A method and system for detecting the processing quality of resolver wire harnesses based on machine vision

Abstract

The present invention relates to the field of image analysis, and particularly to a method and system for detecting the processing quality of resolver wire harnesses based on machine vision. The method first obtains the image of the inner core wire head of the target resolver wire harness, then uses the edge detection method to identify all the inner core edges in the inner core wire head image, then draws an identification frame corresponding to each inner core according to the inner core edges, and finally determines whether the inner core edges meet the following conditions: any inner core edge extends beyond the corresponding identification frame in the width direction; the inclination angle of any inner core edge exceeds the inclination interval; if any condition is met, it is determined that the target resolver wire harness is unqualified. Compared with the prior art, the present invention can adaptively draw identification frames for a variety of different resolver wire harnesses, and solves the problems of poor recognition reliability and low efficiency in the prior art.

Description

Technical Field

[0001] The present invention relates to the field of image analysis, and more specifically, to a method and system for detecting the processing quality of resolver wire harnesses based on machine vision. Background Art

[0002] A resolver wire harness, also known as a resolver transformer wire harness, is an essential part of a motor control system, mainly used for transmitting current and detection signals, and is generally composed of multiple wires. The two ends of the resolver wire harness usually need to be connected to the power supply system and the above-mentioned motor control system through connectors respectively, which means that the wire ends of the resolver wire harness need to be electrically connected and fastened to the connector busbars. Before connecting the wire ends to the connector busbars, it is necessary to cut open the rubber sleeve on the surface of the resolver wire harness to expose its inner core, so as to facilitate the inner core of the wire end to be inserted into the connector busbar. In the actual application environment, the stability of power transmission or signal transmission of the resolver wire harness is closely related to the reliability of the connection between the inner core of the wire end and the connector. During the process of cutting open the rubber sleeve, if there is residual rubber sleeve or impurities on the inner core of the wire end, the conductivity of the inner core of the wire end will decrease. If the sharpness of the cutting mechanism is insufficient, the pressure on the wire end during cutting will be unstable, resulting in the bending of the wire end and inability to directly insert into the connector busbar. Therefore, it is necessary to detect the quality of the inner core of the wire end during the process of peeling the rubber sleeve from the wire end of the resolver wire harness.

[0003] In the prior art, there is a quality detection method for the inner core, and its detection logic is as follows: first, select the recognition range, then identify the inner core according to the preset graphic template, and then calculate the width of the inner core. If the width of a certain section of the inner core exceeds the threshold, it is determined that the inner core is a defective product. The above technology mainly performs image recognition on a single inner core. For a wire harness with multiple inner cores such as a resolver wire harness, the teaching and template storage of the graphic template are required for the recognition of the inner core of each type of resolver wire harness. This method has the following defects:

[0004] First, the judgment of the teaching effect mainly depends on the experience of the staff, and the inclination of multiple wire harnesses cannot be recognized, resulting in the inability to detect the inclination and bending of the head end of the wire in time. This method has poor reliability for the quality detection of resolver inner cores.

[0005] Second, it is necessary to replace the recognition formula of the image template according to the model of the resolver wire harness, and the recognition efficiency is low. Summary of the Invention

[0006] In order to solve the technical problems of poor recognition reliability and low efficiency in the prior art, the present invention provides solutions in the following aspects.

[0007] In a first aspect, the present invention provides a method for detecting the processing quality of resolver wire harnesses based on machine vision, including:

[0008] Obtain the image of the inner core wire head of the target resolver harness;

[0009] Use the edge detection method to identify all inner core edges in the inner core wire head image;

[0010] Draw an identification box corresponding to each inner core according to the inner core edges;

[0011] Determine whether the inner core edges meet the following conditions: any inner core edge exceeds the corresponding identification box in the width direction;

[0012] The inclination angle of any inner core edge exceeds the inclination interval;

[0013] Among them, the calculation formula for the inclination angle is:

[0014]

[0015] In the formula, represents the inclination angle of the length direction of the j th inner core edge relative to the X-axis, arctan() is the arctangent function, is the ordinate of the j th edge point of the i th inner core edge, is the ordinate of the midpoint of the top edge of the j th inner core edge, is the j total number of edge points of the th inner core edge, j is the i th abscissa of the th edge point of the j th inner core edge, is the i upper limit value of the serial number;

[0016] If any condition is met, it is determined that the target resolver harness is unqualified.

