A method for part feature recognition in vertical machining centers

Abstract

The present application relates to the field of image processing technology, and in particular to a method for part feature recognition in a vertical machining center. The method includes: obtaining an image set of the same flange produced by the vertical machining center in a top-down perspective; obtaining feature points of the image in the image set, and taking the feature points belonging to the edge pixel points as edge feature points, so as to determine the image recognition difficulty value according to the number of edge feature points; for an image pair consisting of any two images in the image set, feature matching is performed on the two images in the image pair, and the matching degree value of the image pair is determined according to the number of matching feature points and the recognition difficulty value of the image in the image pair; when the average value of the matching degree values ​​of different image pairs is less than a first preset threshold value, outputting a prompt first information. Through the above technical scheme, the accuracy of identifying possible processing defects of the flange can be improved.

Description

Technical Field

[0001] The present application relates to the field of image processing technology, and in particular to a method for identifying part features of a vertical machining center. Background Art

[0002] A vertical machining center is a common CNC machine tool, which is mainly used for milling, drilling and tapping of metal or other hard materials. For example, a vertical machining center can be used to process metal components such as flanges. Flanges are pipe fittings used for pipe connections, usually made of metal materials such as stainless steel, carbon steel, aluminum or titanium.

[0003] A flange is a detachable connector. Bolt holes are usually provided on the flange body so that two pipes, valves, pumps or other equipment can be connected by bolts and gaskets to ensure the sealing of the pipeline system and the stability of the connection.

[0004] In order to detect possible defects of metal parts such as flanges during the production process, a method for detecting defects of metal components is provided in the related art, for example, in a Chinese patent application document with publication number CN114705824A, the method includes: obtaining a scanned image of the metal component, graying the scanned image and performing Gaussian filtering to obtain a grayed image; performing a watershed transformation on the grayed image by a watershed algorithm to obtain each watershed in the grayed image as an edge line, and segmenting the grayed image into a plurality of irregular segmented images as a segmented image set by the edge lines; performing a natural defect adsorption process on the scanned image by the segmented image set to obtain a detection prototype image marked with a non-defective area; searching for a defect position in the detection prototype image; and performing pixel filling on the corresponding position of the defect position in the scanned image to obtain a defect image.

[0005] When performing defect detection on metal components in the related art, it relies on detection prototype images marked with non-defective areas, but flanges without defects may be quite diverse. It is difficult to accurately recognize the features of defects that may exist in the flanges through detection prototype images marked with non-defective areas. Therefore, it is difficult to accurately detect defects in the flanges. Summary of the invention

[0006] In order to overcome the problem that it is difficult to accurately detect defects in flanges in related technologies, the present application provides a part feature recognition method for a vertical machining center, including: obtaining an image set of the same flange produced by the vertical machining center in a top-down perspective; the interval of the rotation angle of the flange in the image set is equal to an integer multiple of the minimum angle that makes the flange rotationally symmetrical; obtaining feature points of the images in the image set, and taking the feature points belonging to edge pixels as edge feature points, so as to determine the recognition difficulty value of the image according to the number of edge feature points; for an image pair consisting of any two images in the image set, performing feature matching on the two images in the image pair, and determining the matching degree value of the image pair according to the number of matched feature points and the recognition difficulty value of the images in the image pair; when the average value of the matching degree values ​​of different image pairs is less than a first preset threshold value, outputting a first prompt information; the first prompt information is used to prompt that there is a processing defect in the flange; wherein the matching degree value , is the number of matching feature points in either image of the image pair, and are the number of feature points of different images in the image pair, exp is an exponential function with a natural constant as the base, as well as They are the recognition difficulty values ​​of different images in the image pair.

[0007] In this way, a set of images of the same flange produced by a vertical machining center in a top-down perspective is obtained; the interval of the rotation angle of the flange in the image set is equal to an integer multiple of the minimum angle that makes the flange rotationally symmetrical; for an image pair consisting of any two images in the image set, the matching degree value of the image pair is determined according to the number of matched feature points and the recognition difficulty value of the image in the image pair, and when the average value of the matching degree values ​​of different image pairs is less than a first preset threshold, a first prompt information is output, which can realize the identification of the processing defects of the flange according to the rotational symmetry of the flange, avoid the detection process from relying on the detection prototype image marked with non-defective areas, and can improve the accuracy of identifying possible processing defects of the flange.

[0008] Optionally, the recognition difficulty value of the image is determined by: , where D is the recognition difficulty value of the image, N is the number of edge lines in the image, and norm is the normalization function. is the number of feature points located on the i-th edge line among the feature points of the image determined by the feature detection algorithm, is the variance of the smoothness values ​​of all flange edge lines in the image.

