Machining monitoring method for automobile gearbox shell based on machine vision

Abstract

The invention relates to the technical field of image data processing, in particular to an automobile gearbox shell machining monitoring method based on machine vision. The method comprises the following steps: acquiring an outer surface image of a to-be-detected automobile gearbox shell, and determining a noise probability value of a target pixel point in the outer surface image; dividing pixel points which have pixel values smaller than a preset threshold value and are adjacent to each other in the outer surface image into the same candidate area so as to obtain a plurality of candidate areas in the outer surface image; determining a defect probability value of the target pixel point according to a position relationship between the target pixel point and the candidate region; according to the defect probability value and the noise probability value of the pixel point, carrying out filtering processing on the pixel point to obtain a target image; and processing of the automobile gearbox shell is monitored according to the target image. By means of the technical scheme, monitoring of the machining process of the automobile gearbox shell can be better achieved.

Description

Technical Field

[0001] This application relates to the technical field of image data processing, and particularly to a method for monitoring the machining of an automotive transmission housing based on machine vision. Background Art

[0002] The automotive transmission housing is a component of the automotive transmission system and is used to install the transmission mechanism and its accessories; in addition to providing an installation space for components such as the gears, shafts, and bearings of the transmission, the transmission housing can also keep the key components such as the load-bearing gears, shafts, and bearings of the transmission in precise relative positions during high-speed operation, ensuring that the components of the transmission can remain stable during high-speed operation.

[0003] During the manufacturing process of the automotive transmission housing, the production of the transmission housing can be achieved through multiple processes such as casting, cutting, and grinding. The surface of the produced automotive transmission housing may have defects such as cracks, pores, and sand holes.

[0004] The defects existing on the surface of the automotive transmission housing can reflect the defects existing inside the automotive transmission housing. When there are defects on the surface of the automotive transmission housing, it will not only affect the structural strength of the automotive transmission housing, but may also cause the relative positions of the key components such as the load-bearing gears, shafts, and bearings of the transmission carried by the automotive transmission housing to be in an unbalanced state during high-speed operation, thus threatening the driving safety of the vehicle.

[0005] Detecting the completed automotive transmission housing during the processing stage can not only prevent defective automotive transmission housings from being installed in the vehicle, thus ensuring the driving safety of users, but also facilitate the production manufacturer to optimize the production process or technology of the automotive transmission housing and ensure the quality of the produced automotive transmission housing. Therefore, it is necessary to monitor the machining process of the automotive transmission housing. Summary of the Invention

[0006] To monitor the machining process of an automotive transmission housing, the present application provides a method for monitoring the machining of an automotive transmission housing based on machine vision, including: obtaining an outer surface image of the automotive transmission housing to be detected, and determining the noise probability value of a target pixel point in the outer surface image, where the noise probability value is used to characterize the degree of difference in pixel values of the pixel points in the row or column where the target pixel point is located; the target pixel point is any pixel point in the outer surface image; dividing the pixel points in the outer surface image whose pixel values are less than a preset threshold and are adjacent to each other into the same candidate region to obtain multiple candidate regions in the outer surface image; determining the defect probability value of the target pixel point according to the positional relationship between the target pixel point and the candidate region; determining the target filtering side length of the target pixel point according to the defect probability value and the noise probability value of the target pixel point, and performing filtering processing on the target pixel point according to the target filtering side length to obtain a filtered target image; determining the defect detection result of the automotive transmission housing to be detected according to the target image, so as to monitor the machining of the automotive transmission housing by using the defect detection result.

[0007] In this way, it is possible to better monitor the machining process of the automotive transmission housing, thereby ensuring the quality of the obtained automotive transmission housing.

[0008] Optionally, the noise probability value of the target pixel point is determined by the following method: , where N is the noise probability value of the target pixel point, norm is the normalization function, is the number of pixel points in the row where the target pixel point is located, and are the pixel values of the b-th and the (b - 1)-th pixel points in the row where the target pixel point is located respectively, L is the number of pixel points in the column where the target pixel point is located, and are the pixel values of the a-th and the (a - 1)-th pixel points in the column where the target pixel point is located respectively; P is the neighborhood difference value of the target pixel point, which is used to characterize the degree of difference in gray values between the target pixel point and the pixel points within the neighborhood range.

[0009] In this way, it is possible to comprehensively determine the noise probability value of the target pixel point from three dimensions: row, column, and neighborhood, so that the noise probability value can better characterize the probability that the target pixel point belongs to noise.

