Rapid extracting method for structure light welding seam image characteristic points

Abstract

The invention relates to a structure light vision detection technology, in particular to a rapid extracting method for structure light welding seam image characteristic points. In the invention, an on-site programmable gate array and a soft-core processing system are used together for improving processing speed and reducing power consumption; and the rapid extracting task for structure light welding seam image characteristic points is divided by software and hardware to extract characteristic points by software and hardware cooperation. The invention comprises the following steps: building a Programmable System-on-Chip (PSoC); carrying out mathematical morphology processing and image preprocessing on structure light stripe-containing image which is acquired by a camera sensor by the on-site programmable gate array, wherein the mathematical morphology processing comprises filtering, image enhancement, dilation and erosion, and the image preprocessing comprises edge extraction and centre line extraction; and extracting characteristic points by the soft-core processing system. In the invention, production-line working way is used among steps, thereby improving image processing speed and meeting real-time requirement of seam tracking. The invention can be widely used in the field of laser automatic welding.

Description

technical field [0001] The invention belongs to the field of machine vision, and relates to a structured light visual inspection technology; in particular, it relates to a method for quickly extracting feature points of structured light welding seam images based on an on-chip programmable system (SOPC). Background technique [0002] Structured light vision inspection technology can realize intelligent, fast and low-cost image acquisition and feature recognition. This method mainly projects the structured light onto the surface of the workpiece, collects the structured light fringe image through the visual sensor, and extracts the feature information of the weld fringe through image processing, so as to realize the position recognition and quality inspection of the weld. [0003] However, due to the complexity of structured light weld image processing, the real-time performance of weld seam recognition and detection applications in weld seam tracking and quality inspection is...

AI Technical Summary

Problems solved by technology

[0004] In order to overcome the deficiency that the current structured light weld image processing speed cannot meet real-time processing, the present invention proposes a fast and accurate structured light weld image feature point fast track based on an on-chip programmable system, which can achieve the purpose of tracking real-time, fast and accurate. Extraction Method

Images