A sprocket machining reference alignment method based on machine vision

Abstract

The application discloses a sprocket machining reference alignment method based on machine vision, and relates to the technical field of intelligent manufacturing.The method is based on the full-image imaging of the front view section of a to-be-detected sprocket obtained by a machine vision method, and performs intelligent size analysis, so that the machining deviation of the to-be-detected sprocket is determined to determine whether the to-be-detected sprocket is qualified, and the qualified to-be-detected sprocket is intelligently matched with a machining baseline template, angular positioning is realized, and machining baseline positioning information is output, so that automatic reference alignment is realized, and intelligent detection support is provided for a full-automatic large sprocket machining process.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of intelligent manufacturing, in particular to a sprocket machining reference alignment method based on machine vision. BACKGROUND

[0002] A sprocket is a wheel with embedded chain teeth used to mesh with block bodies with accurate pitch on a chain ring or cable, and belongs to a common driving part device in the mechanical field. The machining quality of the sprocket will affect the assembly of the mechanical equipment, and even directly cause the operation failure of the mechanical equipment, thereby seriously affecting the service life of the mechanical equipment.

[0003] A large sprocket is produced and machined by using a gas cutting process, and there is usually a certain degree of deviation between the actual profile of the product after machining and the design profile. After the front cutting process is completed, a deviation alignment process needs to be performed. If the deviation between the actual profile of the product and the design profile is too large, the product is classified as unqualified. If the deviation between the actual profile of the product and the design profile is within the allowable range, the actual profile of the product needs to be fitted with the design profile to determine the best cutting reference line for the next process.

[0004] At present, when a manufacturing enterprise performs a deviation alignment process on a machining semi-finished product of a large sprocket, a manual measurement and scribing tool such as a right-angle bending ruler, a scribing gauge and a scribing needle is usually used by an artificial to align the center hole and the reference line, and to determine the machining reference. When the artificial performs the measurement and scribing, the manual measurement efficiency is low, the operation standardization is difficult to control, the measurement precision error fluctuation is uncontrollable, and the like, so that the manufacturing process is difficult to be standardized, and the product quality control difficulty of the enterprise is increased. SUMMARY

[0005] In view of the above problems and technical requirements, the applicant proposes a sprocket machining reference alignment method based on machine vision, and the technical scheme of the application is as follows:

[0006] A sprocket machining reference alignment method based on machine vision, the sprocket machining reference alignment method comprising:

[0007] An image acquisition system is used to acquire the full image of the front view section of the sprocket to be detected and perform binary processing, so as to obtain the contour image of the sprocket to be detected.

[0008] Product geometric data of the sprocket to be detected is determined based on the contour image to be detected, and the product geometric data at least includes each edge point pitch line segment of the sprocket to be detected. Each edge point pitch line segment is a line segment connected between the edge lines of adjacent two teeth.

[0009] The machining deviation of the sprocket to be detected is detected based on the product geometric data of the sprocket to be detected.

[0010] When the machining deviation of the chain wheel to be detected reaches the deviation threshold, it is determined that the chain wheel to be detected is unqualified; when the machining deviation of the chain wheel to be detected does not reach the deviation threshold, the information of the machining baseline indicated by the machining baseline template of the chain wheel to be detected based on the product geometric data of the chain wheel to be detected in the contour image to be detected is determined, and the machining baseline positioning information of the chain wheel to be detected is output.

[0011] A further technical solution is that the method for determining the point-tooth distance line segment between each adjacent tooth of the chain wheel to be detected comprises:

[0012] Identifying the original point of each edge line in the contour image to be detected, the original point of each edge line being the intersection point of the upper edge line segment and the lower edge line segment of the edge line;

[0013] The original point on each edge line is corrected to obtain the corrected point corresponding to the edge line;

[0014] Connecting the corrected point corresponding to the edge line on one side of each tooth with the corrected point corresponding to the edge line on one side of the adjacent tooth to obtain the point-tooth distance line segment between the two teeth.

[0015] A further technical solution is that the method for obtaining the corrected point corresponding to each edge line comprises:

[0016] Identifying the inner circle center of the chain wheel to be detected in the contour image to be detected, and fitting the original points on each edge line in a circle with the inner circle center as the center to obtain the point circle;

[0017] Directly taking the original point located on the point circle as the corrected point corresponding to the edge line, and projecting the original point not located on the point circle to the point circle along the radial direction of the chain wheel to be detected to obtain the corrected point corresponding to the edge line.