[0009] In this way, the recognition difficulty value obtained by the variance of the smoothness values ​​of all edge lines of the flange in the image and the number of feature points located on the edge lines among the feature points of the image determined by the feature detection algorithm can better characterize the difficulty of using the image to identify defects in the flange in the subsequent process of matching the flange image.

[0010] Optionally, the smoothness value of the edge line is determined by: ,in, is the smoothness value of the target edge line in the image, exp is an exponential function with a natural constant as the base, norm() is the normalization function, is the range of curvature of all points on the target edge line, is the variance of the sine value of the gradient direction angle of all points on the target edge line. The smoothness value is used to characterize the smoothness of the curvature of the pixel points on the edge line.

[0011] In this way, through the extreme difference of the curvature of all points on the target edge line and the variance of the sine value of the gradient direction angle of all points on the target edge line, the smoothness of the edge line can be evaluated from the two dimensions of the shape characteristics of the edge line and the direction of change of the grayscale value, so as to characterize the smoothness of the image through the smoothness value.

[0012] Optionally, the image set is acquired in the following manner: when the flange is located at an initial position in a capture screen of an image acquisition device, an image of the bolt holes of the flange in a top-down perspective is acquired through the image acquisition device; when the acquisition of a previous image of the bolt holes of the flange in a top-down perspective is completed, the flange is controlled to rotate along the rotational symmetry center of the flange, and the rotation angle is equal to the minimum angle that makes the flange rotationally symmetrical, so as to obtain a subsequent image of the bolt holes of the flange in a top-down perspective; when the number of images of the bolt holes of the flange in a top-down perspective acquired reaches a preset number, the acquired images of the bolt holes of the flange in a top-down perspective are combined into the image set.

[0013] Optionally, the image set is acquired in the following manner: when the flange is located at an initial position in a capture screen of an image acquisition device, an image of the flange in a top-down perspective is acquired through the image acquisition device; when the acquisition of the previous image of the flange in a top-down perspective is completed, the flange is controlled to rotate along the rotational symmetry center of the flange, and the rotation angle is equal to the minimum angle that makes the flange rotationally symmetrical, so as to obtain a subsequent image of the flange in a top-down perspective; when the number of images of the flange in a top-down perspective acquired reaches a preset number, the acquired images of the flange in a top-down perspective are combined into the image set.

[0014] In this way, different images of the bolt holes of the flange from a top-down perspective can not only make it easier to find possible defects in the flange, but also further reduce the amount of calculation required for matching between images compared to obtaining an image of the entire flange area.

[0015] Optionally, the method also includes: for a target image in the image set, taking the average value of the matching degree between the target image and other images in the image set except the target image as the characteristic value of the target image; when the characteristic value of the target image is less than or equal to a second preset threshold, outputting a second prompt information; the second prompt information is used to prompt that there is a defect in the flange at the position corresponding to the target image.

[0016] In this way, when the characteristic value of the target image is less than or equal to the second preset threshold, by outputting the second prompt information for prompting that there is a defect in the flange at the position corresponding to the target image, in addition to letting the user know that there is a defect in the flange, the user can further know the location of the defect in the flange, so that the user can process the defective flange or the defective location in the flange.

[0017] Optionally, the method also includes: when the average value of the matching degree values ​​of different image pairs is greater than or equal to a first preset threshold, determining that the inspection result of the flange is qualified; when it is determined that the inspection result of the flange is qualified, controlling the flange moving mechanism to move the flange to an accommodating space, wherein the accommodating space is used to accommodate the flange whose inspection result is qualified.

[0018] Optionally, an image pair consisting of any two images in an image set is determined in the following manner: taking any image in the image set as the first image, and selecting any image from other images in the image set except the first image that has not formed an image pair with the first image as the second image; and forming the image pair with the first image and the second image.

[0019] The technical solution provided by the embodiments of the present application may include the following beneficial effects: obtaining an image set of the same flange produced by a vertical machining center in a top-down perspective; the interval of the rotation angle of the flange in the image set is equal to an integer multiple of the minimum angle that makes the flange rotationally symmetrical; for an image pair consisting of any two images in the image set, determining the matching degree value of the image pair based on the number of matching feature points and the recognition difficulty value of the image in the image pair, and outputting a first prompt information when the average value of the matching degree values ​​of different image pairs is less than a first preset threshold value, which can realize the identification of the processing defects of the flange based on the rotational symmetry of the flange, avoid the detection process from relying on the detection prototype image marked with non-defective areas, and can improve the accuracy of identifying possible processing defects of the flange.

[0020] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0022] Figure 1 is a flow chart of a method for identifying part features of a vertical machining center according to an exemplary embodiment;

[0023] Figure 2 is a schematic diagram of a rotation angle of a flange according to an exemplary embodiment. DETAILED DESCRIPTION

[0024] Here, exemplary embodiments are described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application.

