A welding method and system suitable for multi-dimensional force sensors
Through tilted line laser scanning and image recognition technology, the welding point position is automatically calculated and the parameters are adjusted, which solves the problems of low welding efficiency and unstable quality of multi-dimensional force sensors and realizes an efficient and accurate welding process.
Patent Information
- Application Number
- CN202511065830.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-31
- Publication Date
- 2025-10-24
- Estimated Expiration
- 2045-07-31
AI Technical Summary
Existing multi-dimensional force sensors have low welding efficiency, low quality stability, cumbersome operation, and are prone to errors.
Adopting tilted line laser scanning multi-dimensional force sensor, the camera takes images to identify the welding area, calculate the welding point position and automatically adjust the welding parameters to achieve fast and accurate welding.
The efficiency and quality stability of multi-dimensional force sensor welding are improved, manual intervention is reduced, and the operation process is simplified.
Smart Images

Figure CN120551663B_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of intelligent welding, in particular to a welding method and system suitable for a multi-dimensional force sensor. BACKGROUND
[0002] The structure of the multi-dimensional force sensor is shown in the figure, which comprises an outer annular fixing ring 01 and a mounting block 02 located in the middle of the fixing ring, a plurality of connecting arms 03 are located between the fixing ring 01 and the mounting block 02, connecting the fixing ring 01 and the mounting block 02 into one body, and the top and bottom surfaces of the connecting arms 03 will be attached with deformation sensors. Figure 1 The mounting block 02 is used to connect with the measured object, when the measured object is stressed, it will drive the plurality of connecting arms 03 to produce different deformations, and through the deformation sensors, the size and direction of the deformation of each connecting arm 03 can be detected, so that the size and direction of the stress of the mounting block 02 and the measured object can be calculated.
[0003] The deformation sensor and the data transmission line need to be connected through welding, and the positioning of each welding point is very important during welding. The existing method usually fixes the multi-dimensional force sensor with a positioning tool, and then welds it. This positioning method of repeated installation and disassembly is complicated to operate, and each time only the same side of one sensor can be welded, and the parameters of the welding equipment need to be adjusted frequently, so the welding efficiency is low, and the error probability is also increased due to the numerous steps in the welding process, resulting in low stability of the welding quality. SUMMARY
[0004] The present application provides a welding method and system suitable for a multi-dimensional force sensor, which can effectively solve the problems of low welding efficiency and low stability of welding quality in the background art.
[0005] The welding method provided by the present application comprises the following steps:
[0006] Set a welding area, and use a camera to take pictures from a top view angle of the welding area;
[0007] A plurality of identification surfaces are arranged on the vertical surface of the mounting block of the multi-dimensional force sensor, and the adjacent identification surfaces are at a certain angle;
[0008] Put the multi-dimensional force sensor into the welding area, and irradiate the linear laser from top to bottom at an angle, and move at a constant speed through the welding area; when the linear laser moves, the camera takes a picture every set time;
[0009] Identify the laser pixel points in each picture, which will change into a shape of multiple line segments after being irradiated on the multi-dimensional force sensor, and determine the grouping of each multi-dimensional force sensor and its corresponding contour according to the shape of the laser;
[0010] The profiles of the respective multi-dimensional force sensors in the corresponding images are grouped to form profile images of the respective multi-dimensional force sensors, and the center points of the respective multi-dimensional force sensors and the face profiles are calculated according to the profile images, so as to determine the welding point positions and the welding action parameters of the respective multi-dimensional force sensors;
[0011] According to the welding point positions and the welding action parameters, the respective welding points are welded.
[0012] Further, the laser region in each image is identified, specifically including:
[0013] The line laser moving direction is taken as the y-axis, and the line laser width direction is taken as the x-axis;
[0014] The gray value of the pixel point in the xth row and yth column in the image is denoted as G(x, y);
[0015] The average value Gave(x, y) of the gray values of the multiple pixel points adjacent to the pixel point in the y-axis direction is obtained;
[0016] If the difference between G(x, y) and Gave(x, y) is greater than a set threshold value, the pixel point is recorded as a laser pixel point.
