Machine vision detection method, detection device, and detection system

Abstract

A machine vision detection method, detection device, and detection system. The detection method includes the steps of receiving a three-dimensional image from a line laser, converting the three-dimensional image into a two-dimensional grayscale image, obtaining a boundary of a first member and a boundary of a second member from the two-dimensional grayscale image, determining N perpendicular lines located between the boundary of the first member and the boundary of the second member, and calculating an average value of the lengths of the N perpendicular lines as a gap between the first member and the second member. The three-dimensional image includes at least a part of the boundary of the first member, the boundary of the second member, and at least one welding point located on the boundary of the first member and the second member, where N is a positive integer. The method can more accurately calculate the gap between the two members by performing multiple detections and calculating the average value, thereby improving the detection accuracy. It also supports continuous sampling of the detection member, and does not require pre-calibration of the camera.