[0025] First, a brief introduction to the application scenarios of the embodiments of the present application is given. In the application scenarios of the present application, the produced flanges and other metal components are inspected through machine vision, which can identify possible processing defects in the flanges and other metal components, thereby reducing the labor intensity of inspectors in detecting possible processing defects in the metal components.

[0026] However, in the related art, the detection of possible defects in metal components is mainly achieved through the detection prototype image marked with non-defective areas, while the flanges without defects may be more diverse. For example, in order to increase the friction between the flange and the seal, there may be some rotationally symmetrical patterns on the surface of the flange, making it difficult to accurately identify the possible defects in the metal components through the detection prototype image marked with non-defective areas.

[0027] In view of the above technical problems, the present application provides a method for identifying part features of a vertical machining center. Figure 1 is a flow chart of a method for identifying part features of a vertical machining center according to an exemplary embodiment. Figure 1 As shown, the method includes the following steps.

[0028] In step S101, a set of images of the same flange produced by a vertical machining center in a top view is obtained.

[0029] The interval of the rotation angle of the flange in the image set is equal to an integer multiple of the minimum angle that makes the flange rotationally symmetric; since the flange has rotational symmetry, the minimum angle that makes the flange rotationally symmetric can be used as the rotation angle.

[0030] Figure 2 is a schematic diagram of the rotation angle of the flange in the embodiment of the present application, such as Figure 2 As shown, a plurality of rotationally symmetrical bolt holes are provided on the flange; Figure 2 There are 8 bolt holes on the flange, and the rotation angle of the flange is equal to 360° / 8=45°.

[0031] For example, when the minimum angle that makes the flange rotationally symmetric is equal to 45 degrees, the rotation angles corresponding to different images in the flange image set are 0°, 45°, 90°, 135°, 180°, 225°, 270° and 315° respectively.

[0032] Under normal circumstances, after the flange is rotated along the rotational symmetry center of the flange according to an integer multiple of the rotation angle, the shape of the flange and the position of the bolt holes before and after the rotation are consistent. By utilizing the rotational symmetry of the flange and acquiring images of the flange from a top-down perspective at different rotation angles, it is possible to facilitate the detection of processing defects in the flange.

[0033] In one embodiment, the image set is acquired in the following manner: when the flange is located at the initial position in the capture screen of the image acquisition device, an image of the flange in a top-down perspective is acquired by the image acquisition device; when the acquisition of the previous image of the flange in a top-down perspective is completed, the flange is controlled to rotate along the rotational symmetry center of the flange, and the rotation angle is equal to the minimum angle that makes the flange rotationally symmetrical, so as to obtain a subsequent image of the flange in a top-down perspective; when the number of images of the flange in a top-down perspective acquired reaches a preset number, the acquired images of the flange in a top-down perspective are combined into the image set.

[0034] When the previous image of the flange in a top-down perspective is acquired, the flange can be controlled to rotate along the rotational symmetry center of the flange through the rotating mechanism of the vertical machining center, for example. The rotation angle is equal to the minimum angle that makes the flange rotationally symmetrical. When the rotation of the flange is completed, the next image of the flange in a top-down perspective is acquired.

[0035] In adjacent images of the flange acquired, the interval between the rotation angles of the flange is equal to the rotation angle of the flange, and the shape of the flange at different rotation angles and the position of the bolt holes remain consistent, which can minimize the difference between different images of the flange, thereby reducing the difficulty and calculation amount of matching between images, and facilitating the identification of possible defects in the flange.

[0036] The preset number of images of the top-view perspective can be set according to actual needs. For example, it can be equal to the number of times the flange rotates along the rotational symmetry center according to the rotation angle and returns to the initial position.

[0037] In this way, since the bolt holes arranged in the flange have rotational symmetry, by rotating the flange along the rotational symmetry center according to the rotation angle and capturing images from a top-down perspective after each rotation, it is possible to easily reflect the defects of the flange in different images of the flange, reduce the amount of calculation required for defect detection of the flange, and improve the accuracy of defect detection of the flange.

[0038] In another embodiment, the image set is acquired in the following manner: when the flange is located at an initial position in a capture screen of an image acquisition device, an image of the bolt holes of the flange in a top-down perspective in the capture screen is acquired by the image acquisition device; when the acquisition of a previous image of the bolt holes of the flange in a top-down perspective is completed, the flange is controlled to rotate along the rotational symmetry center of the flange, and the rotation angle is equal to the minimum angle that makes the flange rotationally symmetrical, so as to obtain a subsequent image of the bolt holes of the flange in a top-down perspective; when the number of images of the bolt holes of the flange in a top-down perspective acquired reaches a preset number, the acquired images of the bolt holes of the flange in a top-down perspective are combined into the image set.