[0017] Further, before the average value Gave(x, y) is calculated, the maximum value of the gray values of the multiple pixel points adjacent to the pixel point in the y-axis direction is removed.
[0018] Further, the determination of the respective multi-dimensional force sensors and the corresponding profile groups is specifically as follows:
[0019] In each image, the line laser moving direction is taken as the y-axis, and the line laser width direction is taken as the x-axis;
[0020] The theoretical position of the line laser irradiated in the welding area is denoted as y=y0;
[0021] The laser pixel points with the y value within a set range from y0 are found out, and these laser pixel points are grouped, so that the laser pixel points in the group are adjacent in the x-axis direction, and the number of the laser pixel points in the group is greater than a set threshold value, thereby forming baseline groups;
[0022] When the baseline groups are only one, the calculation of the next image is directly performed;
[0023] When the baseline groups are multiple, the laser pixel points between each adjacent two baseline groups are combined and recorded as the profile group of the same multi-dimensional force sensor;
[0024] If there is no image with only one baseline group in the adjacent multiple images, the profile groups between the adjacent two baseline groups in the multiple images are recorded as the profile group of the same multi-dimensional force sensor corresponding to the previous image; otherwise, the profile groups are recorded as the profile group of a new multi-dimensional force sensor.
[0025] Further, the profile image of each multi-dimensional force sensor is specifically:
[0026] For each image, the laser pixel points in the profile grouping of the same multi-dimensional force sensor are grouped, so that the y value difference of each group of laser pixel points is within a set range, forming a contour grouping;
[0027] The number of pixel points in each contour grouping is counted, and the contour grouping with a number less than a set threshold is removed;
[0028] The coordinates of the pixel points at both ends of each contour grouping are taken as the profile coordinates;
[0029] Combining all the profile coordinates of the same multi-dimensional force sensor in all images, the profile image of the multi-dimensional force sensor can be obtained.
[0030] Further, the center point and the identification surface profile of each multi-dimensional force sensor are specifically:
[0031] For a multi-dimensional force sensor, a square minimum bounding box of the profile image of the multi-dimensional force sensor is calculated, and the center of the minimum bounding box is the center point of the multi-dimensional force sensor;
[0032] The distance between the center point and the points in the contour grouping corresponding to the mounting block of the multi-dimensional force sensor is calculated, the radius of the mounting block is determined according to the mode of the distance, and the points with a distance not equal to the radius are filtered out, so as to determine the position of each identification surface profile.
[0033] Further, the welding action parameters of each multi-dimensional force sensor are specifically:
[0034] A comparison library of welding action parameters is set;
[0035] For a multi-dimensional force sensor, any one image with the most contour groupings is selected, the difference between the y value of each contour grouping and y0 is calculated, and it is arranged from small to large as △y1, △y2……△yq;
[0036] According to the value of q and the values of △y1, △y2……△yq, the corresponding welding action parameters are searched in the comparison library.
[0037] Further, before welding, the line laser is used to irradiate and take images without placing the multi-dimensional force sensor, and it is confirmed whether the laser pixel points in each image are on the same straight line, if yes, the subsequent normal welding is performed, if not, the machine is stopped and an alarm is issued.
[0038] Further, the irradiation direction of the line laser is 45° with the horizontal plane.
[0039] The application further provides a welding system suitable for the multi-dimensional force sensor, which comprises:
[0040] A device rack is provided with a welding platform, and the welding platform is provided with a welding area;
[0041] A camera is fixedly arranged above the welding area;
[0042] A laser emitter is arranged above the welding area and irradiates linear laser on the welding area;
[0043] A conveying assembly is arranged to drive the multi-dimensional force sensor to enter and move out of the welding area;
[0044] A welding arm is arranged to move above the welding area and is used for welding the multi-dimensional force sensor;
[0045] A processor is arranged to realize the welding method suitable for the multi-dimensional force sensor.