[0017] Beneficial effects: The method of the present invention first uses the edge detection method to simultaneously identify the edges of multiple inner cores in the target resolver wire harness, then draws the recognition frames corresponding to the inner cores based on the edges of the multiple inner cores, and then determines whether any one of the inner core edges meets any condition. If the inner core edge exceeds the corresponding recognition frame in the width direction, it indicates that there is glue sleeve residue or impurity residue on a certain bare inner core of the resolver wire harness. If the inclination angle of any inner core edge exceeds the inclination interval, it indicates that the bending degree of a certain bare inner core of the resolver wire harness is too large. Compared with the prior art, the method of the present invention no longer relies on the template teaching method to draw the recognition frame, can adaptively draw the recognition frame for a variety of different resolver wire harnesses, and the method of the present invention also incorporates an algorithm dedicated to calculating the bending degree of the inner core, which can quickly and relatively accurately calculate the bending degree of the inner core, solving the problems of poor reliability and low efficiency in the identification and judgment of the prior art.

[0018] Preferably, the edge recognition formula of the edge detection method is:

[0019]

[0020]

[0021]

[0022] In the formula, is the image of the inner core wire head, is the convolution operation, is the operator matrix in the x-axis direction, is the operator matrix in the y-axis direction, is the gradient of the inner core wire head image in the x-axis direction, is the gradient of the inner core wire head image in the y-axis direction, is the pixel gradient of the pixel at coordinates (x, y) in the inner core wire head image; if , the pixel at coordinates (x, y) is a non-edge pixel; if , the pixel at coordinates (x, y) is an edge pixel; is the lower limit of the edge threshold.

[0023] Beneficial effects: The above edge recognition formula first convolves the pixels in the inner core wire head image, then calculates the pixel gradients in the x and y axis directions respectively, then integrates the gradients in the x and y directions as the pixel gradient of the pixel at coordinates (x, y), and finally uses the single-threshold discrimination method to distinguish edge pixels and non-edge pixels. This method reduces the computational amount of edge pixel operations and can improve the recognition efficiency of inner core edges.

[0024] Preferably, the edge pixels include weak edge pixels and strong edge pixels. If , it is determined that the pixel at coordinates (x, y) is an edge pixel. Specifically:

[0025] If , it is determined that the pixel with coordinates (x, y) is a weak edge pixel;

[0026] If , it is determined that the pixel with coordinates (x, y) is a strong edge pixel;

[0027] Wherein, is the strong edge threshold, and the strong edge pixels are the edge pixels that are preferentially connected for pixel connection.

[0028] Beneficial effects: By using the double-threshold determination method to distinguish strong edge pixels and weak edge pixels, and preferentially connecting the strong edge pixels, this method can obtain the edge of the inner core more quickly.

[0029] Preferably, after drawing the recognition frame corresponding to each inner core, the method of the present invention further includes:

[0030] Connecting a plurality of adjacent weak edge pixels within the recognition frame to form an edge to be measured;

[0031] Judging whether the length of the edge to be measured exceeds the crack depth threshold;

[0032] If so, it is determined that the target resolver harness is unqualified.

[0033] Beneficial effects: After connecting the strong edge pixels, the weak edge pixels are connected within the recognition frame as the edge of the crack to be measured. If the crack edge exceeds the crack depth threshold, it indicates that the conductivity of the inner core is seriously affected by the crack. Compared with the prior art, the method of the present invention can quickly identify the inner core with obvious cracks, thereby improving the reliability of the quality detection of the resolver inner core.

[0034] Preferably, after obtaining the edge pixels, the method of the present invention further includes:

[0035] Connecting a plurality of adjacent edge pixels to form a plurality of closed inner core edges;

[0036] Sifting out the inner core edges that are close to each other in the x-axis direction.