[0039] For example, the position of the image acquisition device is fixed, and the position of the image acquisition device enables the image acquisition to capture an image including at least one bolt hole of the flange; when the acquisition of the previous image is completed, the position of the image acquisition device is kept unchanged, the flange is controlled to rotate along the rotational symmetry center of the flange according to the rotation angle, and the image acquisition device is used to capture the next image. Each rotation of the flange can be performed clockwise or counterclockwise.

[0040] In the captured image, the bolt holes of the flange in the top-down perspective image may include at least one bolt hole in the flange, and since the rotation angle of the flange each time during image capture is equal to the minimum angle that makes the flange rotationally symmetrical, the position of at least one bolt hole included in different images is the same.

[0041] Furthermore, the illumination intensity and light angle at the same position are highly consistent. By fixing the position of the image acquisition device and rotating the flange along the center of rotational symmetry to achieve the minimum angle of rotational symmetry of the flange, the illumination conditions and light angles of at least one bolt hole of the flange can be kept consistent at different rotation angles, thereby avoiding adverse effects on the image of the flange due to different lighting conditions or light angles.

[0042] In this way, different images of the bolt holes of the flange from a top-down perspective can not only make it easier to find possible defects in the flange, but also further reduce the amount of calculation required for matching between images compared to obtaining an image of the entire flange area.

[0043] In step S102, feature points of the images in the image set are obtained, and feature points belonging to edge pixels are used as edge feature points, so as to determine the recognition difficulty value of the image according to the number of edge feature points.

[0044] The image recognition difficulty value is used to characterize the difficulty of identifying defects that may exist in the image. The larger the image recognition difficulty value, the more difficult it is to determine the defects from the image; conversely, the smaller the image recognition difficulty value, the lower the difficulty of determining the defects from the image.

[0045] The recognition difficulty value of an image can be determined based on the number of feature points of the image determined by a feature detection algorithm in the image, and the number of edge pixels in the image.

[0046] Feature detection algorithms may include SIFT (Scale-Invariant Feature Transform), SURF (Speeded Up Robust Features), ORB (Oriented FAST and Rotated BRIEF), and Harris corner detection algorithms.

[0047] In one embodiment, the recognition difficulty value of the image is determined by: , where D is the recognition difficulty value of the image, N is the number of edge lines in the image, and norm is the normalization function. is the number of feature points located on the i-th edge line among the feature points of the image determined by the feature detection algorithm, is the variance of the smoothness values ​​of all flange edge lines in the image. The smoothness value is used to characterize the smoothness of the curvature of the pixel points on the edge line.

[0048] When the feature detection algorithm is used to detect feature points in an image, the more edge pixels in the obtained feature points are, the richer the features contained in the edge of the image are, and the easier it is to identify the features of the flange image.

[0049] The smoothness value is used to characterize the smoothness of the curvature of the pixel points on the edge line. The smoothness value can reflect the smoothness characteristics of the shape of the edge line in the flange image. The higher the smoothness value of the edge line, the smoother the shape of the edge line in the flange image, or the shape change degree of the same edge line at different positions is more consistent.

[0050] For possible defects in the flange, compared with the differences in edge shape between different edges of the flange itself, the shape changes of the edge lines of the defects at different positions are more inconsistent, making the difference between the edge shape of the defect and the edge shape of the flange itself greater. Therefore, the variance of the smoothness values ​​of all edge lines of the flange in the image can characterize the differences in the smoothness of the curvature of the pixel points on different edge lines in the image.

[0051] When comparing different images of the flange, compared with comparing an image of the flange without defects with an image of the flange with defects, when comparing an image of the flange with defects with other images of the flange with defects, due to the lower contrast between the images of the flange with defects, it is more difficult to successfully determine the defects in the image. Therefore, the greater the variance of the smoothness values ​​of all edge lines of the flange in the image, the more difficult it is to identify the defects using an image with a larger variance of the smoothness values ​​of the edge lines in the subsequent image comparison process.

[0052] On the contrary, the smaller the variance of the smoothness values ​​of all edge lines of the flange in the image, the better the image with a smaller variance of the smoothness value can serve as a reference image or the image of the flange itself when comparing images of the flange at different rotation angles, thereby better identifying whether there are processing defects in the image of the flange. Therefore, the smaller the variance of the smoothness values ​​of all edge lines of the flange in the image, the lower the difficulty of using the image to identify defects in the flange in the subsequent image comparison process.