[0046] The technical scheme of the application can realize the following technical effects:
[0047] The method can accurately display the contour shape of each multi-dimensional force sensor in the welding area through the inclined linear laser, can accurately calculate the welding point position and the placement state corresponding to each multi-dimensional force sensor, and can realize rapid and accurate welding of each multi-dimensional force sensor. The intervention demand of personnel can be greatly reduced during the welding process, and the number and placement of the multi-dimensional force sensors are relatively free during each welding, which greatly simplifies the operation of personnel, thereby effectively improving the production efficiency. BRIEF DESCRIPTION OF DRAWINGS
[0048] In order to more clearly illustrate the technical scheme in the embodiments of the application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or the prior art description. Obviously, the drawings in the following description are only some embodiments described in the application, and other drawings can be obtained by those skilled in the art without creative labor.
[0049] Figure 1 It is a structural schematic diagram of the multi-dimensional force sensor in the application;
[0050] Figure 2 It is a step flow chart of the welding method suitable for the multi-dimensional force sensor in the application;
[0051] Figure 3 It is a schematic diagram of the multi-dimensional force sensor when the top surface of the multi-dimensional force sensor is irradiated by the laser;
[0052] Figure 4 It is a schematic diagram of the multi-dimensional force sensor when the bottom surface of the multi-dimensional force sensor is irradiated by the laser;
[0053] Fig. 01, fixed ring; 02, mounting block; 03, connecting arm. DETAILED DESCRIPTION
[0054] The basic principles and main features of the technical solutions of the present application will be described below in combination with the accompanying drawings in the embodiments of the present application. The description will be more intuitive through one or more embodiments, and the described embodiments are only some of the embodiments of the present application, not all the embodiments.
[0055] In the description of the present application, the words indicating the orientation or positional relationship (such as up, down, left, right, etc.) are based on the orientation shown in the drawings or some conventional positional relationship, only for the convenience of describing the present application and simplifying the description, and are not indicative or suggestive of the features indicated must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application.
[0056] The present application provides a welding method suitable for multi-dimensional force sensors, mainly including hardware setting step, image acquisition step, welding point calculation step and welding step, the specific process of each step and the overall process are as shown in Figure 2 , specifically including:
[0057] Hardware setting step: set the welding area, the welding area is the area range that the welding head can reach, all multi-dimensional force sensors that need to be welded need to be placed in the welding area;
[0058] Use a camera to take a top-down view of the welding area, so that the camera can obtain a top-down image of the entire welding area. When the welding area is large, multiple cameras can be used to splice the images taken by them;
[0059] A plurality of identification surfaces are arranged on the vertical surface of the mounting block 02 of the multi-dimensional force sensor, the identification surfaces are in a planar structure and are obviously distinguished from the circular structure on the outer side of the mounting block 02, the adjacent identification surfaces are at a certain angle, which is convenient for subsequent identification of the identification surfaces; the number of identification surfaces corresponds to the number of connecting arms 03, each identification surface corresponds to a connecting arm 03, and the identification surface is used to correspond to the welding point position on the connecting arm 03.
[0060] Image acquisition step: place one or more multi-dimensional force sensors in the welding area, and arrange the multi-dimensional force sensors in an array as much as possible; the line laser is inclined from top to bottom and moves uniformly through the welding area, and the inclination angle of the line laser is aligned with the moving direction of the line laser; when the line laser moves, the camera takes a picture every set time.