[0010] Optionally, the neighborhood difference value of the target pixel point is determined by the following method: , P is the neighborhood difference value of the target pixel point, M is the number of pixel points within the neighborhood range of the target pixel point, Q is the set composed of the pixel values of the pixel points within the neighborhood range of the target pixel point, is the pixel value of the pixel point located in the i-th row and the j-th column within the neighborhood range of the target pixel point, is the average value of the pixel values of the pixels within the neighborhood range of the target pixel point, and exp is the exponential function with the natural constant as the base; is the distance between the pixel at the i-th row and the j-th column within the neighborhood range of the target pixel point and the target pixel point.

[0011] In this way, by comparing the pixel value of the pixel within the neighborhood range of the target pixel point with the average value of the pixel values of the pixels within the neighborhood range of the target pixel point, the neighborhood difference value can better represent the probability that the target pixel point belongs to a noise pixel point.

[0012] Optionally, determining the defect probability value of the target pixel point according to the positional relationship between the target pixel point and the candidate region includes: when the target pixel point is within the candidate region, determining the defect probability value of the target pixel point according to the range difference of the pixel values of the neighborhood pixels of the outer edge pixels of the candidate region where the target pixel point is located; when the target pixel point is outside the candidate region, determining the absolute value of the difference between the preset threshold and the pixel value of the target pixel point, and taking the normalized result of the absolute value as the defect probability value of the target pixel point.

[0013] In this way, according to whether the target pixel point is within the candidate region, the defect probability value of the target pixel point can be adaptively determined to use the defect probability value to represent the probability that the target pixel point belongs to a defective pixel point.

[0014] Optionally, determining the defect probability value of the target pixel point according to the range difference of the pixel values of the neighborhood pixels of the outer edge pixels of the candidate region where the target pixel point is located includes: , where is the defect probability value of the target pixel point, is the normalization processing function, T is the pixel value of the target pixel point, t is the preset threshold, is the number of pixels on the outer edge of the candidate region where the target pixel point is located, is the range difference of the pixel values of the neighborhood pixels of the y-th pixel on the outer edge of the candidate region where the target pixel point is located.

[0015] In this way, when the target pixel point is within the candidate region, the defect probability value of the target pixel point can comprehensively consider the situation of the target pixel point itself and the situation of the candidate region where it is located.

[0016] Optionally, determining the target filtering side length of the target pixel point according to the defect probability value and the noise probability value of the target pixel point includes: , where is the target filtering side length of the target pixel point, is the noise probability value of the target pixel point, is the defect probability value of the target pixel point, is the initial filtering side length.

[0017] Optionally, determining the defect detection result of the automotive transmission housing to be detected according to the target image, including: using the Otsu threshold to determine the target threshold of the target image, and taking the pixel points with gray values less than the target threshold in the target image as defect pixel points to obtain a defect area composed of the defect pixel points; outputting the defect detection result of the automotive transmission housing to be detected according to the position and area of the defect area in the target image.

[0018] Optionally, determining the defect detection result of the automotive transmission housing to be detected according to the target image, including: inputting the target image into a pre-trained defect detection model to obtain the defect detection result of the automotive transmission housing in the target image output by the defect detection model; the defect detection model is used to output the defect detection result of the automotive transmission housing in the input image.

[0019] Optionally, the defect detection model is obtained by training in the following manner: using the sample pictures of the automotive transmission housing as the input of a pre-constructed network model, and using the defect detection labels corresponding to the sample pictures as the output of the network model, and training the network model to obtain the defect detection model.

[0020] Optionally, the defect detection result includes the position where the defect is located and the area of the defect; using the defect detection result to monitor the processing of the automotive transmission housing, including: when the area of the defect is greater than a preset area threshold, moving the automotive transmission housing to be detected to the space where the defect handler is located, and sending the position where the defect is located to the terminal device bound to the defect handler.

[0021] The technical solutions provided by the embodiments of the present application may include the following beneficial effects: obtaining the outer surface image of the automotive transmission housing to be detected, and adaptively determining the side length when filtering the target pixel points according to the situation of the target pixel points in the outer surface image, and performing filtering processing with a side length adapted to the target pixel points, which can avoid the possible noise pixel points in the outer surface image. Therefore, the obtained target image can better reflect the surface defect situation of the automotive transmission housing to be detected, and the processing process of the automotive transmission housing can be better monitored using the target image.