[0018] A further technical solution is that the method for detecting the machining deviation of the chain wheel to be detected based on the product geometric data of the chain wheel to be detected comprises:

[0019] Determining the first reference midpoint of the first reference line segment according to the product geometric data of the chain wheel to be detected, the first reference line segment being one of the point-tooth distance line segments;

[0020] Determining the second reference line segment intersecting the connecting line formed by the first reference midpoint and the inner circle center of the chain wheel to be detected determined based on the contour image to be detected, the second reference line segment being one of the point-tooth distance line segments or one of the point-tooth surface line segments, each point-tooth surface line segment being a line segment connected between two edge lines of one tooth;

[0021] Determining the third reference point of the connecting line formed by the second reference midpoint of the second reference line segment and the inner circle center of the chain wheel to be detected on the first reference line segment;

[0022] The Euclidean distance between the first reference midpoint and the third reference point is taken as the machining deviation of the chain wheel to be detected.

[0023] A further technical solution is that the method for determining the first reference line segment according to product geometric data of the chain wheel to be detected comprises:

[0024] determining the standard deviation between each edge point pitch line segment and the standard pitch line segment;

[0025] taking the edge point pitch line segment with the maximum standard deviation from the standard pitch line segment as the first reference line segment.

[0026] A further technical solution is that the method for outputting the machining baseline positioning information of the chain wheel to be detected comprises:

[0027] aligning the template center of the machining baseline template of the chain wheel to be detected with the inner circle center of the chain wheel to be detected determined based on the contour image to be detected;

[0028] positionally aligning the machining baseline template and the contour image to be detected according to the product geometric data of the chain wheel to be detected and the positioning line on the machining baseline template;

[0029] projecting the machining baseline on the machining baseline template onto the positionally aligned contour image to be detected to obtain the machining baseline positioning information of the chain wheel to be detected.

[0030] A further technical solution is that the method for positionally aligning the machining baseline template and the contour image to be detected comprises:

[0031] determining the fourth reference midpoint of the fourth reference line segment according to the product geometric data of the chain wheel to be detected, the positioning line on the machining baseline template being a radial ray with the template center of the machining baseline template as an endpoint, and the positioning line on the machining baseline template indicating the middle position between two teeth, and the fourth reference line segment being one of the edge point pitch line segments;

[0032] determining the fourth connecting line between the inner circle center of the chain wheel to be detected and the fourth reference midpoint, and determining the rotation angle between the positioning line on the machining baseline template and the fourth connecting line;

[0033] rotating the machining baseline template according to the rotation angle so that the positioning line on the machining baseline template coincides with the fourth connecting line, thereby completing the positional alignment between the machining baseline template and the contour image to be detected.

[0034] A further technical solution is that the method for determining the fourth reference line segment of the chain wheel to be detected comprises:

[0035] determining the standard deviation between each edge point pitch line segment and the standard pitch line segment;

[0036] The ridge point pitch line segment with the smallest standard deviation from the standard pitch line segment is taken as the fourth reference line segment.

[0037] A further technical solution is that a method for obtaining a full-view imaging of a front view section of a chain wheel to be detected comprises:

[0038] The chain wheel to be detected is rotated circumferentially by a rotating platform in an image acquisition system, so that different regions of the front view section of the chain wheel to be detected are rotated into the field of view of a line-scan camera in the image acquisition system; during the process that the rotating platform rotates the chain wheel to be detected circumferentially for one round, the line-scan camera in the image acquisition system obtains local images of the regions of the chain wheel to be detected in the field of view, and the line-scan camera faces the rotating platform directly.

[0039] Each of the local images obtained by the line-scan camera is reconstructed to obtain a reconstructed local image, and the reconstructed local images are spliced along the circumferential rotation direction to obtain the full-view imaging of the front view section of the chain wheel to be detected.

[0040] A further technical solution is that a method for reconstructing an i-th frame of local image obtained successively along the circumferential rotation direction to obtain an i-th frame of reconstructed local image comprises:

[0041] determining a reconstruction initial angle of the i-th frame of local image I is the total number of frames of the local images obtained along the circumferential rotation direction.

[0042] A pixel at a coordinate (p_x, p_y) in the i-th frame of reconstructed local image is mapped from a pixel (org_x, frame_h-org_y-1) at an original coordinate (org_x, org_y) in the i-th frame of local image:

[0043]

[0044] wherein a center coordinate in the i-th frame of reconstructed local image is (cx, cy), and rebuild_w is an image width of the i-th frame of reconstructed local image and rebuild_w=frame_w*2+1, rebuild_h is an image height of the i-th frame of reconstructed local image and rebuild_h=frame_w*2+1; rho=frame_w-rebuild_h+org_y+1, frame_w is an image width of the i-th frame of local image.