[0061] Welding point calculation step: identify the laser pixel points in each image, such as Figure 3 and Figure 4As shown, due to the different heights of the parts on the multi-dimensional force sensor, the inclined laser light will become a shape of multiple lines when it is irradiated, and the height difference between different parts will also cause different intervals between the multiple lines, so the position of each multi-dimensional force sensor and its corresponding profile group (i.e. the outer contour point set of each part of the multi-dimensional force sensor) at this moment can be determined according to the shape of the laser, and it can be distinguished which side of the multi-dimensional force sensor is upward; there can be multiple profile groups of multi-dimensional force sensors on each image, and after calculation, they need to be corresponded, such as the profile groups of multi-dimensional force sensors numbered 1, 2 and 3 from left to right on the 10th image, and the correspondence needs to be recorded:
[0062] Let the set of profile images of the No. 1 multi-dimensional force sensor be L1 = {(x1101, y1101), (x1102, y1102) …} (the points in the set are the profile groups of the No. 1 multi-dimensional force sensor);
[0063] Let the set of profile images of the No. 2 multi-dimensional force sensor be L2 = {(x2101, y2101), (x2102, y2102) …} (the points in the set are the profile groups of the No. 2 multi-dimensional force sensor);
[0064] Let the set of profile images of the No. 3 multi-dimensional force sensor be L3 = {(x3101, y3101), (x3102, y3102) …} (the points in the set are the profile groups of the No. 3 multi-dimensional force sensor);
[0065] After calculating an image, if there are corresponding profile groups of multi-dimensional force sensors on the image, the new profile group is added to the set of profile images of the corresponding multi-dimensional force sensor, and finally the profile groups of each multi-dimensional force sensor in the corresponding image are merged into a large set, such as L1 = {(x1101, y1101), (x1102, y1102), …, (x1111, y1111), (x1112, y1112), …, (x1131, y1131), (x1132, y1132), …}, so that the profile images of each multi-dimensional force sensor are formed;
[0066] According to the profile images, the center points of each multi-dimensional force sensor and the identification of the face profile are calculated, and then the welding point position is determined, and the welding action parameters of each multi-dimensional force sensor are determined, which mainly include the descending height of the welding head and the welding force, etc. These parameters will have different values when corresponding to different models of multi-dimensional force sensors or the front and back of the same multi-dimensional force sensor, and need to be determined according to the size of the multi-dimensional force sensor calculated on the graph.
[0067] Soldering step: according to the soldering position and the soldering action parameters, each soldering point is soldered.
[0068] It can be seen that the method can accurately display the contour shape of each multi-dimensional force sensor in the welding area by the inclined line laser, can accurately calculate the corresponding welding point position and placement state of each multi-dimensional force sensor, and can realize rapid and accurate welding of each multi-dimensional force sensor. The intervention demand of personnel can be greatly reduced during the welding process, and the number and placement of the multi-dimensional force sensors are relatively free each time of welding, greatly simplifying the operation of personnel, thereby effectively improving the production efficiency.
[0069] Preferably, the laser region in each image is identified, specifically including:
[0070] The moving direction of the line laser is the y-axis (refer to the vertical direction of Figure 3 and Figure 4 ), and the width direction of the line laser is the x-axis (refer to the horizontal direction of Figure 3 and Figure 4 );
[0071] For a pixel point in the xth row and yth column in the image, its gray value is recorded as G(x, y). The pixel origin in the image can be freely set, but cannot be changed after being set;
[0072] In general cases, the welding area will be blocked, so that there is only one light source, i.e. the line laser, in the welding area. Therefore, in the regions where the line laser does not irradiate, the gray values in the image are all dark, and the regions where the line laser irradiates are displayed as bright in the image. By finding the pixel points that are brighter than the surrounding pixel points, these pixel points are the regions irradiated by the line laser, specifically:
[0073] Since the moving direction of the line laser is the y-axis, there is and only one bright pixel point in each y column in the image. Therefore, the pixel points in the y-axis direction are most suitable as the basis for comparison, and the average value Gave(x, y) of the gray values of the multiple pixel points adjacent to the pixel point in the y-axis direction is obtained. If the difference between G(x, y) and Gave(x, y) is greater than a set threshold value, the pixel point is recorded as a laser pixel point.
[0074] Before calculating the average value Gave(x, y), the maximum value of the gray values of the multiple pixel points adjacent to the pixel point in the y-axis direction is preferably removed, which is the bright point, to avoid the interference of the laser bright point on the average value Gave(x, y) and affect the accuracy of the judgment.