[0037] Beneficial effects: During the generation of edge pixels, due to the influence of external environmental factors, misjudgment is inevitable, resulting in the connection of multiple closed contours, which is not conducive to recognition. By adopting the above solution, sifting out multiple close edges can overcome the recognition interference caused by external factors and improve the accuracy of the method of the present invention.

[0038] Preferably, according to the inner core edge, drawing the recognition frame corresponding to each inner core specifically includes:

[0039] Obtaining the inner diameter parameter of each inner core according to the model of the target resolver harness as the frame selection width of the recognition frame;

[0040] Take the single wire cutting stroke of the wire cutting device as the selected length of the recognition frame;

[0041] According to the selected width and the selected length, with the geometric center of the corresponding inner core edge as the reference center, draw the corresponding recognition frame.

[0042] Beneficial effects: The method of the present invention can adaptively adjust the selected size of the recognition frame according to the model of the target resolver wire harness and the single wire cutting stroke of the wire cutting device, without manual intervention for teaching. Compared with the prior art, the recognition efficiency of the method of the present invention is higher.

[0043] Preferably, after obtaining the image of the inner core lead of the target resolver wire harness, the method of the present invention further includes:

[0044] Draw a detection frame for the image of the inner core lead;

[0045] Wherein, the recognition frame is located within the detection frame.

[0046] Beneficial effects: After obtaining the image of the inner core lead, drawing the detection frame limits the range of image recognition, which can reduce the data operation amount of the method of the present invention, and thus improve the recognition efficiency of the method of the present invention.

[0047] Preferably, after drawing the recognition frame for each inner core, the method of the present invention further includes:

[0048] Output the number of recognition frames as the number of inner cores to the monitoring terminal;

[0049] Calculate and output the width data of all inner core edges to the monitoring terminal;

[0050] Calculate and output the length data of all inner core edges to the monitoring terminal;

[0051] According to the width data and the length data, calculate and output the area data of the corresponding inner core to the monitoring terminal.

[0052] Beneficial effects: After the recognition is completed, the method of the present invention will output data such as the number of inner cores, the width of the inner cores, the length of the inner cores, and the area of the inner cores to the monitoring terminal, which is conducive to professional personnel for rejudgment or evaluation.

[0053] Preferably, after obtaining the image of the inner core lead of the target resolver wire harness, the method of the present invention further includes:

[0054] Perform filtering processing on the image of the inner core lead.

[0055] Preferably, after obtaining the image of the inner core lead of the target resolver wire harness, the method of the present invention further includes:

[0056] Perform binarization processing on the image of the inner core lead.

[0057] Beneficial effects: By filtering and binarizing the image of the inner core wire end, the image features of the inner core wire end image can be enhanced, which is beneficial to improving the accuracy of edge recognition, and further improving the reliability of the method of the present invention for detecting the quality of the inner core.

[0058] In a second aspect, the present invention further provides a detection system for the processing quality of resolver wire harnesses based on machine vision, including a processor and a memory. The above-mentioned memory stores computer program instructions, and when the computer program instructions are executed by the processor, the method for detecting the processing quality of resolver wire harnesses based on machine vision in the first aspect is implemented.

[0059] The beneficial effects of the present invention are as follows:

[0060] (1) Compared with the prior art, the method of the present invention no longer relies on the template method to draw the recognition frame, and can adaptively draw the recognition frame for a variety of different resolver wire harnesses, solving the problems of poor recognition reliability and low efficiency in the prior art.

[0061] (2) The method of the present invention can adaptively adjust the frame selection size of the recognition frame according to the model of the target resolver wire harness and the single wire cutting stroke of the wire cutting device, without manual intervention for teaching. Compared with the prior art, the recognition efficiency of the method of the present invention is higher.