[0022] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a flowchart of a method for monitoring the processing of an automotive transmission housing based on machine vision shown according to an exemplary embodiment. Detailed implementation manners

[0024] First, a brief introduction to the application scenario of the embodiments of the present application is given. In the application scenario of the present application, there may be defects such as cracks, pores, and sand holes on the surface of the automotive transmission housing. To ensure the quality of the produced automotive transmission housing, it is necessary to monitor the processing process of the automotive transmission housing.

[0025] In view of the above technical problems, the embodiments of the present application provide a method for monitoring the processing of an automotive transmission housing based on machine vision. Figure 1 It is a flowchart of a method for monitoring the processing of an automotive transmission housing based on machine vision shown according to an exemplary embodiment. As Figure 1 shown, the method includes the following steps.

[0026] In step S101, an outer surface image of the automotive transmission housing to be detected is acquired, and the noise probability value of the target pixel point in the outer surface image is determined.

[0027] An outer surface image of the automotive transmission housing to be detected can be acquired by using an image acquisition device; in order to avoid the influence of the possible background in the image, the pixel values of the pixel points of other parts except the automotive transmission housing in the outer surface image can be equal to 0.

[0028] In order to reduce the amount of calculation required in the subsequent calculation process, the outer surface image of the automotive transmission housing to be detected can be a grayscale image after grayscale processing.

[0029] The noise probability value is used to characterize the degree of difference in pixel values of the pixel points in the row or column where the target pixel point is located; the target pixel point is any pixel point in the outer surface image.

[0030] There may be some noise pixel points in the outer surface image of the automotive transmission housing. In order to avoid the influence of the possible noise pixel points in the outer surface image of the automotive transmission housing on the detection result, the noise probability value of the target pixel point in the outer surface image can be determined.

[0031] The noise probability value is used to characterize the degree of difference in pixel values of the pixel points in the row or column where the target pixel point is located. Since the noise pixel points have a certain randomness and isolation, and the possible defects in the outer surface image of the automotive transmission housing are usually concentrated in one or more regions, therefore, the degree of difference in the grayscale values of the pixel points in the row where the noise pixel point is located is greater, or the degree of difference in the grayscale values of the pixel points in the column where the noise pixel point is located is greater. Through the noise probability value of the target pixel point, the probability that the target pixel point belongs to noise can be characterized.

[0032] Referring to the determination process of the noise probability value of the target pixel, the noise probability value of each pixel of the outer surface image of the automotive transmission housing can be obtained respectively.

[0033] In one embodiment, the noise probability value of the target pixel is determined in the following manner: , where N is the noise probability value of the target pixel, norm is the normalization function, is the number of pixels in the row where the target pixel is located, and are the pixel values of the b-th and the (b - 1)-th pixels in the row where the target pixel is located respectively, L is the number of pixels in the column where the target pixel is located, and are the pixel values of the a-th and the (a - 1)-th pixels in the column where the target pixel is located respectively; P is the neighborhood difference value of the target pixel, which is used to characterize the difference degree of the gray value between the target pixel and the pixels in the neighborhood range.

[0034] Due to the isolation of noise pixels, compared with the pixel row or pixel column without noise pixels, the difference degree of the gray values of the pixels in the pixel row or pixel column with noise pixels is greater. Therefore, by comparing the pixel values of the pixels in the row where the target pixel is located and comparing the pixel values of the pixels in the column where the target pixel is located, the probability of the existence of noise pixels in the row or column where the target pixel is located can be characterized.

[0035] For example, when the probability of the existence of noise pixels in the row where the target pixel is located is relatively large and the probability of the existence of noise pixels in the column where the target pixel is located is relatively large, the probability that the target pixel is a noise pixel is relatively large. Therefore, by comparing the pixel values of the adjacent pixels in the column where the target pixel is located and comparing the pixel values of the adjacent pixels in the row where the target pixel is located, the obtained noise probability value can better distinguish between noise pixels and non-noise pixels, and the noise probability value can better characterize the probability that the target pixel belongs to a noise pixel.

[0036] When the target pixel is a non-noise pixel, since the defects existing on the outer surface of the automotive transmission housing usually do not appear as single points but as defect regions with a certain concentration, if the difference degree between the target pixel and its surrounding pixels is relatively large, it indicates that the target pixel has a relatively high probability of being generated by noise data points.

[0037] In this way, by considering the neighborhood difference value of the target pixel, the noise probability value of the target pixel is obtained, and the noise probability value of the target pixel can be comprehensively determined from three dimensions: row, column, and neighborhood, so that the noise probability value can better represent the probability that the target pixel belongs to noise.