[0075] Preferably, the determination of each multi-dimensional force sensor and the corresponding contour grouping in each image is specifically:
[0076] The moving direction of the line laser is the y-axis (refer to the vertical direction of Figure 3 andFigure 4 The vertical direction of the line laser is the y-axis (reference Figure 3 The horizontal direction of the line laser is the x-axis (reference Figure 4 The horizontal direction of the line laser is the x-axis (reference
[0077] If there is no object in the welding area, the line laser will form a straight line parallel to the x-axis with a constant y value, which can be calculated according to the position of the line laser emitter, which is the theoretical position of the line laser. The theoretical position of the line laser irradiating in the welding area is y=y0. Note that the y0 value of each image is different, and y0 needs to be updated before calculating a new image.
[0078] When the line laser is blocked by the multi-dimensional force sensor, the part of the area at the theoretical position y=y0 will move up, as shown in Figure 3 and Figure 4 It can be seen that each moved interval corresponds to a multi-dimensional force sensor. Therefore, by finding the laser pixel points with a y value within a set range from y0 and grouping these laser pixel points, the laser pixel points in each group are still adjacent in the x-axis direction (i.e., the part of the laser directly irradiating the bottom of the welding area on both sides of the multi-dimensional force sensor), and the number of laser pixel points in each group needs to be greater than a set threshold, the very short laser that passes through the hollow part of the multi-dimensional force sensor and irradiates to the bottom can be removed, thereby forming one or more baseline groups.
[0079] When the baseline group of this image is only one, it means that there is no multi-dimensional force sensor at the irradiation position, and the laser area in the figure is a straight line y=y0. Therefore, the calculation of the next image can be directly skipped.
[0080] When the baseline group of this image has multiple, it means that there is a multi-dimensional force sensor at the irradiation position. The combination of laser pixel points between each adjacent baseline group is recorded as the contour group of the same multi-dimensional force sensor. For example, if there are three baseline groups numbered 1, 2, and 3 from left to right in the image, there will be two laser areas moving upward, which are the contour group of the No. 1 multi-dimensional force sensor and the contour group of the No. 2 multi-dimensional force sensor, respectively. In this way, the contour groups in each image can be corresponded to the multi-dimensional force sensors.
[0081] Of course, if you want to obtain the final contour of each multi-dimensional force sensor, you also need the contour group information on other images. For the corresponding relationship between each image, there is the following finding method:
[0082] If there is no image with only one baseline group in the adjacent images, it means that the lasers of these images are scanning the same multi-dimensional force sensor, and the contour group between the adjacent two baseline groups in the multiple images can be recorded as the contour group of the same multi-dimensional force sensor corresponding to the previous image; otherwise, it is recorded as a new contour group of a multi-dimensional force sensor; for example:
[0083] Now that the three baseline groups numbered 1, 2, and 3 from left to right have been calculated in the fifth image, there will be corresponding contour groups of the first multi-dimensional force sensor and the second multi-dimensional force sensor;
[0084] Now calculate the adjacent sixth image, and find that there are still three baseline groups numbered 1, 2, and 3 from left to right, and there is no image with only one baseline group, so the two contour groups between the three baseline groups are still corresponding to the first multi-dimensional force sensor and the second multi-dimensional force sensor, respectively;
[0085] From the fifth image to the fifteenth image, there are still three baseline groups numbered 1, 2, and 3 from left to right, and there is no image with only one baseline group, so the two contour groups in these images are still corresponding to the first multi-dimensional force sensor and the second multi-dimensional force sensor, respectively;
[0086] Now the sixteenth image has an image with only one baseline group, which means that the contour groups of the first multi-dimensional force sensor and the second multi-dimensional force sensor have been found; when the multi-dimensional force sensor is placed, it will be noted that there needs to be a space between them, so there will be multiple images with only one baseline group.