[0045] The beneficial technical effects of the present application are:

[0046] This application discloses a machine vision-based sprocket machining datum alignment method. This method uses machine vision to perform intelligent dimensional analysis based on the full-view imaging of the frontal cross-section of the sprocket to be inspected, thereby determining the machining deviation of the sprocket to be inspected to determine whether the sprocket to be inspected is qualified. The qualified sprocket to be inspected is intelligently matched with the machining baseline template to achieve angular positioning and alignment, and output machining baseline positioning information. This can realize automated datum alignment and provide intelligent detection support for fully automated large sprocket machining processes. Attached Figure Description

[0047] Figure 1 This is a flowchart of a sprocket machining datum alignment method according to one embodiment of this application.

[0048] Figure 2 This is a schematic diagram of the full-view image of the frontal cross-section of the sprocket to be detected obtained in an example, as well as the binarized image of the contour of the sprocket to be detected.

[0049] Figure 3 This is a schematic diagram of an image acquisition system from one perspective in one embodiment of this application.

[0050] Figure 4 This is a schematic diagram of an image acquisition system from another perspective in one embodiment of this application.

[0051] Figure 5 This is a schematic diagram of the local images of each frame obtained in an example, as well as a schematic diagram of the reconstructed local image of one frame and the final full-image image obtained.

[0052] Figure 6 This is a flowchart illustrating a method for obtaining the tooth pitch segments and tooth surface segments of each edge point in one embodiment of this application.

[0053] Figure 7 This is a schematic diagram of the tooth pitch line segments and tooth surface line segments of each edge point obtained in an example.

[0054] Figure 8 This is a flowchart of a method for determining the machining deviation of a sprocket to be tested in one embodiment of this application.

[0055] Figure 9 This is a schematic diagram of the machining deviation of the sprocket to be tested in an example.

[0056] Figure 10 This is a schematic diagram of a processing baseline template in an example.

[0057] Figure 11 This is a flowchart of a method for obtaining the machining baseline positioning information of the sprocket to be tested in one embodiment of this application.

[0058] Figure 12Fig. 4 is a schematic diagram of a rotation angle between a positioning line on the machining baseline template and a fourth connecting line of the profile image to be detected in an example, and a schematic diagram of a projection of a machining baseline on the machining baseline template on the full image after the position alignment of the machining baseline template and the profile image to be detected. DETAILED DESCRIPTION

[0059] The specific embodiments of the present application will be further described below with reference to the accompanying drawings.

[0060] The present application discloses a chain wheel machining reference alignment method based on machine vision, please refer to the flow chart shown in Fig. 1, which comprises the following steps: Figure 1

[0061] Step 1, the full image of the front view section of the chain wheel to be detected is obtained by the image acquisition system and is subjected to binary processing, so that the chain wheel to be detected can be distinguished from the background image, and the profile image of the front view section of the chain wheel to be detected is obtained.

[0062] Figure 2 Fig. 2 shows a schematic diagram of the full image of the front view section of the chain wheel 1 to be detected and the profile image after binary processing obtained in an example. The front view section of the chain wheel to be detected is a section perpendicular to the rotation axis of the chain wheel to be detected, and the full image of the front view section of the chain wheel 1 to be detected obtained includes a circular structure of the chain wheel to be detected and a plurality of gear teeth 1a distributed circumferentially. Figure 2 Taking the chain wheel 1 containing 7 gear teeth 1a along the circumference as an example.

[0063] The chain wheel to be detected targeted by the present application is generally a large-scale industrial chain wheel. Since the size of the chain wheel to be detected is large, it is difficult to directly obtain the full image of the front view section of the chain wheel to be detected as shown in Fig. 1, so in an embodiment, the image acquisition system for obtaining the full image of the front view section of the chain wheel to be detected comprises a rotating platform 2 and a line scan camera 3, please refer to the structure of the image acquisition system shown in Figs. 3 and 4. Figure 2 Figure 4 The chain wheel 1 to be detected is placed on the rotating platform 2, and the front view section of the chain wheel 1 to be detected is parallel to the plane of the rotating platform 2. In order to facilitate positioning, a protruding positioning fixed shaft is provided on the rotating platform 2, and the inner circle of the chain wheel 1 to be detected is sleeved on the positioning fixed shaft of the rotating platform 2 to realize positioning and fixing.