[0038] In one embodiment, the neighborhood difference value of the target pixel is determined in the following manner: , where P is the neighborhood difference value of the target pixel, M is the number of pixels in the neighborhood range of the target pixel, and Q is the set composed of the pixel values of the pixels in the neighborhood range of the target pixel. is the pixel value of the pixel located in the i-th row and the j-th column within the neighborhood range of the target pixel. is the mean value of the pixel values of the pixels in the neighborhood range of the target pixel, and exp is the exponential function with the natural constant as the base. is the distance between the pixel located in the i-th row and the j-th column within the neighborhood range of the target pixel and the target pixel.

[0039] The size of the neighborhood range can be set according to actual needs. For example, the neighborhood range can be a range of sizes such as 3×3, 5×5, or 7×7.

[0040] The distance between the pixel in the neighborhood range of the target pixel and the target pixel can be determined by the Euclidean distance or the Manhattan distance between two pixels. The specific algorithm for distance calculation in the embodiments of the present application is not limited.

[0041] For the first pixel and the second pixel among the pixels other than the target pixel in the neighborhood range of the target pixel, the distance from the first pixel to the target pixel is greater than the distance from the second pixel to the target pixel, so that the probability that the first pixel is associated with the target pixel is less than the probability that the second pixel is associated with the target pixel; or, the correlation between the first pixel and the target pixel is lower than the correlation between the second pixel and the target pixel.

[0042] Compared with the distance from the second pixel to the target pixel, since the distance from the first pixel to the target pixel is greater, in the calculation formula of the neighborhood difference value of the target pixel, the contribution of the first pixel to the value of the neighborhood difference value of the target pixel is smaller, so that the target pixel with a higher correlation with the target pixel contributes more to the neighborhood difference value.

[0043] Compared with the pixel values of two non - adjacent position points on the outer surface of the automotive transmission housing, the influence degree or correlation between the pixel values of two adjacent position points on the outer surface of the automotive transmission housing is greater. Therefore, by determining the neighborhood difference value of the target pixel point according to the distance from the pixel points within the neighborhood range to the target pixel point in the outer surface image, the degree of influence of the target pixel point by noise can be better determined.

[0044] Compare the pixel values of the pixel points within the neighborhood range of the target pixel point with the average value of the pixel values of the pixel points within the neighborhood range of the target pixel point. The greater the difference between the pixel values of the pixel points within the neighborhood range of the target pixel point and the average value of the pixel values of the pixel points within the neighborhood range of the target pixel point, the greater the probability that there are noise pixel points within the neighborhood range where the target pixel point is located, and the greater the probability that the target pixel point belongs to a noise pixel point.

[0045] In this way, by comparing the pixel values of the pixel points within the neighborhood range of the target pixel point with the average value of the pixel values of the pixel points within the neighborhood range of the target pixel point, and considering the distance from the pixel points within the neighborhood range of the target pixel point to the target pixel point, the obtained neighborhood difference value can better characterize the probability that there are noise pixel points within the neighborhood range where the target pixel point is located, and thus better characterize the probability that the target pixel point belongs to a noise pixel point through the neighborhood difference value.

[0046] In step S102, the pixel points on the outer surface image whose pixel values are less than the preset threshold and are adjacent to each other are divided into the same candidate region to obtain multiple candidate regions in the outer surface image; according to the positional relationship between the target pixel point and the candidate region, the defect probability value of the target pixel point is determined.

[0047] Since the defect regions existing on the outer surface of the automotive transmission housing have a certain degree of aggregation, the pixel values between the pixel points with defects and adjacent to each other in the automotive transmission housing have a certain degree of similarity.

[0048] The defects on the outer surface of the automotive transmission housing are mainly manifested as cracks, pores or sand holes. Therefore, compared with the non - defect position points on the outer surface of the automotive transmission housing, the brightness of the defect position points on the outer surface of the automotive transmission housing is lower.

[0049] Since the brightness of the defect position points on the outer surface of the automotive transmission housing is lower, and the pixel values between the pixel points with defects and adjacent to each other in the automotive transmission housing have a certain degree of similarity, therefore, the defect regions on the outer surface of the automotive transmission housing are manifested as having smaller pixel values and a certain degree of aggregation in the outer surface image.