[0062] (3) Compared with the prior art, the method of the present invention can screen out multiple adjacent edges to overcome the recognition interference caused by external factors, thereby improving the accuracy of the recognition of the method of the present invention. Description of the Drawings

[0063] Figure 1 is a flowchart schematically showing the method for detecting the processing quality of resolver wire harnesses based on machine vision in Embodiment 1 of the present invention;

[0064] Figure 2 is an effect diagram schematically showing the recognition of qualified products in Embodiment 1 of the present invention;

[0065] Figure 3 is an effect diagram schematically showing the recognition of unqualified products in Embodiment 1 of the present invention;

[0066] Figure 4 is a parameter summary diagram of the monitoring end schematically showing in Embodiment 1 of the present invention;

[0067] Figure 5 is a structural schematic diagram schematically showing the detection system for the processing quality of resolver wire harnesses based on machine vision in Embodiment 3 of the present invention. Detailed Embodiments

[0068] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts fall within the protection scope of the present invention.

[0069] Embodiment 1

[0070] As Figure 1 shown, this embodiment discloses a method for detecting the processing quality of a resolver wire harness based on machine vision, including:

[0071] S10: Obtain the image of the inner core wire head of the target resolver wire harness.

[0072] In this embodiment, the image of the inner core wire head refers to the image at the junction of the bare electric core and the rubber sleeve after the rubber sleeve is peeled off. The above-mentioned image of the inner core wire head can be obtained by a CCD camera.

[0073] S20: Use the edge detection method to identify all inner core edges in the image of the inner core wire head.

[0074] S30: Draw an identification frame corresponding to each inner core according to the inner core edge.

[0075] S40: Determine whether the inner core edge meets the following conditions:

[0076] Condition 1: Any inner core edge extends beyond the corresponding identification frame in the width direction;

[0077] Condition 2: The inclination angle of any inner core edge exceeds the inclination interval;

[0078] S50: If any condition is met, determine that the target resolver wire harness is unqualified.

[0079] S60: If all conditions are not met, determine that the target resolver wire harness is qualified.

[0080] As Figure 2 and Figure 3 shown, through the above steps S10 - S60, the method of the present invention relies on the edge detection method to adaptively identify the edges of the inner cores, and can automatically use the identification frames (i.e., the green or red frames in Figure 2 and 3 ) to frame the inner core edges and analyze whether the inner core edges extend beyond the identification frames in the width direction, and can also adaptively calculate the inclination angle of each inner core edge. The whole process does not require professionals to provide a teaching template according to the specifications or models of the target resolver wire harness, realizing intelligence and automation, and solving the problems of poor recognition reliability and low efficiency in the prior art.

[0081] In this embodiment, the above-mentioned inclination range is generally 87°-103° relative to the X-axis. In other embodiments, the above-mentioned inclination range can be adjusted adaptively according to actual situations.

[0082] Preferably, before the above-mentioned S10, the method of the present invention further includes:

[0083] S100: Filter the image of the inner core wire end.

[0084] S200: Binarize the filtered image of the inner core wire end.

[0085] In this embodiment, the above-mentioned filtering process can be performed based on a Gaussian filter, a median filter, or a bilateral filter, and the specific selection is made according to the image quality requirements.

[0086] Through the above step S100, the noise in the original image of the inner core wire end can be removed, and the image quality can be improved. And the above step S200 can convert the RGB image of the inner core wire end image into a black and white image to enhance the image features. In this way, the inner core edge features in the image are more prominent, the accuracy of edge detection is improved, which is beneficial to improving the detection efficiency of steps S20-S30.

[0087] Furthermore, after step S10, the method of the present invention further includes:

[0088] S300: Draw a detection frame for the image of the inner core wire end.

[0089] It should be noted that the above-mentioned detection frame (i.e., Figure 2 or the blue frame in 3) is a preset detection range. After obtaining the image of the inner core wire end, it can be adjusted appropriately according to the maximum width of the image of the inner core wire end, so that the recognition frame falls within the detection frame. In this way, the amount of data calculation can be reduced, and thus the recognition efficiency of the method of the present invention can be improved.

