 111; According to the detection image, a pit feature 120 of the target gypsum board 110 is identified to obtain a pit defect value; According to the detection image, a crack feature 130 of the target gypsum board 110 is identified to obtain a crack defect value; Selecting a plurality of target measurement points 140 on the bevel 111 of the target gypsum board 110 based on a preset selection rule; Performing position deviation detection on each target measurement point 140 to obtain a slope defect value of the bevel edge; Inputting the pit defect value, the crack defect value and the bevel slope defect value into a preset defect assessment model to obtain a defect assessment value; Determine whether the defect evaluation value is greater than a preset standard threshold. If so, identify the target gypsum board 110 as a defective product. If not, identify the target gypsum board 110 as a qualified product.

[0023] In this embodiment, in the detection image obtained based on the downward view angle of the target gypsum board 110, the characteristic information of the main detection areas such as the upper surface and the hypotenuse 111 of the gypsum board can be displayed, including the pit feature 120 and the crack feature 130. Thus, the corresponding pit defect value and crack defect value can be calculated respectively. When detecting the slope of the hypotenuse 111, if directly detecting the slope of the entire surface of the hypotenuse 111, not only the detection difficulty is large, but also it is difficult to accurately grasp the local deformation information during the whole surface detection, resulting in insufficient detection accuracy. Therefore, in this embodiment, first, multiple target measurement points 140 are selected on the hypotenuse 111 of the target gypsum board 110 based on a preset selection rule. Then, position deviation detection is performed on each target measurement point 140. If there is a slope error, the position of the corresponding target measurement point 140 will deviate relative to the theoretical position. Thus, the slope defect value of the hypotenuse 111 is calculated based on the deviation situation, and the slope error of the hypotenuse 111 can be characterized laterally according to this defect value. The method of using multiple measurement points for position detection in this embodiment can effectively and accurately detect the local deformation of the hypotenuse 111, more precisely grasp the local deformation situation of the hypotenuse 111, and the detection difficulty is also relatively low. Finally, the pit defect value, the crack defect value, and the slope defect value of the hypotenuse 111 are input into a preset defect evaluation model to obtain a defect evaluation value, and it is determined whether the defect evaluation value is greater than a preset standard threshold. If so, the target gypsum board 110 is identified as a defective product. If not, the target gypsum board 110 is identified as a qualified product. Thus, the online defect detection of the gypsum board can be realized, and multiple defects can be comprehensively detected, overcoming the problems of low accuracy and low efficiency of manual detection.

[0024] It should be noted that an industrial camera and other detection instruments can be configured at the corresponding position of the conveyor rack for transporting the gypsum board, so as to realize the online image acquisition and recognition of the moving gypsum board. The industrial camera is electrically connected to an industrial control computer, and the industrial control computer can process and analyze the acquired images, and finally obtain and display the detection data for easy viewing.

[0025] As an optional implementation manner, performing position deviation detection on each target measurement point 140 to obtain the slope defect value of the hypotenuse includes: Obtain the first detection distance from the first detection point 150 to the corresponding target measurement point 140, and at the same time obtain the second detection distance from the second detection point 160 to the same target measurement point 140; wherein, the detection direction of the first detection point 150 is the downward view direction of the target gypsum board 110, and the detection direction of the second detection point 160 is the side view direction of the target gypsum board 110; Obtain the first deviation value between the first detection distance and a preset first theoretical distance; Obtain the second deviation value between the second detection distance and the preset second theoretical distance; Output the average value of the first deviation value and the second deviation value as the defect characteristic value corresponding to the target measurement point 140; Output the maximum value of the defect characteristic values corresponding to multiple target measurement points 140 as the hypotenuse slope defect value.

[0026] In this embodiment, the corresponding first detection point 150 and second detection point 160 are respectively constructed based on the top view direction and the side view direction of the target gypsum board 110, and the first detection distance h1 from the first detection point 150 to the corresponding target measurement point 140 and the second detection distance h2 from the second detection point 160 to the same target measurement point 140 are respectively calculated. Here, by obtaining detection data using two different detection directions, the risk of inaccurate calculation caused by distortion of a single detection data can be reduced, and the data accuracy is improved. If there is a slope deviation, the position of the target measurement point 140 is correspondingly shifted. At this time, the first deviation value |h1 - d1| between the first detection distance h1 and the corresponding first theoretical distance d1, and the second deviation value |h2 - d2| between the second detection distance h2 and the corresponding second theoretical distance d2 are respectively calculated. Finally, the average value of the first deviation value and the second deviation value is output as the defect characteristic value corresponding to the target measurement point 140. According to the above measurement method, the defect characteristic values corresponding to multiple target measurement points 140 are calculated, and the maximum value among them is output as the hypotenuse slope defect value, so as to effectively characterize the error situation of the hypotenuse slope.