[0087] If the calculation reaches the twentieth image, there are three baseline groups numbered 1, 2, and 3 from left to right, so the two contour groups are now new contour groups of multi-dimensional force sensors, and a new number needs to be set, i.e., the two contour groups correspond to the third multi-dimensional force sensor and the fourth multi-dimensional force sensor, respectively.
[0088] Preferably, the contour image of each multi-dimensional force sensor is specifically formed as follows:
[0089] For each image, group the laser pixel points in the contour group of the same multi-dimensional force sensor so that the y value difference of each group of laser pixel points is within a set range to form a contour group, such as Figure 3 and Figure 4 In the above, the surface of the fixed ring 01 is a height, so the laser irradiated on the fixed ring 01 will be on a straight line; the surface of the mounting block 02 is a height, so the laser irradiated on the mounting block 02 will be on a straight line;
[0090] The number of pixels in each contour group is counted, and if the number of pixels in the contour group is too small, it means that these places are likely to be laser pixels formed by the laser irradiated into the hole of the fixing ring 01 and the mounting block 02, and the contour group composed of these laser pixels needs to be removed, and only the main contour group shown in Figs. 1-3 is left. Figure 3 and Figure 4 ;
[0091] The coordinates of the pixels at both ends of each contour group are taken as the contour coordinates;
[0092] All contour coordinates of the same multi-dimensional force sensor in all images are combined, and the contour image of the multi-dimensional force sensor can be obtained.
[0093] Preferably, the center point and the identification surface contour of each multi-dimensional force sensor are calculated as follows:
[0094] For a multi-dimensional force sensor, a square minimum bounding box of the contour image of the multi-dimensional force sensor is calculated, that is, the center of the multi-dimensional force sensor is determined by the outermost contour of the fixing ring 01, and the center of the minimum bounding box is the center point of the multi-dimensional force sensor. The outer side of the fixing ring 01 is not provided with an identification surface, so the information of the circular shape can be retained as much as possible, and the found center point is more accurate;
[0095] Then, the distance between the center point and all points in the contour group corresponding to the mounting block 02 of the multi-dimensional force sensor is calculated. Since the mounting block 02 has the same circular arc shape except the identification surface, the distance between most points on the mounting block 02 and the center point is equal to the radius. Therefore, the radius of the mounting block 02 can be determined according to the mode of the distance, and the points with a distance not equal to the radius are screened out, which are the points on the identification surface, so as to determine the position of each identification surface contour.
[0096] Preferably, the welding action parameters of each multi-dimensional force sensor are determined as follows:
[0097] A comparison library of welding action parameters is set, and the physical parameters and welding action parameters of each multi-dimensional force sensor are one-to-one corresponding; for example, record:
[0098] The physical parameters of the A-type multi-dimensional force sensor are: the height difference between the fixing ring 01 and the bottom surface, and the height difference between the mounting block 02 and the bottom surface, which are arranged from large to small and recorded as △y0a1 and △y0a2; and the corresponding welding action parameters are: welding head drop La and welding pressure Pa;
[0099] The physical parameters of the B-type multi-dimensional force sensor are: the height difference between the fixed ring 01 and the bottom surface, and the height difference between the mounting block 02 and the bottom surface, which are arranged from large to small and recorded as △y0b1 and △y0b2; and the corresponding welding action parameters are: the welding head descending Lb and the welding pressure Pb.
[0100] …
[0101] In this way, the physical parameters and the welding action parameters of all multi-dimensional force sensors are recorded correspondingly, so that a comparison library of welding action parameters is formed.
[0102] During calculation, for a multi-dimensional force sensor, any image with the largest number of contour groupings is selected, the difference between the y value of each contour grouping and y0 is calculated, and the difference is arranged from small to large and recorded as △y1, △y2, …, △yq.
[0103] According to the value of q and the values of △y1, △y2, …, △yq, the corresponding welding action parameters are searched in the comparison library.