[0050] Dividing the pixel points in the outer surface image whose pixel values are less than a preset threshold and are adjacent to each other into the same candidate region can achieve a preliminary screening of the image region of the defective region on the surface that may correspond to the automotive transmission housing in the outer surface image, so as to narrow the search range for the defective region.

[0051] Multiple candidate regions in the outer surface image may be scattered in multiple independent image regions in the outer surface image. The multiple candidate regions may respectively correspond to defective regions at different positions on the outer surface of the automotive transmission housing. To further determine a more accurate defective region, the defective probability value of the target pixel point can be determined according to the positional relationship between the target pixel point and the candidate region.

[0052] For example, if the target pixel point is within the candidate region, the target pixel point is more likely to correspond to a defective pixel point on the surface of the automotive transmission housing. Therefore, according to the positional relationship between the target pixel point and the candidate region, the defective probability value of the target pixel point can be determined to more accurately determine the surface defect of the automotive transmission housing.

[0053] In one embodiment, determining the defective probability value of the target pixel point according to the positional relationship between the target pixel point and the candidate region includes: when the target pixel point is within the candidate region, determining the defective probability value of the target pixel point according to the range of the pixel values of the neighborhood pixel points of the outer edge pixel points of the candidate region where the target pixel point is located; when the target pixel point is outside the candidate region, determining the absolute value of the difference between the preset threshold and the pixel value of the target pixel point, and taking the normalization result of the absolute value as the defective probability value of the target pixel point.

[0054] When there is a defective region on the surface of the automotive transmission housing, there is a certain difference in the thickness at the outer edge line and the thickness between the two sides of the outer edge line. When the target pixel point is within the candidate region, for the outer edge pixel points of the candidate region where the target pixel point is located, the greater the range of the pixel values of the neighborhood pixel points of these outer edge pixel points, the greater the probability that the position point on the outer surface of the automotive transmission housing is actually at the edge of the defective region, and the greater the probability that the target pixel point within the candidate region belongs to a defective pixel point.

[0055] At the same time, the edges in the defective region existing on the outer surface of the automotive transmission housing are usually more obvious. Paying attention to the neighborhood pixel points of the outer edge pixel points of the candidate region helps to more keenly discover the early defects existing on the outer surface of the automotive transmission housing.

[0056] When the target pixel is outside the candidate region, it indicates that the number of other position points with similar pixel values within a certain range around the position point corresponding to the target pixel on the outer surface of the automotive transmission housing is small; and since the candidate region includes pixels with pixel values less than a preset threshold and adjacent to each other, when the target pixel is outside the candidate region, the pixel value of the target pixel has a high probability of being greater than the preset threshold, and the probability that the target pixel belongs to the defective region is smaller.

[0057] In this way, when the target pixel is within the candidate region, by determining the defect probability value of the target pixel according to the range of pixel values of the neighborhood pixels of the outer edge pixels of the candidate region where the target pixel is located, the defect probability value can better represent the probability that the target pixel belongs to the defective pixel.

[0058] In one embodiment, determining the defect probability value of the target pixel according to the range of pixel values of the neighborhood pixels of the outer edge pixels of the candidate region where the target pixel is located includes: , where is the defect probability value of the target pixel, is the normalization processing function, T is the pixel value of the target pixel, t is the preset threshold, is the number of pixels on the outer edge of the candidate region where the target pixel is located, is the range of pixel values of the neighborhood pixels of the y-th pixel on the outer edge of the candidate region where the target pixel is located.

[0059] The normalization processing function is used to normalize the variable to be normalized within the range of 0 to 1; when the target pixel is within the candidate region, the range of pixel values of the neighborhood pixels of the y-th pixel on the outer edge of the candidate region where the target pixel is located can reflect the probability that the region corresponding to the candidate region on the surface of the automotive transmission housing actually has a defect, and through the difference between the target pixel and the preset threshold, the probability that the target pixel has a defect on the surface of the automotive transmission housing can be reflected. Therefore, the defect probability value can combine the situation of the target pixel itself with the situation of the candidate region where the target pixel is located.

[0060] In this way, when the target pixel is within the candidate region, the defect probability value of the target pixel can comprehensively consider the situation of the target pixel itself and the situation of the candidate region where the target pixel is located. Therefore, the defect probability value can better represent the probability that the target pixel belongs to the defective pixel.

[0061] In step S103, according to the defect probability value and the noise probability value of the target pixel, determine the target filtering side length of the target pixel, and perform filtering processing on the target pixel according to the target filtering side length to obtain the filtered target image.