[0027] It should be noted that after determining the positions of the first detection point 150 and the second detection point 160, corresponding detection instruments can be installed at these positions, such as industrial cameras, laser rangefinders, etc.

[0028] As an alternative embodiment, obtaining the first detection distance from the first detection point 150 to the corresponding target measurement point 140 and simultaneously obtaining the second detection distance from the second detection point 160 to the same target measurement point 140 includes: Based on laser ranging, obtain the first measurement distance from the first detection point 150 to the corresponding target measurement point 140, and simultaneously based on machine vision recognition, obtain the second measurement distance from the first detection point 150 to the corresponding target measurement point 140; Assign weights to the first measurement distance and the second measurement distance through confidence levels to obtain the first detection distance after data fusion; Based on laser ranging, obtain the third measurement distance from the second detection point 160 to the same target measurement point 140, and simultaneously based on machine vision recognition, obtain the fourth measurement distance from the second detection point 160 to the same target measurement point 140; Assign weights to the third measurement distance and the fourth measurement distance through confidence levels to obtain the second detection distance after data fusion.

[0029] In this embodiment, when calculating the first detection distance, at the first detection point 150, the first measurement distance from the first detection point 150 to the corresponding target measurement point 140 is calculated based on the laser ranging principle (i.e., using a laser rangefinder), and a depth image is collected based on the machine vision recognition principle (i.e., using a depth camera), so as to calculate the second measurement distance from the pixel point corresponding to the target measurement point 140 to the first detection point 150. For the same detection point, data is collected using two measurement instruments respectively, which can further reduce the risk of measurement data distortion or large accidental errors caused by using a single measurement instrument. Then, weights are assigned to the first measurement distance and the second measurement distance through confidence. Here, the weight values can be assigned based on the training of historical big data. For example, the weight value corresponding to the first measurement distance is 0.6, and the weight value corresponding to the second measurement distance is 0.4, so as to calculate the first detection distance after data fusion, providing an accurate data source for accurately calculating the bevel slope defect value in the subsequent process. Similarly, when calculating the second detection distance, data is also collected using the above two measurement instruments and weights are assigned according to the confidence, so as to calculate a second detection distance with higher accuracy.

[0030] It should be noted that the first detection point 150 and the second detection point 160 mainly refer to the same spacing horizontal line of the target gypsum board 110. Any point on this same spacing horizontal line can be used as the first detection point 150 or the second detection point 160. Therefore, the above two measurement instruments can be respectively set on the same spacing horizontal line. For example, when the target gypsum board 110 moves to the measurement area of the laser rangefinder to complete the measurement, it then moves to the measurement area of the depth camera for measurement. The laser rangefinder and the depth camera located on the same spacing horizontal line are at the same height position, and then the two measurement data can be integrated.

[0031] As an alternative embodiment, the selection rules are as follows: Avoid the area with the pit feature 120 on the bevel edge 111 of the target gypsum board 110; Avoid the area with the crack feature 130 on the bevel edge 111 of the target gypsum board 110; Select a plurality of target measurement points 140 along both the length direction and the slope direction of the bevel edge 111 of the target gypsum board 110.

[0032] In this embodiment, when selecting the target measurement point 140, the areas corresponding to the previously identified pit feature 120 and crack feature 130 should be avoided, so as to avoid the distortion of the detection data. At the same time, a plurality of target measurement points 140 are selected along both the length direction and the slope direction of the bevel edge 111, so as to ensure the discreteness of the target measurement points 140, so as to ensure that the measurement data can simultaneously represent the overall and local error conditions of the bevel edge 111, and improve the accuracy of the detection data.

[0033] As an alternative implementation, based on the detected image, the pit feature 120 of the target gypsum board 110 is identified to obtain the pit defect value, including: Based on the detected image, the pit feature 120 of the target gypsum board 110 is identified and extracted; wherein, the pit feature 120 includes the pit area S, the pit depth H, and the number of pits N1; Based on the pit feature 120, the pit defect value F is obtained, and the expression of F is: F = K1*S + K2*H + K3*N1; In the formula, K1 is the first adjustment coefficient, K2 is the second adjustment coefficient, and K3 is the third adjustment coefficient.

[0034] In this implementation, the detected image should be a depth image, so that the pit area S, the pit depth H, and the number of pits N1 can be identified and calculated, and then input into the expression of the pit defect value F. Considering the different attributes of the above parameters, corresponding adjustment coefficients (which may include weight coefficients) are respectively assigned, so that the three parameters with different attributes can be quantitatively superimposed. Therefore, based on the above formula, the corresponding pit defect value F can be finally quantitatively calculated. Among them, if the pit area S is too large, the pit depth H is too deep, or the number of pits N1 is too large, it will affect the surface quality of the gypsum board. Therefore, based on the above formula, it can represent the influence of a single defect manifestation form on the overall quality, and can also represent the superposition of all defect manifestation forms on the overall quality. Therefore, it can accurately represent the influence degree of the pit defect on the surface quality of the gypsum board, and the calculation is accurate and effective.

