Gluing information detection method based on structured light measurement

Abstract

The invention discloses a gluing information detection method based on structured light measurement. The method comprises the steps of (1) carrying out image processing on a glue strip contour structured light image, segmenting a bottom straight line section, and carrying out the fitting to obtain a bottom reference line, (2) calculating an included angle theta between the bottom reference line and an x axis of a light plane coordinate system, rotating the glue strip contour -theta degrees, removing the bottom straight line section, and calculating a glue height feature point and a glue widthfeature point in a glue shape contour section formed by remaining contour points, (3) obtaining a plurality of glue height feature points, calculating a normal line, constructing a plane I according to the normal line, selecting a preset frame image, projecting glue height feature points and the glue width feature points in the preset frame image to the plane I, calculating a glue height by utilizing the glue height feature points, recording a distance between the glue width feature points as a glue width, and (4) repeating the steps (1), (2) and (3) for other to-be-detected positions in a gluing track to complete the detection of glue strip information in the whole gluing track. According to the method, the angle deviation is corrected, and the gluing three-dimensional contour informationcan be accurately obtained.

Description

technical field [0001] The invention relates to the field of glue coating detection, in particular to a glue coating information detection method based on structured light measurement. Background technique [0002] In the process of automobile manufacturing, the quality of glue coating is a key factor determining the sealing performance of the vehicle. The traditional glue coating detection method only relies on the original colloidal image collected by the camera for processing. This is a passive vision detection method, which can detect the glue width, Whether the glue is broken or not, but the glue height cannot be detected, and it is limited by the color of the glue and the glued bottom surface. When the two colors are similar or consistent, the traditional glue detection method will not be able to effectively detect the glue strip information, such as the current windshield glass. Most of the models with glue at the place have black background and black glue. For glue i...

AI Technical Summary

Problems solved by technology

[0002] In the process of automobile manufacturing, the quality of glue application is the key factor to determine the airtightness of the vehicle. The traditional glue inspection method only relies on the original glue image collected by the camera for processing. This is a passive visual inspection method, which can detect the width of the glue, Whether the glue is broken or not, but the height of the glue cannot be detected, and it is limited by the color of the glue and the bottom surface of the glue. When the two colors are similar or consistent, the traditional glue detection method will not be able to effectively detect the information of the glue strip, such as the current windshield. Most car models have black background and black glue. For this kind of glue, the advantage of using structured light measurement method is obvious. It uses single-line structured light, double-line structured light or multi-line structured light to measure The object projects the laser, and with the cooperation of the camera, the modulated structured light image (such as figure 1 shown); in order to accurately obtain the gluing information, the projected laser is required to be perpendicular to the measured rubber strip, but in the actual application process, due to the sensor assembly error, robot path, angle change and other reasons, there is an angular deviation between the projected laser and the rubber strip , that is, there are pitch angles (pitch), yaw angles (yaw), and rotation angles (roll); this will affect the accuracy of glue information detection

Images