[0064] ​​The line scan camera 3 is fixed above and opposite to the rotating platform 2. When the rotating platform 2 rotates circumferentially, the chain wheel 1 placed on the rotating platform also rotates circumferentially, so that different regions of the front view section of the chain wheel 1 rotate into the field of view of the line scan camera 3 in the image acquisition system. During the circumferential rotation of the chain wheel 1 driven by the rotating platform 2, the line scan camera 3 acquires partial images of different regions of the chain wheel 1.

[0065] After the rotating platform 2 drives the chain wheel 1 to rotate circumferentially by 350 degrees, the line scan camera 3 can acquire multiple partial images of different regions of the chain wheel 1 in sequence. In actual application, the rotating platform 2 drives the chain wheel 1 to rotate each tooth into the field of view of the line scan camera 3 in sequence, so that each partial image acquired by the line scan camera 3 includes an image of a tooth and the surrounding region, and therefore in the example shown in Figure 2 Figure 5 Each of the seven partial images acquired by the line scan camera 3 covers a region of a tooth. Then, each of the partial images acquired by the line scan camera is reconstructed to obtain a reconstructed partial image.

[0066] In one embodiment, the method for reconstructing the i-th partial image acquired in sequence along the circumferential rotation direction to obtain the i-th reconstructed partial image includes determining a reconstruction initial angle of the i-th partial image I is the total number of partial images acquired along the circumferential rotation direction. Then, the pixel at the coordinate (p_x, p_y) in the i-th reconstructed partial image is mapped from the pixel (org_x, frame_h-org_y-1) at the original coordinate (org_x, org_y) in the i-th partial image:

[0067]

[0068] wherein the center coordinate in the i-th reconstructed partial image is (cx, cy), and rebuild_w is the image width of the i-th reconstructed partial image and rebuild_w = frame_w*2+1, rebuild_h is the image height of the i-th reconstructed partial image and rebuild_h = frame_w*2+1; rho = frame_w-rebuild_h+org_y+1, frame_w is the image width of the i-th partial image.

[0069] ​Then the reconstructed partial images are spliced along the circumferential rotation direction to obtain the full view imaging of the front view section of the chain wheel to be detected. In order to facilitate splicing, when the partial images are collected, the field of view range of the line-scan camera 3 and the shooting position can be adjusted so that the acquired partial images are connected but not overlapped, and then the reconstructed partial images are directly spliced. Alternatively, the adjacent reconstructed partial images can also have overlaps, and the existing image splicing technology can also be used to complete the image splicing. Figure 5 The schematic diagram of the reconstructed partial image obtained by reconstructing one frame of partial image is shown, and the reconstructed partial images obtained by reconstructing other frames of partial image are not all shown. Finally, the reconstructed partial images are spliced along the circumferential rotation direction, and the full view imaging of the front view section of the chain wheel to be detected as shown in Figure 2 is obtained.

[0070] In addition, in order to improve the quality of the acquired image, a visual light source 4 is also fixed above the rotating platform 2, and the visual light source 4 irradiates the field of view range of the line-scan camera 3 to improve the image quality of the partial image acquired by the line-scan camera 3.

[0071] The line-scan camera 3 and the visual light source 4 can be fixed on the gantry 9. In addition, in the image acquisition system, in order to facilitate the taking and placing of the chain wheel 1 to be detected, as shown in Figure 3 and Figure 4 , the rotating platform 2 is placed on the conveying platform 5, and the conveying platform 5 is arranged in the conveying slide rail 6 and can slide along the conveying slide rail 6. When the conveying platform 5 moves along the conveying slide rail 6 to the outside of the gantry 9, the chain wheel 1 to be detected can be conveniently hoisted and placed on the rotating platform 2 or hoisted and taken off from the rotating platform 2. When the chain wheel 1 to be detected is placed on the rotating platform 2 and needs to be detected, the conveying platform 5 moves along the conveying slide rail 6 to the inside of the gantry 9 and is located below the line-scan camera 3 and the visual light source 4, and the full view imaging can be collected according to the above method. As shown in Figure 3 and 4 , the case that the conveying platform 5 moves along the conveying slide rail 6 to the inside of the gantry 9 and is located below the line-scan camera 3 and the visual light source 4 is taken as an example for illustration.