[0090] Furthermore, the edge recognition formula of the edge detection method in the above step S20 is:

[0091]

[0092]

[0093]

[0094] In the formula, is the image of the inner core wire end, is the convolution operation, is the operator matrix in the x-axis direction, is the operator matrix in the y-axis direction, is the gradient of the image of the inner core wire end in the x-axis direction, is the gradient of the inner core wire end image in the y-axis direction, is the pixel gradient of the coordinate (x, y) in the inner core wire end image; if , the pixel at the coordinate (x, y) is a non-edge pixel; if , the pixel at the coordinate (x, y) is an edge pixel; is the lower limit of the edge threshold.

[0095] In this embodiment, the above-mentioned is specifically:

[0096]

[0097] In this embodiment, the above-mentioned is specifically:

[0098]

[0099] In this embodiment, and can adopt the Sobel operator to achieve better operation effects. In other embodiments, and can also adopt the Prewitt operator or the Roberts operator to adapt to different application environments.

[0100] Preferably, after obtaining the edge pixels by using the edge recognition formula and the single-threshold discrimination method, the method of the present invention further includes:

[0101] S21: Connect multiple adjacent edge pixels to form multiple closed inner core edges.

[0102] S22: Screen out the inner core edges that are close to each other in the x-axis direction.

[0103] It should be explained that after determining the edge pixels by the single threshold, the shortest path traversal method can be used to connect multiple adjacent edge pixels to form multiple closed inner core edges. Generally, if the image features are obvious and the indicators of the inner core are normal, there will be no situation where two inner core edges are close to each other, and whether step S22 is executed will not affect the final result. In abnormal situations (such as severe bending of the inner core, foreign object coverage, or dirty CCD lens, etc.), there may be a situation where two inner core edges are close to each other. At this time, the close edges are screened out, and only the complete and independent inner core edges are retained for judgment. It should be further supplemented that the above-mentioned edge screening may cause some recognition frames to be missing, resulting in the number of output wire cores being less than the reference value. At this time, it can also be determined that the target resolver harness is a non-conforming product, and a "rejudgment" warning is sent to the monitoring terminal to prompt professionals to check the equipment.

[0104] Preferably, to determine whether the edges of two inner cores are close, the following method can be adopted:

[0105] Calculate the centroid spacing distance between the edges of the two inner cores.

[0106] If the centroid spacing distance is lower than a preset spacing threshold, it is determined that the edges of the two inner cores are close.

[0107] Compared with the prior art, the method of the present invention can overcome the recognition interference caused by external factors and improve the recognition accuracy of the method of the present invention.

[0108] Further, the above step S30 is specifically as follows:

[0109] S31: According to the model of the target resolver harness, obtain the inner diameter parameter of each inner core as the selection width of the recognition frame.

[0110] S32: Take the single cutting stroke of the wire cutting device as the selection length of the recognition frame.

[0111] S33: According to the selection width and selection length, draw the corresponding recognition frame with the geometric center of the corresponding inner core edge as the reference center.

[0112] It should be noted that the wire cutting device is a device dedicated to wire harness cutting. It mainly consists of a wire cutting knife and a servo unit for driving the knife to move back and forth. During the process of the servo unit driving the knife to move back and forth, it will automatically record each cutting stroke. When drawing the recognition frame corresponding to each inner core, the database information or device process parameters can be called, and the inner diameter parameter of the inner core corresponding to the model of the target resolver harness is used as the selection width of the recognition frame, and then the single cutting stroke is used as the selection length of the recognition frame, so as to realize the adaptive adjustment of the recognition frame size. This method does not require manual teaching and improves the recognition efficiency of the method of the present invention. After determining the size of the recognition frame, draw the recognition frame according to the centroid of each inner core edge, so as to realize a more accurate recognition frame selection.

[0113] It should be added that a certain compensation value can be added to the width or length of the above recognition frame to improve the general performance of the method of the present invention.