[0104] For example: if the height difference between the fixed ring 01 and the bottom surface, and the height difference between the mounting block 02 and the bottom surface are arranged from large to small as △y1=△y0a1 and △y2=△y0a2, respectively, it indicates that the multi-dimensional force sensor is of type A, and when welding the welding point on the multi-dimensional force sensor, the welding action parameters of the welding head descending La and the welding pressure Pa are required to be used for welding operation.
[0105] Preferably, before welding, the linear laser is used to irradiate and take images without placing the multi-dimensional force sensor, and it is confirmed whether the laser pixel points in each image are on the same straight line, if yes, the subsequent normal welding is performed, and if not, it indicates that the bottom surface of the welding area is uneven, and the machine needs to be stopped to issue an alarm to remind personnel to debug.
[0106] Preferably, the irradiation direction of the linear laser forms an angle of 45° with the horizontal plane, so that when the multi-segment laser pixel points are formed, the distance between the laser pixel points in the y-axis direction is directly equal to the actual height difference, so that calculation is not required, and calculation is facilitated.
[0107] The application also provides a welding system suitable for multi-dimensional force sensors, which comprises:
[0108] The device rack is provided with a welding platform, and the welding platform is provided with a welding area;
[0109] The camera is fixedly arranged above the welding area;
[0110] The laser emitter slides above the welding area and irradiates linear laser on the welding area;
[0111] The conveying assembly is used to drive the multi-dimensional force sensor to enter and exit the welding area.
[0112] a welding arm, which is moved over the welding area, and which has a welding head that is movable up and down, for welding the multi-dimensional force sensor;
[0113] a processor, such as a computer or a PLC, for implementing the above-mentioned welding method for a multi-dimensional force sensor.
[0114] The above shows and describes the basic principles, main features and advantages of the present application. Those skilled in the art should understand that the present application is not limited to the above-mentioned embodiments, and the above-mentioned embodiments and descriptions in the specification are only to illustrate the principles of the present application. Without departing from the spirit and scope of the present application, various changes and improvements can be made to the present application, and these changes and improvements all fall within the scope of the claimed present application. The scope of protection of the present application is defined by the appended claims and their equivalents.
Claims
1. A welding method suitable for a multi-dimensional force sensor, characterized by the steps of The application relates to a welding point position recognition method and device. The method comprises the following steps: Setting a welding area, and a camera shoots the welding area from a top view; A plurality of identification surfaces are arranged on the vertical surface of a mounting block of a multi-dimensional force sensor, and the adjacent identification surfaces are arranged at a certain angle; The multi-dimensional force sensor is placed in the welding area, a linear laser is obliquely irradiated from top to bottom, and the linear laser moves at a constant speed through the welding area; When the linear laser moves, the camera shoots an image every set time; The laser pixel points in each image are identified, the laser irradiated on the multi-dimensional force sensor becomes a shape of a plurality of line segments, the multi-dimensional force sensor and the corresponding profile grouping are determined according to the laser shape; The profile groupings of each multi-dimensional force sensor in the corresponding images are combined to form a profile image of each multi-dimensional force sensor, the center point of each multi-dimensional force sensor and the identification surface profile are calculated according to the profile image, the welding point position is determined, and the welding action parameters of each multi-dimensional force sensor are determined; According to the welding point position and the welding action parameters, each welding point is welded; The method for identifying the laser region in each image comprises the following steps: The moving direction of the linear laser is taken as the y-axis, and the width direction of the linear laser is taken as the x-axis; The gray value of the pixel point in the xth row and the yth column of the image is recorded as G (x, y); The average value Gave (x, y) of the gray values of a plurality of adjacent pixel points along the y-axis direction is obtained; If the difference between G (x, y) and Gave (x, y) is greater than a set threshold value, the pixel point is recorded as a laser pixel point; The determination of each multi-dimensional force sensor and the corresponding profile grouping comprises the following steps: In each image, the moving direction of the linear laser is taken as the y-axis, and the width direction of the linear laser