[0072] In order to adapt to different sizes of the chain wheel 1 to be detected, the line-scan camera 3 is fixed on the gantry 9 through the three-axis motion platform 8, so that the line-scan camera 3 can be adjusted in position through the three-axis motion platform 8 to better shoot the chain wheel 1 to be detected. Similarly, the visual light source 4 is fixed on the gantry 9 through the three-axis motion platform 7, so that the line-scan camera 3 can be adjusted in position through the three-axis motion platform 7 to better irradiate the chain wheel 1 to be detected.

[0073] The image acquisition system also includes a central controller 10, which connects to and controls the operation of each component in the image acquisition system, and acquires multiple frames of local images captured by the line scan camera 3, as well as obtains a full-image image through the above method. The sprocket machining datum alignment method of this application can also be considered as a method executed by the central controller 10.

[0074] Step 2: Determine the product geometry data of the sprocket to be inspected based on the contour image to be inspected. The product geometry data includes at least the tooth pitch line segments of each edge point of the sprocket to be inspected.

[0075] Each tooth of the sprocket 1 to be tested includes two side lines. Each edge point tooth pitch line segment is a line segment connecting the side lines of two adjacent teeth. It is a line segment connecting the right side edge line of any tooth A and the left side edge line of its adjacent tooth B. Tooth B is the tooth located to the right of tooth A.

[0076] In one embodiment, the method for determining the tooth pitch segment of each edge point includes, please refer to... Figure 6 The flowchart shown below:

[0077] Step 2.1: Identify the original edge points on each edge line of each tooth in the contour image to be detected. The original edge points on each edge line are the intersection points of the upper and lower edge line segments of the edge line.

[0078] Step 2.2 involves correcting the original edge points on each edge line to obtain the corrected edge points for that edge line. This includes:

[0079] (1) Identify the inner circle center of the sprocket to be detected in the contour image to be detected, and perform circumferential fitting on the original edge points on each edge line with the inner circle center as the center to obtain the edge point circle. Theoretically, due to the symmetrical structure of the sprocket, the original edge points on each edge line should all be on the edge point circle. However, due to the actual machining error, some gear teeth may have machining errors, which will result in some original edge points not being on the edge point circle.

[0080] For example, please refer to Figure 7 ,from Figure 2 In the example shown, the inner circle center O and 14 original edge points N1 to N14 can be identified in the contour image to be detected. The edge point circles obtained by fitting the original edge points with O as the center are shown by the dashed lines. Figure 7 It can be seen that due to machining deviations in the gear teeth, the original edge points N7 and N8 are not on the edge point circle, while the other 12 original edge points are on the edge point circle. It should be noted that the actual contour image to be detected is as follows... Figure 2 The image is a binary representation, but for ease of illustration, it is shown below. Figure 7 The following illustrations only show the outer contour of the tooth to be detected as indicated by the image of the contour to be detected.

[0081] (2) The original vertex point located on the vertex point circle is directly taken as the modified vertex point corresponding to the edge line where the original vertex point is located. The original vertex point not located on the vertex point circle is projected onto the vertex point circle along the radial direction of the chain wheel to be detected to obtain the modified vertex point corresponding to the edge line where the original vertex point is located. The projection onto the vertex point circle along the radial direction of the chain wheel to be detected means that the intersection point of the radial ray between the original vertex point and the inner circle center O and the vertex point circle is the modified vertex point. For example, in Figure 7 , the intersection point N8' of the radial ray between the original vertex point N8 and the inner circle center O and the vertex point circle is taken as the modified vertex point corresponding to the edge line where the original vertex point N8 is located. The intersection point N7' of the radial ray between the original vertex point N7 and the inner circle center O and the vertex point circle is taken as the modified vertex point corresponding to the edge line where the original vertex point N7 is located. The original vertex points not on the vertex point circle are directly taken as the modified vertex points corresponding to the edge lines where the original vertex points are located.

[0082] Step 2.3, connecting the modified vertex points corresponding to the edge lines on one side of each tooth with the modified vertex points corresponding to the edge lines on the adjacent side of the adjacent tooth to obtain the vertex point pitch line segment between the two teeth. For example, Figure 7 , seven vertex point pitch line segments L11, L12, L13, L14, L15, L16 and L17 respectively located between the seven teeth are obtained.