[0114] Further, the calculation formula for the tilt angle in the above step S40 is:

[0115]

[0116] In the formula, represents the tilt angle of the j th inner core edge length direction relative to the X-axis, arctan() is the arctangent function, is the j th inner core edge iThe ordinate of an edge point is the j ordinate of the midpoint of the top edge of the is the j total number of edge points of the inner core edge is the j abscissa of the i th edge point of the inner core edge is the j abscissa of the midpoint of the top edge of the inner core edge is the upper limit value of the serial number i of

[0117] It should be noted that the above formula for calculating the tilt angle can calculate the centroid coordinates of each inner core edge according to the coordinates of each pixel point in the inner core edge, and then use the centroid coordinates of the inner core edge and the midpoint coordinates of the corresponding fixed edge of the inner core edge to calculate the slopes of the two coordinate points relative to the X-axis. Finally, the tilt angles of each inner core edge are calculated using the arctangent function respectively. Compared with the prior art, the above method is more accurate in calculating the tilt angle of the inner core, improving the reliability of the method of the present invention.

[0118] Furthermore, as Figure 4 shown, after the above step S40, the method of the present invention further includes:

[0119] S400: Output the number of recognition frames as the number of inner cores to the monitoring end.

[0120] S401: Calculate and output the width data of all inner core edges to the monitoring end.

[0121] S402: Calculate and output the length data of all inner core edges to the monitoring end.

[0122] S403: Calculate and output the area data of the corresponding inner core to the monitoring end according to the width data and the length data.

[0123] Through the above steps S400 - S403, after the recognition is completed, the method of the present invention will output data such as the number of inner cores, the width of the inner cores, the length of the inner cores, and the area of the inner cores to the monitoring end, so as to facilitate professional personnel to conduct rejudgment and evaluation. After the above data is output, it will be transmitted to the storage medium for storage, thereby realizing the archiving of product information.

[0124] Embodiment 2

[0125] Compared with Embodiment 1, this embodiment uses a dual-threshold determination method to determine edge pixels, specifically:

[0126] If , determine that the pixel is a weak edge pixel.

[0127] If , determine that the pixel is a strong edge pixel.

[0128] Among them, is the strong edge threshold.

[0129] By using the double-threshold determination method to distinguish strong edge pixels and weak edge pixels, and preferentially connecting the strong edge pixels, this method can more accurately identify the actual edge of the inner core, reduce the possibility of the closed inner core edge showing an irregular shape, and make the detection result of the method of the present invention more reliable.

[0130] Furthermore, in order to implement the inner core crack detection function, after drawing the recognition frame corresponding to each inner core, the method of the present invention further includes:

[0131] Inside the recognition frame, connect multiple adjacent weak edge pixels to form an edge to be measured.

[0132] Judge whether the length of the edge to be measured exceeds the crack depth threshold.

[0133] If so, judge that the target resolver harness is unqualified.

[0134] It should be explained that during the process of cutting the rubber sleeve of a certain inner core, if the cutting tool becomes dull or the tool pressure is too high, scratches or cracks often will be left on the inner core at the starting point of the cut. If the depth of this part of the scratches or cracks is relatively shallow, it has little impact on the electrical conductivity of the inner core. However, once the above-mentioned scratches or cracks reach a certain depth, it will greatly reduce the electrical conductivity of the inner core, and tiny electric arcs will be generated in a high-frequency current environment, which has a certain degree of danger.

[0135] Compared with the prior art, the above technical solution utilizes the characteristic that the gray value of the inner core scratch or crack is relatively low in the image, and uses the weak edge to judge whether the crack depth exceeds the depth threshold. This method can quickly identify the inner core with obvious cracks, thereby improving the reliability of the quality inspection of the resolver inner core.

[0136] Embodiment III

[0137] As Figure 5 shown, on the basis of Embodiment I, this embodiment provides a resolver harness processing quality detection system based on machine vision, including a processor and a memory. The above memory stores computer program instructions, and when the computer program instructions are executed by the processor, the resolver harness processing quality detection method described in Embodiment I is implemented.

[0138] The inventive system further includes other components well known to those skilled in the art such as a communication interface, and its settings and functions are known in the art, so they will not be described in detail here.