[0062] When performing filtering processing on pixel points at different filtering side lengths, the considered range is different. For example, the larger the filtering side length used for filtering, the better the filtering effect on noise pixel points; the larger the filtering side length used for filtering, the better the retention effect on the features of pixel points. Therefore, according to the defect probability value and noise probability value of the target pixel point, the target filtering side length of the target pixel point can be adaptively determined.

[0063] In one embodiment, determining the target filtering side length of the target pixel point according to the defect probability value and noise probability value of the target pixel point includes: , where is the target filtering side length of the target pixel point, is the noise probability value of the target pixel point, is the defect probability value of the target pixel point, is the initial filtering side length.

[0064] In this way, since the larger the defect probability value of the target pixel point, the greater the probability that the position point corresponding to the target pixel point on the surface of the automotive transmission housing actually has a defect; the larger the noise probability value of the target pixel point, the greater the probability that the position point corresponding to the target pixel point on the surface of the automotive transmission housing belongs to a noise pixel point. Therefore, by combining the defect probability value and noise probability value of the target pixel point, the target filtering side length that better matches the target pixel point can be determined.

[0065] The filtering processing performed on the target pixel point can be Gaussian filtering, mean filtering, median filtering, etc. The embodiments of the present application do not limit the filtering algorithm used, and those skilled in the art can make an adaptive selection of the filtering algorithm according to actual needs.

[0066] In step S104, determine the defect detection result of the automotive transmission housing to be detected according to the target image, so as to monitor the processing of the automotive transmission housing by using the defect detection result.

[0067] After performing filtering processing on the pixel points in the outer surface image, and the filtering side lengths of different pixel points are determined according to the actual situation of the pixel points, the pixel values of the obtained pixel points can avoid the influence of noise pixel points, and the pixel values of the pixel points in the target image can better reflect the defect characteristics of the surface of the automotive transmission housing.

[0068] In one embodiment, determining a defect detection result for an automotive transmission housing to be detected based on a target image includes: using the Otsu threshold to determine a target threshold for the target image, and taking the pixel points in the target image with gray values less than the target threshold as defect pixel points to obtain a defect area composed of the defect pixel points; outputting a defect detection result for the automotive transmission housing to be detected according to the position and area of the defect area in the target image.

[0069] The Otsu method can adaptively select an optimal threshold according to the statistical characteristics of image data. Therefore, using the Otsu method to determine the target threshold of the target image improves the objectivity and accuracy of detection; it can facilitate the detection process of the outer surface defects of the automotive transmission housing to adapt to different detection lighting conditions.

[0070] Since the pixel values of the pixel points in the target image can better reflect the probability of defects existing at the surface position points of the automotive transmission housing, when the pixel value of a pixel point in the target image is less than the target threshold, taking the pixel points in the target image with pixel values less than the preset threshold as defect pixel points can simply and effectively screen out the defect areas existing on the surface of the automotive transmission housing.

[0071] According to the position and area of the defect area in the target image, a defect image with one or more defect areas distributed in the outer surface image can be output; or, one or more defect areas can be marked in the outer surface image to obtain a marked image, and the marked image with defects can be output so that the user can process the automotive transmission housing with defects.

[0072] Since the position where defects exist in the automotive transmission housing to be detected is output, when the user processes the defects, the time or labor intensity required for visual search for defects can be reduced. Therefore, the processing efficiency or experience of the automotive transmission housing with defects can be improved.

[0073] In one embodiment, determining a defect detection result for an automotive transmission housing to be detected based on a target image includes: inputting the target image into a pre-trained defect detection model to obtain a defect detection result for the automotive transmission housing in the target image output by the defect detection model; the defect detection model is used to output a defect detection result for the automotive transmission housing in the input image.

[0074] Since different defects existing on the surface of the automotive transmission housing have different characteristics, for example, there are differences in the area characteristics or distribution characteristics of defects such as cracks, pores, and sand holes that may exist on the surface of the automotive transmission housing, and the target image is obtained after adaptive filtering processing of the pixel points in the outer surface image, and the possible existing defects can be more clearly reflected in the target image. Therefore, inputting the target image into a pre-trained defect detection model can obtain a more accurate defect detection result.

[0075] The model structure of the defect detection model can be, for example, a deep residual network and an FCN (Fully Convolutional Network), etc. The embodiments of the present application do not limit the model structure of the defect detection model.