[0083] In another embodiment, the product geometric data of the chain wheel to be detected includes not only the vertex point pitch line segments of the chain wheel to be detected but also the vertex point tooth surface line segments of the chain wheel to be detected. Each vertex point tooth surface line segment is a line segment connected between two edge lines of a tooth. Then, as Figure 6 shown in the flowchart, after the modified vertex points corresponding to each edge line are determined, the modified vertex points corresponding to the edge lines on both sides of each tooth are connected to obtain the vertex point tooth surface line segment connected between the two edge lines of the tooth. Please refer to the example shown in Figure 7 , seven vertex point tooth surface line segments L21, L22, L23, L24, L25, L26 and L27 respectively located on the tooth surfaces of the seven teeth are obtained.

[0084] In addition to the vertex point pitch line segment and the vertex point tooth surface line segment, the product geometric data of the chain wheel to be detected also includes other parameters for characterizing the geometric features of the chain wheel to be detected, including the angle between the radial ray from the inner circle center to each original vertex point and the pre-marked positive direction, the angle between the radial ray from the inner circle center to each modified vertex point and the pre-marked positive direction, and the like. The specific output parameters can be configured according to the actual situation, thereby providing the required parameters for the subsequent process.

[0085] Step 3, detecting the machining deviation of the chain wheel to be detected based on the product geometric data of the chain wheel to be detected. The steps include the following steps, please refer to the flowchart shown in Figure 8 and the example shown in Figure 9The example diagram shown is as follows:

[0086] Step 3.1: Determine the first reference midpoint Node1 of the first reference line segment based on the product geometry data of the sprocket to be tested. The first reference line segment is one of the edge point tooth pitch segments. Based on actual machining experience, the machining deviation of the sprocket to be tested is mainly the deviation in the tooth pitch between the teeth, and the machining deviation in the size of the teeth is not significant. Therefore, this application uses the edge point tooth pitch segment as a reference to detect the machining deviation of the sprocket to be tested.

[0087] To better reflect the machining deviation of the sprocket under test, and to minimize the number of correction steps required when subsequently correcting the sprocket, in one embodiment, the standard deviation between each edge point tooth pitch segment and the standard tooth pitch segment is first determined. Then, the edge point tooth pitch segment with the largest standard deviation from the standard tooth pitch segment is used as the first reference segment. For example, in... Figure 9 In the example, the edge point tooth pitch segment L14, which has the largest standard deviation from the standard tooth pitch segment, is taken as the first reference segment.

[0088] Step 3.2: Determine the second reference line segment that intersects the line formed by the first reference midpoint Node1 and the inner circle center O of the sprocket to be detected, which is determined based on the contour image to be detected.

[0089] Depending on the structure of the sprocket being tested, the second reference line segment could be either a tooth pitch segment at one edge point or a tooth surface segment at one edge point. For example, in... Figure 9 In the example, the second reference line segment is the edge point tooth surface line segment L21.

[0090] Step 3.3: Determine the third reference point Node3 on the first reference line segment by connecting the second reference midpoint Node2 of the second reference line segment with the inner circle center O of the sprocket to be tested.

[0091] Step 3.4: The Euclidean distance between the first reference midpoint Node1 and the third reference point Node3 is used as the machining deviation of the sprocket to be tested.

[0092] Step 4: When the machining deviation of the sprocket to be tested reaches the deviation threshold, the sprocket to be tested is determined to be unqualified, and the benchmark alignment method ends.

[0093] Step 5: When the machining deviation of the sprocket to be tested does not reach the deviation threshold, determine the information of the machining baseline indicated by the machining baseline template of the sprocket to be tested in the image of the sprocket to be tested based on the product geometry data of the sprocket to be tested, and output the machining baseline positioning information of the sprocket to be tested.

[0094] Please refer to Figure 10A schematic diagram of the machining baseline template for the sprocket to be inspected is shown. The machining baseline template is a standard design drawing of the sprocket to be inspected. It includes several machining baselines to indicate the positions to be machined on the sprocket. In addition to the machining baselines, the template also includes a positioning line Ls. The positioning line Ls is a radial ray with the center O' of the template as its endpoint. In practical applications, the positioning line Ls indicates the midpoint between two teeth or the midpoint of a single tooth, and can be pre-configured. Generally, the positioning line on the machining baseline template indicates the midpoint between two teeth. This step includes the following steps; please refer to the flowchart in Figure 1:

[0095] Step 5.1: Align the center O' of the machining baseline template of the sprocket to be inspected with the center O of the inner circle of the sprocket to be inspected, which is determined based on the contour image to be inspected.