[0035] As an alternative implementation, based on the detected image, the crack feature 130 of the target gypsum board 110 is identified to obtain the crack defect value, including: Based on the detected image, the crack feature 130 of the target gypsum board 110 is identified and extracted; wherein, the crack feature 130 includes the crack length L and the number of cracks N2; Based on the crack feature 130, the crack defect value R is obtained, and the expression of R is: R = K4*L + K5*N2; In the formula, K4 is the fourth adjustment coefficient, and K5 is the fifth adjustment coefficient.

[0036] In this implementation, based on the detected image, the crack length L and the number of cracks N2 can be identified and calculated. Inputting the above formula, the crack defect value R can be quantitatively calculated. Similarly, considering the different attributes of the parameters, they are respectively transformed through the corresponding adjustment coefficients, so that the two parameters can be quantitatively superimposed, and finally the corresponding crack defect value R can be accurately calculated.

[0037] As an alternative implementation, the expression of the defect evaluation model is: Q = W1 * E + W2 * F + W3 * R; In the formula, Q is the defect evaluation value, E is the bevel slope defect value, F is the pit defect value, R is the crack defect value, W1 is the first weight value, W2 is the second weight value, and W3 is the third weight value.

[0038] In this embodiment, after calculating the pit defect value, crack defect value, and bevel slope defect value, substitute them into the above formula. Considering that different defects have different degrees of influence on the overall quality, different weight values are assigned here. According to the training of historical big data, preferably W1 = 0.4, W2 = 0.35, W3 = 0.25, and the defect evaluation value Q can be calculated quantitatively, which has strong reference and guidance.

[0039] Example 2 Based on the same inventive concept as the previous embodiment, this embodiment also provides an on-line detection system for beveled gypsum boards, including: An image acquisition module for acquiring a detection image based on the downward viewing angle of the target gypsum board 110; wherein, bevels 111 are provided on the four side walls of the target gypsum board 110; A pit detection module for identifying the pit feature 120 of the target gypsum board 110 according to the detection image to obtain the pit defect value; A crack detection module for identifying the crack feature 130 of the target gypsum board 110 according to the detection image to obtain the crack defect value; A measurement point selection module for selecting a plurality of target measurement points 140 on the bevel 111 of the target gypsum board 110 based on a preset selection rule; A bevel detection module for detecting the position deviation of each target measurement point 140 to obtain the bevel slope defect value; A defect evaluation module for inputting the pit defect value, crack defect value, and bevel slope defect value into a preset defect evaluation model to obtain the defect evaluation value; A data processing module for determining whether the defect evaluation value is greater than a preset standard threshold. If so, identify the target gypsum board 110 as a defective product; if not, identify the target gypsum board 110 as a qualified product. For the relevant explanations and examples of each module in the system of this embodiment, reference can be made to the method of the previous embodiment, which will not be elaborated here.

[0040] Example 3 Based on the same inventive concept as the previous embodiment, this embodiment provides a computer device, which includes a memory and a processor. A computer program is stored in the memory, and the processor executes the computer program to implement the above method.

[0041] Example 4 Based on the same inventive concept as the foregoing embodiments, this embodiment provides a computer-readable storage medium, on which a computer program is stored, and when the processor executes the computer program, the above-mentioned method is implemented.

[0042] The above are only the preferred embodiments of the present invention, and do not limit the patent scope of the present invention accordingly. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present invention, or directly or indirectly applied in other related technical fields, shall be equally included in the patent protection scope of the present invention.

Claims

1. An on-line detection method for bevel gypsum board, characterized in that It includes the following steps: Obtain a detection image based on the top-down view of the target gypsum board; wherein, bevels are provided on all four side walls of the target gypsum board; According to the detection image, identify the pit features of the target gypsum board to obtain the pit defect value; According to the detection image, identify the crack features of the target gypsum board to obtain the crack defect value; Select a plurality of target measurement points on the bevel of the target gypsum board based on a preset selection rule; Perform position deviation detection on each target measurement point to obtain the bevel slope defect value; Input the pit defect value, the crack defect value, and the bevel slope defect value into a preset defect evaluation model to obtain a defect evaluation value; Judge whether the defect evaluation value is greater than a preset standard threshold. If so, identify the target gypsum board as a defective product; if not, identify the target gypsum board as a qualified product.