[0076] The defect detection result can be used to indicate one or more types of defects existing on the outer surface of the automotive transmission housing, so that the user can perform targeted processing on the defective automotive transmission housing according to the types of defects existing on the outer surface of the automotive transmission housing.

[0077] For example, for the slight scratches existing on the surface of the automotive transmission housing, the user can polish or buff the position where the slight scratches are located on the surface of the automotive transmission housing.

[0078] In one embodiment, the defect detection model can be obtained through the following method: taking the sample pictures of the automotive transmission housing as the input of a pre-constructed network model, and taking the defect detection labels corresponding to the sample pictures as the output of the network model, and training the network model to obtain the defect detection model.

[0079] According to the different conditions of the outer surface of the automotive transmission housing, a matching defect detection label can be determined for the sample pictures of the outer surface of the automotive transmission housing, and the defect detection label is used to prompt at least one type of defect existing in the sample pictures.

[0080] When the number of training times of the network model reaches a preset number of times, or the accuracy rate of the defect detection labels output by the network model reaches a preset accuracy rate, it can be determined that the training of the network model is completed, so as to use the trained network model as the defect detection model; wherein, the network model can be, for example, a neural network model such as a convolutional neural network or a deep residual network.

[0081] In one embodiment, the defect detection result includes the location of the defect and the area of the defect; monitoring the machining of the automotive transmission housing by using the defect detection result includes: when the area of the defect is greater than a preset area threshold, moving the automotive transmission housing to be detected to the space where the defect handler is located, and sending the location of the defect to the terminal device bound to the defect handler.

[0082] When the defect detection result indicates that the automotive transmission housing to be detected has a defect, a prompt message can be output, and the prompt message is used to prompt the user that the automotive transmission housing to be detected has a defect, so as to facilitate the user to process the automotive transmission housing with a defect.

[0083] Alternatively, according to the defect situation of the automotive transmission housing with a defect, adjust the machining process of the automotive transmission housing based on machine vision; for example, the temperature or speed of different processes of the automotive transmission housing based on machine vision can be adjusted, which will not be elaborated in this embodiment of the present application.

[0084] In one embodiment, the defect detection result includes the location of the defect and the area of the defect; monitoring the machining of the automotive transmission housing by using the defect detection result includes: when the area of the defect is greater than a preset area threshold, moving the automotive transmission housing to be detected to the space where the defect handler is located, and sending the location of the defect to the terminal device bound to the defect handler.

[0085] The preset area threshold can be determined according to the total area of the surface to be detected of the automotive transmission housing to be detected; for example, the preset area threshold can be between 2% and 5% of the total area of the surface to be detected.

[0086] Moving the automotive transmission housing to be detected to the space where the defect handler is located and sending the location of the defect to the terminal device bound to the defect handler can enable the defect handler to focus on processing the automotive transmission housing with a defect in the space where they are located, improve the processing efficiency of the automotive transmission housing with a defect, and can reduce the moving distance required by the defect handler, and can minimize the unnecessary physical consumption of the defect handler.

[0087] After considering the specification and practicing the invention disclosed herein, those skilled in the art will readily conceive of other embodiments of the present application. The present application is intended to cover any variations, uses, or adaptations of the present application, which follow the general principles of the present application and include the common general knowledge or conventional technical means in the technical field not disclosed in the present application. The specification and embodiments are only regarded as exemplary.

[0088] It should be understood that the present application is not limited to the exact structures described above and shown in the drawings, and various modifications and changes can be made without departing from its scope.

Claims

1. A method for monitoring the machining of an automotive transmission housing based on machine vision, characterized in that, Including: Obtain the outer surface image of the automotive transmission housing to be detected, and determine the noise probability value of the target pixel point in the outer surface image; The noise probability value is used to characterize the degree of difference in pixel values of the pixel points in the row or column where the target pixel point is located; the target pixel point is any pixel point in the outer surface image; Divide the pixel points in the outer surface image whose pixel values are less than the preset threshold and are adjacent to each other into the same candidate region to obtain multiple candidate regions in the outer surface image; Determine the defect probability value of the target pixel point according to the positional relationship between the target pixel point and the candidate region; Determine the target filtering side length of the target pixel point according to the defect probability value and the noise probability value of the target pixel point, and perform filtering processing on the target pixel point according to the target filtering side length to obtain the filtered target image; Determine the defect detection result of the automotive transmission housing to be detected according to the target image, so as to monitor the processing of the automotive transmission housing by using the defect detection result.