[0096] Step 5.2: Align the machining baseline template and the contour image to be inspected according to the product geometry data of the sprocket to be inspected and the positioning line Ls on the machining baseline template. This includes:

[0097] (1) Determine the fourth reference midpoint Node4 of the fourth reference line segment based on the product geometry data of the sprocket to be tested. Since the positioning line Ls on the machining baseline template indicates the midpoint between two teeth, in this step, the fourth reference line segment is one of the edge point tooth pitch segments.

[0098] Similarly, in order to facilitate the subsequent correction of the sprocket under inspection with as few correction steps as possible, this step first determines the standard deviation between each edge point tooth pitch segment and the standard tooth pitch segment, and then uses the edge point tooth pitch segment with the smallest standard deviation from the standard tooth pitch segment as the fourth reference segment Node4.

[0099] (2) Determine the fourth line connecting the inner circle center O of the sprocket to be tested and the fourth reference midpoint Node4, and determine the rotation angle α between the positioning line Ls on the machining baseline template and the fourth line.

[0100] In practice, the angle θ between the fourth line and the predetermined positive direction can be determined first, and then the angle β between the positioning line Ls and the predetermined positive direction can be determined. The rotation angle α between the positioning line Ls and the fourth line can be determined by the angle difference between the angles θ and β.

[0101] (3) Rotate the processing baseline template according to the rotation angle α so that the positioning line Ls on the processing baseline template coincides with the fourth connecting line, thus completing the position alignment between the processing baseline template and the contour image to be detected.

[0102] Step 5.3, project the machining base line on the machining base line template which is positionally aligned to the completed position, onto the positionally aligned to-be-detected contour image, to obtain machining base line positioning information of the to-be-detected sprocket. As shown in Fig. 5.3, the machining base line on the machining base line template which is positionally aligned to the completed position is projected onto the positionally aligned to-be-detected contour image, to obtain machining base line positioning information of the to-be-detected sprocket. Figure 12 The projection of the machining base line on the full-image imaging is shown in Fig. 5.3.

[0103] The above only describes the preferred embodiments of the present application, and the present application is not limited to the above embodiments. It can be understood that other improvements and changes directly derived or thought by those skilled in the art without departing from the spirit and concept of the present application shall be considered to be included in the protection scope of the present application.

Claims

1. A machine vision based sprocket machining datum alignment method, characterized in that, The chain wheel machining reference alignment method comprises: An image acquisition system is used to acquire a full image of the front view section of the chain wheel to be detected and perform binary processing to obtain a contour image of the chain wheel to be detected; Product geometry data of the chain wheel to be detected is determined based on the contour image of the chain wheel to be detected, and the product geometry data at least comprises each edge point pitch line segment of the chain wheel to be detected, each edge point pitch line segment being a line segment connected between the edge lines of two adjacent teeth; A machining deviation of the chain wheel to be detected is detected based on the product geometry data of the chain wheel to be detected; When the machining deviation of the chain wheel to be detected reaches a deviation threshold, it is determined that the chain wheel to be detected is unqualified; when the machining deviation of the chain wheel to be detected does not reach the deviation threshold, information of a machining reference line of the chain wheel to be detected indicated by a machining reference line template determined based on the product geometry data of the chain wheel to be detected in the contour image of the chain wheel to be detected is determined, and machining reference line positioning information of the chain wheel to be detected is outputted; The method for determining each edge point pitch line segment between adjacent teeth of the chain wheel to be detected comprises: identifying original edge points on each edge line of each tooth in the contour image of the chain wheel to be detected, the original edge point on each edge line being an intersection point of an upper edge line segment and a lower edge line segment of the edge line; performing correction processing on the original edge points on each edge line to obtain a corrected edge point corresponding to the edge line; connecting the corrected edge point on one side of each tooth to the corrected edge point on one side of an adjacent tooth to obtain an edge point pitch line segment between the two teeth; The method for obtaining the corrected edge point corresponding to each edge line comprises: identifying a center of an inner circle of the chain wheel to be detected in the contour image of the chain wheel to be detected; performing a circumferential fitting on the original edge points on each edge line with the center of the inner circle as the center to obtain an edge point circle; directly taking the original edge point located on the edge point circle as the corrected edge point corresponding to the edge line, and projecting the original edge point not located on the edge point circle to the edge point circle along a radial direction of the chain wheel to be detected to obtain the corrected edge point corresponding to the edge line.

