A method for real-time detection of weld seam targets

Abstract

The invention belongs to the technical field of detection, in particular to a method for real-time detection of weld seam targets. The method comprises the following steps: building a training sampleset: welding seam images of different shapes are collected as source samples and the source samples are pretreated to form a training sample; training detector offline: the neural network is trained under different initial conditions by using the training samples, and the optimal neural network model is obtained by multiple training as a welding seam detector; performing on-line detection: a detection image is acquired, a weld detection is performed by using the weld detector and a detection result is output. The invention overcomes the problem of poor robustness of positioning the weld seam through the morphological method, realizes the accurate positioning of the weld seam of different shapes, classifies the weld seam of different kinds accurately, and has high positioning precision by adopting the weld seam detector obtained through the weld seam image training based on the line laser imaging for the different weld seams.

Description

technical field [0001] The invention relates to a real-time detection method for a welding seam target, in particular to a real-time detection method for a robot line laser welding seam. Background technique [0002] Due to the problems of high labor intensity, harsh working environment, and low efficiency of hand-held welding by workers, welding robots are gradually applied to welding production at home and abroad. Most of the traditional welding robots are teaching and programming robots, and sensors are used to assist in obtaining external information to adjust the welding trajectory. In recent years, with the development of computer vision, the welding seam tracking system based on laser and vision sensor imaging has begun to appear, and its characteristic is to realize the location of welding seam feature points by means of tracking. Since most of its tracking algorithms are based on the way of spatio-temporal context learning, the initial point of the weld needs to be...

AI Technical Summary

Problems solved by technology

[0004] In view of the above problems, the present invention provides a real-time detection method for weld targets, by collecting and preprocessing training samples, training a weld detector based on convolutional neural network, so that it can quickly and accurately identify and locate different types of welds It can effectively solve the problem of inaccurate positioning and poor robustness of the welding starting point by the morphological method used by the automatic welding seam tracking system in the current automatic welding technology

Images