2. The method for monitoring the machining of an automotive transmission housing based on machine vision according to claim 1, characterized in that, The noise probability value of the target pixel point is determined by the following method: , where N is the noise probability value of the target pixel point, norm is the normalization function, is the number of pixel points in the row where the target pixel point is located, and are the pixel values of the b-th and (b - 1)-th pixel points in the row where the target pixel point is located respectively, L is the number of pixel points in the column where the target pixel point is located, and are the pixel values of the a-th and (a - 1)-th pixel points in the column where the target pixel point is located respectively; P is the neighborhood difference value of the target pixel point, which is used to characterize the difference degree of the gray values between the target pixel point and the pixel points within the neighborhood range.

3. The method for monitoring the machining of an automotive transmission housing based on machine vision according to claim 2, wherein, The neighborhood difference value of the target pixel point is determined by the following method: , P is the neighborhood difference value of the target pixel point, M is the number of pixel points within the neighborhood range of the target pixel point, and Q is the set composed of the pixel values of the pixel points within the neighborhood range of the target pixel point. is the pixel value of the pixel point located in the i-th row and the j-th column within the neighborhood range of the target pixel point. is the mean value of the pixel values of the pixel points within the neighborhood range of the target pixel point, and exp is the exponential function with the natural constant as the base. is the distance between the pixel point located in the i-th row and the j-th column within the neighborhood range of the target pixel point and the target pixel point.

4. The method for monitoring the machining of an automotive transmission housing based on machine vision according to claim 1, wherein, Determine the defect probability value of the target pixel point according to the positional relationship between the target pixel point and the candidate region, including: When the target pixel point is within the candidate region, determine the defect probability value of the target pixel point according to the range of the pixel values of the neighborhood pixel points of the outer edge pixel points of the candidate region where the target pixel point is located; When the target pixel point is outside the candidate region, determine the absolute value of the difference between the preset threshold and the pixel value of the target pixel point, and use the normalization result of the absolute value as the defect probability value of the target pixel point.

5. The method for monitoring the machining of an automotive transmission housing based on machine vision according to claim 4, wherein, Determine the defect probability value of the target pixel point according to the range of the pixel values of the neighborhood pixel points of the outer edge pixel points of the candidate region where the target pixel point is located, including: , where is the defect probability value of the target pixel point, is the normalization function, T is the pixel value of the target pixel point, and t is the preset threshold, is the number of pixel points on the outer edge of the candidate region where the target pixel point is located, is the range of the pixel values of the neighborhood pixel points of the y-th pixel point on the outer edge of the candidate region where the target pixel point is located.

6. The method for monitoring the machining of an automotive transmission housing based on machine vision according to claim 1, characterized in that, Determine the target filtering side length of the target pixel point according to the defect probability value and the noise probability value of the target pixel point, including: , where is the target filtering side length of the target pixel point, is the noise probability value of the target pixel point, is the defect probability value of the target pixel point, is the initial filtering side length.

7. The method for monitoring the machining of an automotive transmission housing based on machine vision according to claim 1, wherein, Determine the defect detection result of the automotive transmission housing to be detected according to the target image, including: Use the Otsu threshold to determine the target threshold of the target image, and use the pixel points in the target image whose gray values are less than the target threshold as defect pixel points to obtain the defect region composed of the defect pixel points; Output the defect detection result of the automotive transmission housing to be detected according to the position and area of the defect region in the target image.

8. The method for monitoring the machining of an automotive transmission housing based on machine vision according to claim 1, characterized in that, Determine the defect detection result of the automotive transmission housing to be detected according to the target image, including: Input the target image into a pre-trained defect detection model to obtain the defect detection result of the automotive transmission housing in the target image output by the defect detection model; the defect detection model is used to output the defect detection result of the automotive transmission housing in the input image.

9. The method for monitoring the machining of an automotive transmission housing based on machine vision according to claim 8, wherein The defect detection model is obtained by training in the following way: Use the sample picture of the automotive transmission housing as the input of a pre-constructed network model, and use the defect detection label corresponding to the sample picture as the output of the network model to train the network model to obtain the defect detection model.

10. The method for monitoring the machining of an automotive transmission housing based on machine vision according to claim 1, wherein, The defect detection result includes the location of the defect and the area of the defect; Monitoring the machining of an automotive transmission housing using defect detection results, including: When the area of the defect is greater than a preset area threshold, moving the automotive transmission housing to be detected to the space where the defect handler is located, and sending the position where the defect is located to the terminal device bound to the defect handler.

Images