[0053] In this way, the recognition difficulty value obtained by the variance of the smoothness values ​​of all edge lines of the flange in the image and the number of feature points located on the edge lines among the feature points of the image determined by the feature detection algorithm can better characterize the difficulty of using the image to identify defects in the flange in the subsequent process of matching the flange image.

[0054] In one embodiment, the smoothness value of the edge line is determined by: ,in, is the smoothness value of the target edge line in the image, exp is an exponential function with a natural constant as the base, norm() is the normalization function, is the range of curvature of all points on the target edge line, It is the variance of the sine value of the gradient direction angle of all points on the target edge line.

[0055] During the flange machining process, the disc of the flange body is mainly machined by the rotation of the metal blank and the cutting of the tool set in the vertical machining center, and the fixed disc is drilled by a rotating drill to obtain the bolt holes in the flange. Therefore, in the obtained flange, the curvature of the position points in the edge line of the flange image has a high consistency.

[0056] The target edge line can be any one of the multiple edge lines in the image of the flange. The curvature at the pixel point on the target edge line can reflect the degree of directional change of the pixel point on the target edge line. The extreme difference of the curvature of all points on the target edge line can reflect the difference in the degree of directional change of different pixel points on the target edge line. For edge lines with different shape characteristics, the difference in the degree of directional change of the pixel points on the edge line is different. Therefore, the extreme difference of the curvature of all points on the target edge line can reflect the shape characteristics of the target edge line.

[0057] The larger the range of the curvature of all points on the target edge line, the lower the consistency between the curvatures of the pixels on the target edge line, and the lower the smoothness of the target edge line; conversely, the smaller the range of the curvature of all points on the target edge line, the higher the consistency between the curvatures of the pixels on the target edge line, and the higher the smoothness of the target edge line.

[0058] The gradient direction angle of a pixel point on the edge line of the image may be equal to the angle between the grayscale gradient direction of the pixel point and the horizontal direction; the grayscale gradient direction of the pixel point may be determined using a gradient operator such as a Sobel operator or a Prewitt operator.

[0059] The gradient direction angle of all points on the target edge line can reflect the direction of change of the grayscale value of the pixel points on the target edge line, and the variance of the sine value of the gradient direction angle of all points on the target edge line can reflect the degree of difference in the direction of change of the grayscale value of different pixel points on the same edge line of the flange.

[0060] In this way, through the extreme difference of the curvature of all points on the target edge line and the variance of the sine value of the gradient direction angle of all points on the target edge line, the smoothness of the edge line can be evaluated from the two dimensions of the shape characteristics of the edge line and the direction of change of the grayscale value, so as to characterize the smoothness of the image through the smoothness value.

[0061] In step S103, for an image pair consisting of any two images in the image set, feature matching is performed on the two images in the image pair, and the matching degree value of the image pair is determined according to the number of matched feature points and the recognition difficulty value of the images in the image pair.

[0062] When there are no defects in the flange produced by the machining center, the consistency between different parts of the flange with rotational symmetry is high, so that when the images of the same flange at different rotation angles are matched for features, the images of the same flange at different rotation angles have high consistency and the matching degree value of the image pair is high.

[0063] When defects exist in the flange produced by the machining center, the consistency between different parts of the flange with rotational symmetry is reduced, so that when feature matching is performed on images of the same flange at different rotation angles, the consistency of the images of the same flange at different rotation angles is reduced, and the matching degree value of the image pair is reduced.

[0064] When matching the images included in the image pair, due to the different recognition difficulty values ​​of different images, the probabilities of identifying defects in the flange by matching the images are different. When the matching degree values ​​between different images of the same flange are different, it means that the features of the same flange in different images are different. Therefore, the matching degree value of the image pair can be better determined based on the number of matched feature points and the recognition difficulty values ​​of the images in the image pair.

[0065] In one embodiment, the matching degree value , is the number of matching feature points in either image of the image pair, and are the number of feature points of different images in the image pair, exp is an exponential function with a natural constant as the base, as well as They are the recognition difficulty values ​​of different images in the image pair.

[0066] In addition to being able to identify feature points in an image, feature detection algorithms can also match feature points obtained in different images, thereby achieving matching between different images and feature matching between two images in an image pair. This can be achieved, for example, through algorithms such as SIFT (Scale-Invariant Feature Transform), SURF (Speeded Up Robust Features), and ORB (Oriented FAST and Rotated BRIEF).

[0067] When performing feature matching on two images in an image pair, in addition to the matched feature points, there may be some other feature points that have not been matched. Therefore, the larger the proportion of matched feature points in all feature points, the higher the degree of matching between the two images in the image pair. Therefore, by determining the proportion of matched feature points in the two images in the image pair, the degree of matching between the image pairs can be reflected.