2. The on-line detection method of a bevel gypsum board according to claim 1, characterized in that, Performing position deviation detection on each target measurement point to obtain the bevel slope defect value includes: Obtain the first detection distance from the first detection point to the corresponding target measurement point, and at the same time obtain the second detection distance from the second detection point to the same target measurement point; wherein, the detection direction of the first detection point is the top-down view direction of the target gypsum board, and the detection direction of the second detection point is the side view direction of the target gypsum board; Obtain the first deviation value between the first detection distance and a preset first theoretical distance; Obtain the second deviation value between the second detection distance and a preset second theoretical distance; Output the average value of the first deviation value and the second deviation value as the defect feature value corresponding to the target measurement point; Output the maximum value of the defect feature values corresponding to the plurality of target measurement points as the bevel slope defect value.

3. The on-line inspection method for bevel gypsum board according to claim 2, characterized in that, Obtaining the first detection distance from the first detection point to the corresponding target measurement point and at the same time obtaining the second detection distance from the second detection point to the same target measurement point includes: Obtain the first measurement distance from the first detection point to the corresponding target measurement point based on laser ranging, and at the same time obtain the second measurement distance from the first detection point to the corresponding target measurement point based on machine vision recognition; Assign weights to the first measurement distance and the second measurement distance according to confidence to obtain the first detection distance after data fusion; Obtain the third measurement distance from the second detection point to the same target measurement point based on laser ranging, and at the same time obtain the fourth measurement distance from the second detection point to the same target measurement point based on machine vision recognition; Assign weights to the third measurement distance and the fourth measurement distance according to confidence to obtain the second detection distance after data fusion.

4. The on-line inspection method for bevel gypsum board according to claim 1, characterized in that, The selection rule is: Avoid the area with pit features on the bevel of the target gypsum board; Avoid the area with crack features on the bevel of the target gypsum board; Select a plurality of target measurement points along both the length direction and the slope direction of the bevel of the target gypsum board.

5. The on-line inspection method for bevel gypsum board according to claim 1, characterized in that According to the detection image, identifying the pit features of the target gypsum board to obtain the pit defect value includes: According to the detection image, identify and extract the pit features of the target gypsum board; wherein, the pit features include the pit area S, the pit depth H, and the number of pits N1; According to the pit features, obtain the pit defect value F, and the expression of F is: F = K1*S + K2*H + K3*N1; In the formula, K1 is the first adjustment coefficient, K2 is the second adjustment coefficient, and K3 is the third adjustment coefficient.

6. The on-line inspection method of a bevel gypsum board according to claim 1, characterized in that Based on the detected image, identify the crack characteristics of the target gypsum board to obtain the crack defect value, including: Based on the detected image, identify and extract the crack characteristics of the target gypsum board; among them, the crack characteristics include the crack length L and the crack number N2; Based on the crack characteristics, obtain the crack defect value R, and the expression of R is: R = K4 * L + K5 * N2; In the formula, K4 is the fourth adjustment coefficient, and K5 is the fifth adjustment coefficient.

7. A method for on-line inspection of beveled gypsum boards according to any one of claims 1-6, characterized in that, The expression of the defect evaluation model is: Q = W1 * E + W2 * F + W3 * R; In the formula, Q is the defect evaluation value, E is the bevel slope defect value, F is the pit defect value, R is the crack defect value, W1 is the first weight value, W2 is the second weight value, and W3 is the third weight value.

8. An on-line inspection system for beveled gypsum boards, characterized in that, Including: An image acquisition module, configured to acquire a detected image based on the top view angle of the target gypsum board; among them, bevels are provided on the four side walls of the target gypsum board; A pit detection module, configured to identify the pit characteristics of the target gypsum board according to the detected image to obtain the pit defect value; A crack detection module, configured to identify the crack characteristics of the target gypsum board according to the detected image to obtain the crack defect value; A measurement point selection module, configured to select a plurality of target measurement points on the bevel of the target gypsum board based on a preset selection rule; A bevel detection module, configured to perform position deviation detection on each target measurement point to obtain the bevel slope defect value; A defect evaluation module, configured to input the pit defect value, the crack defect value, and the bevel slope defect value into a preset defect evaluation model to obtain the defect evaluation value; A data processing module, configured to determine whether the defect evaluation value is greater than a preset standard threshold. If so, identify the target gypsum board as a defective product; if not, identify the target gypsum board as a qualified product.

9. A computer device, characterized in that, The computer device includes a memory and a processor. A computer program is stored in the memory, and the processor executes the computer program to implement an on-line detection method for a bevel gypsum board according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, A computer program is stored on the computer-readable storage medium, and the processor executes the computer program to implement an on-line detection method for a bevel gypsum board according to any one of claims 1-7.

Images