is taken as the x-axis; The theoretical position of the linear laser irradiated in the welding area is y=y0; The laser pixel points with the y value within a set range from y0 are found out, and the laser pixel points are grouped, so that the laser pixel points in each group are adjacent in the x-axis direction, and the number of the laser pixel points in each group is greater than a set threshold value, thereby forming a baseline grouping; When the baseline grouping has only one, the calculation of the next image is directly performed; When the baseline grouping has a plurality of groups, the laser pixel points between each two adjacent baseline groupings are combined as the profile grouping of the same multi-dimensional force sensor; If there is no image with only one baseline grouping in adjacent images, the profile groupings between the two adjacent baseline groupings in the plurality of images are recorded as the profile grouping of the same multi-dimensional force sensor corresponding to the last image; otherwise, the profile groupings are recorded as the profile grouping of a new multi-dimensional force sensor; The formation of the profile image of each multi-dimensional force sensor comprises the following steps: For each image, the laser pixel points in the profile grouping of the same multi-dimensional force sensor are grouped, so that the y value difference of each group of laser pixel points is within a set range, thereby forming a contour line grouping; The number of pixel points in each contour line grouping is counted, and the contour line grouping with a number less than a set threshold value is removed; The coordinates of the pixel points at both ends of each contour line grouping are taken as contour coordinates; All the contour coordinates of the same multi-dimensional force sensor in all images are combined, and the profile image of the multi-dimensional force sensor is obtained; The calculation of the center point of each multi-dimensional force sensor and the identification surface profile comprises the following steps: For a multi-dimensional force sensor, a square minimum bounding box of the multi-dimensional force sensor profile image is calculated, and the center of the multi-dimensional force sensor is the middle of the minimum bounding box; The distance between the center point and the points in each contour group corresponding to the mounting block of the multi-dimensional force sensor is calculated, the radius of the mounting block is determined according to the mode of the distance, and the points with distance not equal to the radius are screened out, so as to determine the position of each identification surface profile.
2. The welding method suitable for a multi-dimensional force sensor according to claim 1, characterized by, Before calculating the average value Gave(x,y), the maximum value of the gray values of the adjacent pixel points along the y-axis direction is removed.
3. The welding method suitable for a multi-dimensional force sensor according to claim 1, characterized by, The welding action parameters of each multi-dimensional force sensor are determined, and the welding action parameters of each multi-dimensional force sensor are determined. A comparison library of welding action parameters is set. For a multi-dimensional force sensor, select an arbitrary image with the largest number of contour groups, calculate the difference between the y value of each contour group and y0, and arrange them from small to large as △y1, △y2, …, △yq. According to the value of q and the values of △y1, △y2, …, △yq, the corresponding welding action parameters are searched in the comparison library.
4. The welding method suitable for a multi-dimensional force sensor according to claim 1, characterized by, Before welding, first use a line laser to irradiate and take an image without placing a multi-dimensional force sensor, and confirm whether the laser pixel points in each image are on the same straight line, if so, then proceed to the subsequent normal welding, if not, then stop and issue an alarm.
5. The welding method suitable for a multi-dimensional force sensor according to claim 1, wherein, The irradiation direction of the line laser is 45° with the horizontal plane.
6. A welding system suitable for a multi-dimensional force sensor, characterized in that: It comprises: A device rack is provided with a welding platform, and a welding area is arranged on the welding platform; A camera is fixedly arranged above the welding area; A laser emitter slides above the welding area and irradiates a line laser on the welding area; A carrying assembly is used to drive the multi-dimensional force sensor into and out of the welding area; A welding arm moves above the welding area and is used to weld the multi-dimensional force sensor; A processor is used to realize the welding method for the multi-dimensional force sensor as claimed in any one of claims 1-5.
Citation Information
Patent Citations
Method and system for detecting contour of battery main body based on in-line linear laser illumination
CN109668524A
Steel bar counting method
CN109712180A
Chamfer size measurement method and device, electronic equipment and storage medium
CN115060162A
Detection method and device for identifying and positioning welding seam of steel grating
CN119027398A