[0068] Since the recognition difficulty value of an image is determined based on the number of feature points located at the edge of the image, the recognition difficulty value of an image can reflect the amount of information of the feature points in the image. The greater the difference in the recognition difficulty values ​​between the two images in the image pair, the greater the difference in the amount of information of the feature points between the two images in the image pair. Since the flange has rotational symmetry, the amount of information of the feature points of a flange without defects corresponding to different rotation angles or different bolt holes has a high consistency. Therefore, the greater the difference in the recognition difficulty values ​​between the two images in the image pair, the more likely it is that the image pair includes an image that can identify defects in the flange, or the more likely it is that the image pair includes a defective area in the flange.

[0069] In this way, the recognition difficulty values ​​of the two images in the image pair are compared, and the proportion of matching feature points in the feature points of the two images is determined. The obtained matching degree value can better characterize the matching degree between the two images in the image pair.

[0070] In one embodiment, an image pair formed by any two images in an image set is determined in the following manner: any image in the image set is used as the first image, and any image that has not formed an image pair with the first image is selected from other images in the image set except the first image as the second image; the first image and the second image form the image pair.

[0071] For example, when the image set includes 8 images of a flange, namely P1, P2, P3, P4, P5, P6, P7 and P8, when selecting the first image, if P1 is selected as the first image, then any image that does not form an image pair with image P1 can be selected from the 7 images, P2, P3, P4, P5, P6, P7 and P8, as the second image to obtain an image pair consisting of the first image and the second image.

[0072] In this way, the two images included in the image pair can be selected regularly from the image set while exhausting all image combinations including the two images, avoiding repeated selection or omission of image pairs, and ensuring the effective implementation of the flange image matching process.

[0073] In step S104, when the average value of the matching degree values ​​of different image pairs is less than a first preset threshold, a first prompt information is output.

[0074] The smaller the matching degree value of the image pair, the more likely it is that the image pair includes an image that can identify defects in the flange, or the more likely it is that the image pair includes an area where defects exist in the flange. Therefore, the average value of the matching degree values ​​of different image pairs can comprehensively reflect the probability of defects existing in the flange.

[0075] The average value of the matching degree values ​​of different image pairs is less than the first preset threshold, indicating that there is a defect in the flange. By outputting the first prompt information, the first prompt information can be used to prompt that there is a processing defect in the flange, so that the user can understand the defects in the flange in time so as to deal with the defective flange.

[0076] The first prompt information may be text prompt information, audio prompt information or vibration prompt information, etc. The first preset threshold may be set according to actual needs. For example, after normalizing the matching degree value, the first preset threshold may be set between 0.3 and 0.4.

[0077] In one embodiment, for a target image in an image set, the average value of the matching degree between the target image and other images in the image set except the target image can be used as a characteristic value of the target image; when the characteristic value of the target image is less than or equal to a second preset threshold, a second prompt information is output; the second prompt information is used to prompt that there is a defect in the flange at the position corresponding to the target image.

[0078] The target image can be any one of the multiple images in the flange image set. The target image and different other images in the image set except the target image can correspond to different matching degree values ​​respectively. The average value of the matching degree values ​​between the target image and other images in the image set except the target image can reflect the difference between the target image and other images.

[0079] Since different images of the same flange in the image set correspond to different rotation angles of the flange, these different images may correspond to the characteristics of different rotation directions of the flange under the same lighting conditions, and the difficulty of identifying the flange under different rotation directions is different; or, these different images may correspond to the characteristics of different bolt holes of the flange under the same lighting conditions, and the probability of defects in the local areas where different bolt holes are located is different.

[0080] The average value of the matching degree between the target image and other images in the image set except the target image is taken as the characteristic value of the target image. The characteristic value can characterize the probability of the flange defect existing in the target image.

[0081] Since the eigenvalue of the target image can reflect the difference between other images of the same flange, and since different images of a flange without defects have high consistency, the defects that may exist in the flange reduce the original consistency. Therefore, the smaller the eigenvalue of the target image, the greater the probability of identifying defects in the flange at the rotation angle of the target image.

[0082] When the characteristic value of the target image is less than or equal to the second preset threshold, it means that the defect of the flange can be identified at the rotation angle of the target image; while in other images other than the target image of the flange, the defect in the flange cannot be identified due to different lighting conditions.

[0083] When there are 8 rotationally symmetrical bolt holes in the flange, the bolt holes in the flange can be numbered W1, W2, W3, W4, W5, W6, W7 and W8 in clockwise order from the top; after the flange is rotated 45 degrees clockwise, as the bolt hole W8 is rotated to the top, the bolt holes in the flange are W8, W1, W2, W3, W4, W5, W6 and W7 in clockwise order from the top. The lighting conditions in different areas of the flange may change after the flange is rotated.