2. The sprocket machining reference alignment method of claim 1, wherein, The method for detecting the machining deviation of the chain wheel to be detected based on the product geometry data of the chain wheel to be detected comprises: A first reference midpoint of a first reference line segment is determined according to the product geometry data of the chain wheel to be detected, the first reference line segment being one of the edge point pitch line segments; A second reference line segment intersecting a line formed by the first reference midpoint and a center of an inner circle of the chain wheel to be detected determined based on the contour image of the chain wheel to be detected is determined, the second reference line segment being one of the edge point pitch line segments or one of the edge point tooth surface line segments, each edge point tooth surface line segment being a line segment connected between two edge lines of one tooth; A third reference point of the line formed by the second reference midpoint of the second reference line segment and the center of the inner circle of the chain wheel to be detected on the first reference line segment is determined; A Euclidean distance between the first reference midpoint and the third reference point is taken as the machining deviation of the chain wheel to be detected.

3. The sprocket machining reference alignment method of claim 2, wherein, The method for determining the first reference line segment according to the product geometry data of the chain wheel to be detected comprises: determining standard deviations between each addendum pitch line segment and a standard pitch line segment; taking the addendum pitch line segment with the largest standard deviation from the standard pitch line segment as the first reference line segment.

4. The sprocket machining reference alignment method of claim 1, wherein, The method for outputting the machining base line positioning information of the chain wheel to be detected comprises: aligning a template center of a machining base line template of the chain wheel to be detected with an inner circle center of the chain wheel to be detected determined based on the contour image of the chain wheel to be detected; positionally aligning the machining base line template and the contour image of the chain wheel to be detected according to product geometric data of the chain wheel to be detected and positioning lines on the machining base line template; projecting a machining base line on the machining base line template onto the contour image of the chain wheel to be detected after position alignment to obtain machining base line positioning information of the chain wheel to be detected.

5. The sprocket machining reference alignment method of claim 4, wherein, The method for positionally aligning the machining base line template and the contour image of the chain wheel to be detected comprises: determining a fourth reference midpoint of a fourth reference line segment according to product geometric data of the chain wheel to be detected, the positioning lines on the machining base line template being radial rays with a template center of the machining base line template as an end point, and the positioning lines on the machining base line template indicating intermediate positions between two teeth, the fourth reference line segment being one of the addendum pitch line segments; determining a fourth connecting line between the inner circle center of the chain wheel to be detected and the fourth reference midpoint, and determining a rotation included angle between the positioning lines on the machining base line template and the fourth connecting line; rotating the machining base line template according to the rotation included angle so that the positioning lines on the machining base line template coincide with the fourth connecting line, and completing position alignment between the machining base line template and the contour image of the chain wheel to be detected.

6. The sprocket machining reference alignment method of claim 5, wherein, The method for determining the fourth reference line segment of the chain wheel to be detected comprises: determining standard deviations between each addendum pitch line segment and a standard pitch line segment; taking the addendum pitch line segment with the smallest standard deviation from the standard pitch line segment as the fourth reference line segment.

7. The sprocket machining reference alignment method of claim 1, wherein, The method for obtaining full-image imaging of the front view section of the chain wheel to be detected comprises: driving the chain wheel to be detected placed on a rotating platform in the image acquisition system to rotate circumferentially so that different regions of the front view section of the chain wheel to be detected rotate into the field of view range of a line-scan camera in the image acquisition system; in the process of driving the chain wheel to be detected to rotate circumferentially for one round by the rotating platform, obtaining local images of the regions of the chain wheel to be detected in the field of view range by the line-scan camera in the image acquisition system, the line-scan camera facing the rotating platform; reconstructing each local image obtained by the line-scan camera to obtain reconstructed local images, and splicing the reconstructed local images in the circumferential rotation direction to obtain full-image imaging of the front view section of the chain wheel to be detected.

8. The sprocket machining reference alignment method of claim 7, wherein, The method for reconstructing the partial image of the first frame includes: The method for reconstructing the partial image of the first frame includes: The method for reconstructing the partial image of the first frame includes: determining the first reconstruction initial angle of the frame partial image , is the total number of partial images acquired along the circumferential rotation direction; The first Original coordinates in a frame local image The pixel at that location is mapped to the first pixel. Coordinates in the local image after frame reconstruction Location: ； wherein the first The center coordinates in the local image after frame reconstruction are , and , , is the image width of the first local image after frame reconstruction and , is the image height of the first local image after frame reconstruction and ; , , is the image width of the first local image.

Images