A method for extracting the gluing trajectory of industrial robot soles

Through the extraction method of industrial robot sole glue coating trajectory, the spraying motion trajectory is automatically designed using 3D visual recognition and data processing, which solves the problem of relying on manual operation of the sole glue coating process in the existing technology, and improves the efficiency and intelligence of shoemaking. It realizes automated production of different shoe sizes and left and right feet.

CN112465767BActive Publication Date: 2024-11-05PANDA ELECTRONICS +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202011339092.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-11-25
Publication Date
2024-11-05
Estimated Expiration
2040-11-25

AI Technical Summary

Technical Problem

In the existing shoe leather processing and manufacturing industry, the sole glue coating process relies on manual operation, which is inefficient and difficult to master. In the face of different styles and sizes, it needs to be re-teached and programmed, resulting in a shortage of job types and low production efficiency.

Method used

Using the industrial robot sole glue coating trajectory extraction method, three-dimensional point cloud data is obtained through 3D line laser camera scanning, data processing and dimensionality reduction, sole edge trajectory is extracted, glue coating points and normals are fitted, and glue coating trajectory is optimized to obtain glue coating trajectory Line, realizes automatic spray motion trajectory design.

Benefits of technology

The automatic processing of glue coating for soles has been realized, the efficiency and intelligence of shoemaking are improved, the volatility of glue has been avoided to the personnel, and the mixed production of different shoe sizes and left and right feet has been supported, which solves the limitations of single shoe sizes in traditional production lines.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112465767B_ABST
    Figure CN112465767B_ABST
Patent Text Reader

Abstract

The present invention discloses a method for extracting the gluing trajectory of the sole of an industrial robot, wherein the initial 3D point cloud data is obtained by scanning the sole with a line laser camera, and the gluing trajectory line of the sole and the robot posture are obtained through a series of processing methods such as filtering, fitting, optimization, simplification, and smoothing. The present invention realizes accurate extraction of the gluing line of the edge of the sole without manual intervention, and finally outputs the robot gluing trajectory through data processing.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to an industrial intelligent manufacturing production method, and in particular to a method for extracting the gluing trajectory of an industrial robot sole. Background Art

[0002] At present, in the shoe leather processing and manufacturing industry, most of the gluing process is done manually, due to factors such as frequent changes in shoe styles and difficulty in remodeling production lines. In addition, gluing of soles also has problems such as high shoe body temperature and strong toxicity. It is difficult to recruit employees for gluing soles in shoe factories, which has become the most common pain point in the industry.

[0003] Currently, the more common sole gluing technology mostly adopts the teaching pendant trajectory simulation method, which is heavily dependent on manual participation. The teaching of the spraying surface is basically determined by visual inspection. The process is cumbersome and inefficient, and the accuracy is not easy to master. In addition, for different styles and sizes, as well as left and right feet and unused production lines, re-teaching and programming are required. Summary of the invention

[0004] Purpose of the invention: The purpose of the present invention is to provide a highly intelligent method for extracting the gluing trajectory of the sole of an industrial robot.

[0005] Technical solution: The method for extracting the gluing trajectory of the industrial robot sole of the present invention comprises the following steps:

[0006] (1) Scanning the sole surface to obtain three-dimensional point cloud data, performing initial data processing on the point cloud data, and obtaining a 3D point cloud image of the initial contour of the sole;

[0007] (2) Reduce the dimension of the 3D point cloud image of the initial sole contour and map it into a two-dimensional grayscale image;

[0008] (3) Processing the two-dimensional grayscale image to obtain the initial edge trajectory curve of the sole and calculating the center point of the sole as the subsequent initial point;

[0009] (4) With the initial point as the center, a rectangular slice with the sole plane as the section is established, and the extreme point of the point cloud change in the slice is calculated to find the edge point of the sole. The first glue coating point is offset to the inside of the shoe with this as the starting point, and the point cloud plane is fitted with the glue coating point as the center, and the normal of the plane is calculated;

[0010] (5) Further offset in the direction of the shoe edge to find the edge point, repeat step 4 to find the second and third glue application points and the normal line;

[0011] (6) fitting a straight line using the first to third glue-coating points, finding the edge point in the direction of the next shoe edge and the section perpendicular to the straight line, and calculating the glue-coating points and the normal line;

[0012] (7) Repeat step 6 until all the glue coating points and normal lines of the sole are found;

[0013] (8) Optimize all glue points and normal lines;

[0014] (9) Obtain the gluing trajectory based on all the gluing points and postures.

[0015] Furthermore, in step (1), the scanning is performed by a 3D line laser camera.

[0016] Furthermore, in step (1), the initial data processing includes noise filtering and elimination of isolated points and isolated small areas.

[0017] Furthermore, the noise filtering is 11×11 median filtering; isolated points and isolated small areas are points or areas that are more than 100 pixels away from the outside of the largest closed area.

[0018] Furthermore, in step (2), the mapping is to normalize the height values ​​in the 3D point cloud in the sole surface to between 0 and 255 through the grayscale image pointer, and save them as grayscale values.

[0019] Furthermore, in step (3), the data processing is morphological filtering processing, and the morphological filtering includes corrosion, dilation and edge detection, closed area detection, obtaining the largest trajectory line of the closed area, curve smoothing, and then calculating the center point of the closed area.

[0020] Furthermore, in step (8), the optimization process includes sparse processing, outlier removal and curve smoothing.