[0084] For example, when the rotation angles of the flange relative to the initial position are 0°, 45°, 90°, 135°, 180°, 225°, 270° and 315° respectively, the rotation angle of the flange in the target image can be 90°, for example, indicating that the flange can be determined to be defective when the rotation angle relative to the initial position is 90°, and the bolt holes in the flange are W7, W8, W1, W2, W3, W4, W5 and W6 in clockwise order from the top. The flange can be identified as defective when the bolt holes are W7, W8, W1, W2, W3, W4, W5 and W6 in clockwise order from the top.

[0085] When the captured image of the flange is the area where at least one bolt hole in the flange is located, it means that there is a defect in the area where at least one bolt hole with a corresponding rotation angle of 90° is located; for example, when the image is captured for the area where the topmost bolt hole of the flange is located, it can be determined that there is a defect in the area where the bolt hole W7 at the top of the flange is located.

[0086] In this way, when the characteristic value of the target image is less than or equal to the second preset threshold, by outputting the second prompt information for prompting that there is a defect in the flange at the position corresponding to the target image, in addition to letting the user know that there is a defect in the flange, the user can further know the location of the defect in the flange, so that the user can process the defective flange or the defective location in the flange.

[0087] In one embodiment, the inspection result of the flange can also be determined to be qualified when the average value of the matching degree values ​​of different image pairs is greater than or equal to a first preset threshold value; when it is determined that the inspection result of the flange is qualified, the flange moving mechanism is controlled to move the flange to the accommodating space, and the accommodating space is used to accommodate the flange with a qualified inspection result.

[0088] Since different images of the flange correspond to different rotation angles, when the average value of the matching degree values ​​of different image pairs is greater than or equal to the first preset threshold, it means that after matching the images of the flange at different rotation angles, the consistency of the flange is high and the probability of defects in the flange is low. It can be determined that the inspection result of the flange is qualified, so that the next processing step can be carried out on the flange with a qualified inspection result.

[0089] The flange moving mechanism may be a robot with a mechanical arm, or the flange moving mechanism may be a conveying device such as a conveyor belt for transporting the flange.

[0090] In this way, when the average value of the matching degree values ​​of different image pairs is greater than or equal to the first preset threshold, the flange moving mechanism is controlled to move the flange to the accommodating space, so that the flanges that have passed the inspection can be sorted, so as to execute the next process for the flanges that have passed the inspection, and it is also convenient to continue to inspect the next flange to be inspected.

[0091] Through the part feature recognition method for a vertical machining center provided in an embodiment of the present application, a set of images of the same flange produced by the vertical machining center in a top-down perspective is obtained; the interval of the rotation angle of the flange in the image set is equal to an integer multiple of the minimum angle that makes the flange rotationally symmetrical; for an image pair consisting of any two images in the image set, the matching degree value of the image pair is determined according to the number of matching feature points and the recognition difficulty value of the image in the image pair, and when the average value of the matching degree values ​​of different image pairs is less than a first preset threshold, a first prompt information is output, which can realize the recognition of the processing defects of the flange according to the rotational symmetry of the flange, avoid the detection process from relying on the detection prototype image marked with non-defective areas, and can improve the accuracy of identifying possible processing defects of the flange.

[0092] It should be understood that the features of some embodiments of the various applications described herein may be combined with each other, unless specifically stated otherwise.

[0093] Although terms such as "first", "second", and "third" may be used herein to describe various components, parts, regions, layers, or sections, these components, parts, regions, layers, or sections are not limited to these terms. Instead, these terms are only used to distinguish one component, part, region, layer, or section from another component, part, region, layer, or section. Therefore, without departing from the teachings of the examples described herein, the first component, part, region, layer, or section mentioned in the examples may also be referred to as the second component, part, region, layer, or section.

[0094] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying 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 the description herein, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0095] In addition, the word "exemplary" is used herein to mean serving as an example, instance, or diagram. Any aspect or design described herein as "exemplary" is not necessarily to be understood as being advantageous over other aspects or designs. Instead, the use of the word exemplary is intended to present concepts in a concrete manner. As used herein, the term "or" is intended to mean an inclusive "or" rather than an exclusive "or".

[0096] Likewise, although the present application has been shown and described with respect to one or more implementations, equivalent variations and modifications will occur to those skilled in the art upon reading and understanding this specification and the accompanying drawings. With particular regard to the various functions performed by the components (e.g., elements, resources, etc.) described above, unless otherwise noted, the terms used to describe such components are intended to correspond to any component (functionally equivalent) that performs the specific functions of the described components, even if not structurally equivalent to the disclosed structures.