Claims

1. A method for detecting the processing quality of resolver wire harnesses based on machine vision, characterized in that, Including: Filter the image of the inner core wire head to obtain the image of the inner core wire head of the target resolver harness; use the edge detection method to identify all inner core edges in the image of the inner core wire head; according to the inner core edges, draw an identification frame corresponding to each inner core, specifically: Obtain the inner diameter parameter of each inner core as the selection width of the identification frame according to the model of the target resolver harness; Take the single wire cutting stroke of the wire cutting device as the selection length of the identification frame; According to the selection width and the selection length, draw the corresponding identification frame with the geometric center of the corresponding inner core edge as the reference center; Distinguish strong edge pixels and weak edge pixels by the double-threshold determination method, and connect the strong edge pixels to accurately identify the actual edge of the inner core; Inside the identification frame, connect multiple adjacent weak edge pixels to form an edge to be measured; Judge whether the length of the edge to be measured exceeds the crack depth threshold; If so, judge that the target resolver harness is unqualified; Judge whether the inner core edge meets the following conditions: any inner core edge exceeds the corresponding identification frame in the width direction; the inclination angle of any inner core edge exceeds the inclination interval; where the calculation formula for the inclination angle is: In the formula, represents the inclination angle of the j th inner core edge length direction relative to the X-axis, and arctan() is the arctangent function, is the j th ordinate of the i th edge point of the inner core edge, is the j th ordinate of the midpoint of the top edge of the inner core edge, is the j th total number of edge points of the inner core edge, is the j th abscissa of the i th edge point of the inner core edge, is the j th abscissa of the midpoint of the top edge of the inner core edge, is the upper limit value of the serial number i ; If any condition is met, judge that the target resolver harness is unqualified.

2. The method for detecting the processing quality of a resolver wire harness based on machine vision according to claim 1, wherein The edge recognition formula of the edge detection method is: In the formula, is the inner core wire head image, is the convolution operation, is the operator matrix in the x-axis direction, is the operator matrix in the y-axis direction, is the gradient of the inner core wire head image in the x-axis direction, is the gradient of the inner core wire head image in the y-axis direction, is the pixel gradient of the coordinate (x, y) in the inner core wire head image; if , the pixel at the coordinate (x, y) is a non-edge pixel; if , the pixel at the coordinate (x, y) is an edge pixel; is the lower limit of the edge threshold.

3. The method for detecting the processing quality of a resolver wire harness based on machine vision according to claim 2, wherein The edge pixels include weak edge pixels and strong edge pixels. If , it is determined that the pixel with coordinates (x, y) is an edge pixel, specifically: If , determine that the pixel with coordinates (x, y) is a weak edge pixel; If , determine that the pixel with coordinates (x, y) is a strong edge pixel; Among them, is the strong edge threshold, and the strong edge pixels are the edge pixels that are preferentially connected by pixels.

4. The method for detecting the processing quality of a resolver wiring harness based on machine vision according to claim 2, wherein, After obtaining the edge pixels, the method further includes: Connect multiple adjacent edge pixels to form multiple closed inner core edges; Sift out the inner core edges that are close to each other in the x-axis direction.

5. The method for detecting the processing quality of a resolver wire harness based on machine vision according to claim 1, wherein After obtaining the image of the inner core wire head of the target resolver harness, the method further includes: Draw a detection frame for the image of the inner core wire head; Wherein, the identification frame is located inside the detection frame.

6. The method for detecting the processing quality of a resolver wire harness based on machine vision according to claim 1, characterized in that After drawing the identification frame of each inner core, the method further includes: Output the number of the identification frames as the number of inner cores to the monitoring terminal; Calculate and output the width data of all inner core edges to the monitoring terminal; Calculate and output the length data of all inner core edges to the monitoring terminal; Calculate and output the area data of the corresponding inner core to the monitoring terminal according to the width data and the length data.

7. The method for detecting the processing quality of a resolver wire harness based on machine vision according to claim 1, wherein After obtaining the image of the inner core wire head of the target resolver harness, the method further includes: Filter the image of the inner core wire head.

8. A resolver wire harness processing quality detection system based on machine vision, characterized in that, Including a processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the method for detecting the processing quality of the resolver harness based on machine vision according to any one of claims 1-7 is implemented.

Images