[0021] Furthermore, the thinning process is that the interval between the glue points at the sole and toe positions is not less than 1 mm, and the interval between the glue points at other positions is not less than 2 mm; the outlier points are eliminated when the area of ​​the triangle formed by three consecutive points is not greater than 1.5 times the area of ​​the triangle formed by the first three points.

[0022] Furthermore, the curve smoothing algorithm is Savitzky-Golay.

[0023] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages:

[0024] (1) The present invention uses 3D vision to identify the three-dimensional structural model of the rubber coating surface of the sole, and then designs the spraying motion trajectory on the model through an algorithm to realize the automation of the shoemaking process and achieve mixed production of different shoe sizes and left and right feet, thus avoiding the disadvantage that traditional shoemaking production lines can only produce the same shoe size in the same batch.

[0025] (2) The robot can automatically extract the edge of the shoe sole and accurately track the gluing process without human intervention, thus avoiding the risk of glue volatilizing into the air and causing harm to personnel. The shoemaking efficiency is improved, the degree of intelligence is high, and it is energy-saving and environmentally friendly. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a schematic diagram of the process of the gluing trajectory extraction method of the present invention. DETAILED DESCRIPTION

[0027] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0028] Step 1: Use a 3D line laser camera to scan the sole surface to obtain the original 3D point cloud data, and perform initial data processing on the point cloud data, including noise filtering, and removing isolated points and isolated small areas. Obtain the initial contour 3D point cloud image of the sole.

[0029] Step 2: Reduce the dimension of the 3D point cloud image of the initial contour of the sole and map it into a two-dimensional grayscale image.

[0030] Step 3: Perform data processing such as morphological filtering on the two-dimensional grayscale image to obtain the initial edge trajectory curve of the sole and calculate the center point of the sole as the subsequent initial point.

[0031] Step 4: With the initial point as the center, create a rectangular slice with a length of 20 mm and a width of 1 mm with the sole plane as the section, calculate the extreme points of the point cloud change in the slice, and find the edge points of the sole. Use this as the starting point to offset toward the inside of the shoe (where the slope of the section curve changes) to find the first gluing point, fit the point cloud plane with the gluing point as the center, and calculate the normal of the plane.

[0032] Step 5: Repeat step 4 with a certain offset in the direction of the shoe to find the second and third glue points and the normal.

[0033] Step 6: Use the first to third gluing points to fit a straight line, find the edge point in the direction of the next shoe and the section perpendicular to the straight line, and calculate the gluing points and normal lines.

[0034] Step 7: Repeat step 6 until all the glue points and normals of the sole are found.

[0035] Step 8: Optimize all glue points and normals, including thinning, outlier removal, and curve smoothing.

[0036] Step 9: Finally, the gluing trajectory is obtained based on all the gluing points and postures.

Claims

1. A method for extracting the gluing trajectory of an industrial robot sole, characterized in that: The following steps are involved: (1) Scanning the sole surface to obtain three-dimensional point cloud data, performing initial data processing on the three-dimensional point cloud data, and obtaining a 3D point cloud image of the initial contour of the sole; (2) Reduce the dimension of the 3D point cloud image of the initial sole contour and map it into a two-dimensional grayscale image; (3) performing data processing on the two-dimensional grayscale image to obtain the initial edge trajectory curve of the sole, and calculating the center point of the sole as the subsequent initial point; wherein the data processing is morphological filtering processing, and the morphological filtering processing includes corrosion, expansion and edge detection, closed area detection, obtaining the maximum trajectory line of the closed area, smoothing the curve, and then calculating the center point of the closed area; (4) With the initial point as the center, a rectangular slice with the sole plane as the section is established, and the extreme point of the point cloud change in the slice is calculated to find the edge point of the sole. The first glue coating point is offset to the inside of the shoe with this as the starting point, and the point cloud plane is fitted with the glue coating point as the center, and the normal of the plane is calculated; (5) Further offset in the direction of the shoe edge to find the edge point, repeat step (4) to find the second gluing point, the third gluing point and the normal line; (6) fitting a straight line using the first to third glue-coating points, finding the edge point in the direction of the next shoe edge and the section perpendicular to the straight line, and calculating the glue-coating points and the normal line; (7) Repeat step (6) until all the glue coating points and normal lines of the sole are found; (8) Optimizing all glue points and normals; wherein the optimization includes thinning, outlier removal, and curve smoothing; the thinning is that the interval between glue points at the sole and toe is not less than 1 mm, and the interval between glue points at other positions is not less than 2 mm; outlier removal is that the area of ​​a triangle formed by three consecutive points is not greater than 1.5 times the area of ​​a triangle formed by the first three points; the curve smoothing algorithm is Savitzky-Golay; (9) Obtain the gluing trajectory based on all the gluing points and postures.

2. The method for extracting the gluing trajectory of the industrial robot sole according to claim 1, characterized in that: In step (1), the scanning is performed by a 3D line laser camera.

3. The method for extracting the gluing trajectory of the industrial robot sole according to claim 1, characterized in that: In step (1), the initial data processing includes noise filtering and elimination of isolated points and isolated small areas.

4. The method for extracting the gluing trajectory of the industrial robot sole according to claim 3 is characterized in that: The noise filter is an 11×11 median filter; isolated points and isolated small areas are points or areas that are more than 100 pixels away from the outside of the largest closed area.

5. The method for extracting the gluing trajectory of the industrial robot sole according to claim 1, characterized in that: In step (2), the mapping is to normalize the height value in the 3D point cloud in the sole surface to between 0 and 255 through the grayscale image pointer, and save it as a grayscale value.