[0097] Additionally, while particular features of the present application may have been disclosed with respect to only one of several implementations, such features may be combined with one or more other features of other implementations as may be desired and advantageous for any given or particular application.

[0098] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary technical means in the art that are not disclosed in the present application, and the specification and embodiments are only considered as exemplary.

[0099] It should be understood that the present application is not limited to the exact construction that has been described above and shown in the drawings, and that various modifications and changes may be made without departing from the scope thereof.

Claims

1. A method for identifying part features of a vertical machining center, characterized in that: include: A set of images of the same flange produced by a vertical machining center in a top-down perspective is obtained; the interval of the rotation angle of the flange in the image set is equal to an integer multiple of the minimum angle that makes the flange rotationally symmetrical; Acquire feature points of the images in the image set, and use feature points belonging to edge pixels as edge feature points, so as to determine the recognition difficulty value of the image according to the number of edge feature points; For an image pair consisting of any two images in the image set, feature matching is performed on the two images in the image pair, and a matching degree value of the image pair is determined according to the number of matched feature points and the recognition difficulty value of the images in the image pair; When the average value of the matching degree values ​​of different image pairs is less than the first preset threshold, outputting a first prompt information; the first prompt information is used to prompt that there is a processing defect in the flange; Among them, the matching degree value , is the number of matching feature points in either image of the image pair, and are the number of feature points of different images in the image pair, exp is an exponential function with a natural constant as the base, as well as They are the recognition difficulty values ​​of different images in the image pair.

2. The part feature recognition method for a vertical machining center according to claim 1, characterized in that: The recognition difficulty value of an image is determined in the following way: , where D is the recognition difficulty value of the image, N is the number of edge lines in the image, and norm is the normalization function. is the number of feature points located on the i-th edge line among the feature points of the image determined by the feature detection algorithm, is the variance of the smoothness values ​​of all flange edge lines in the image.

3. The part feature recognition method for a vertical machining center according to claim 2, characterized in that: The smoothness value of the edge line is determined by: ,in, is the smoothness value of the target edge line in the image, exp is an exponential function with a natural constant as the base, norm() is the normalization function, is the range of curvature of all points on the target edge line, is the variance of the sine value of the gradient direction angle of all points on the target edge line. The smoothness value is used to characterize the smoothness of the curvature of the pixel points on the edge line.

4. The part feature recognition method for a vertical machining center according to claim 1, characterized in that: The image set is obtained in the following manner: When the flange is located at an initial position in a capture screen of the image capture device, an image of the bolt holes of the flange in a top-down perspective in the capture screen is acquired by the image capture device; When the previous image of the bolt hole of the flange in the top view is obtained, the flange is controlled to rotate along the rotational symmetry center of the flange, and the rotation angle is equal to the minimum angle that makes the flange rotationally symmetric, so as to obtain the next image of the bolt hole of the flange in the top view; When the number of the collected images of the bolt holes of the flange in the top view angle reaches a preset number, the collected images of the bolt holes of the flange in the top view angle are combined into the image set.

5. The part feature recognition method for a vertical machining center according to claim 1, characterized in that: The image set is obtained in the following manner: When the flange is located at an initial position in a capture screen of an image capture device, an image of the flange in a top-down perspective is acquired by the image capture device; When the previous image of the flange at the top-view angle is acquired, the flange is controlled to rotate along the rotational symmetry center of the flange, and the rotation angle is equal to the minimum angle that makes the flange rotationally symmetric, so as to obtain the next image of the flange at the top-view angle; When the number of the collected images of the flange in the bird's-eye view reaches a preset number, the collected images of the flange in the bird's-eye view are combined into the image set.

6. The part feature recognition method for a vertical machining center according to claim 1, characterized in that: The method further comprises: For a target image in an image set, an average value of the matching degree between the target image and other images in the image set except the target image is taken as a feature value of the target image; When the characteristic value of the target image is less than or equal to the second preset threshold, a second prompt information is output; the second prompt information is used to prompt that there is a defect in the flange at the position corresponding to the target image.

7. The part feature recognition method for a vertical machining center according to claim 1, characterized in that: The method further comprises: When the average value of the matching degree values ​​of different image pairs is greater than or equal to a first preset threshold, determining that the inspection result of the flange is qualified; When it is determined that the inspection result of the flange is qualified, the flange moving mechanism is controlled to move the flange to the accommodating space, where the accommodating space is used to accommodate the flange with the qualified inspection result.

8. The part feature recognition method for a vertical machining center according to claim 1, characterized in that: An image pair consisting of any two images in the image set is determined by: Taking any image in the image set as the first image, and selecting any image that has not formed an image pair with the first image from the other images in the image set except the first image as the second image; The first image and the second image form the image pair.

Images