A visual positioning method, system and storage medium for a mold clamping machine

Abstract

A visual positioning method, system and storage medium of a mold clamping machine, the method includes the steps of: calibrating a binocular camera to obtain internal and external parameters of the camera; using the binocular camera to shoot a group of digital grayscale images for I l and I r ; for image I l and I r After correction, the edge image I' is obtained after edge detection l and I' r ; Obtain the projected image of the target, use the projected image to match and identify the target in the scene, and obtain the image I' l and I' r The pixel coordinates of the target in (x l ,y l ) and (x r ,y r ), the difference between the two abscissas is recorded as the parallax d, d=x l ‑x r ; Compensate the pixel-level parallax d to obtain high-precision sub-pixel parallax d'; use the coordinate mapping matrix K obtained from camera calibration and sub-pixel parallax d' to obtain the Cartesian coordinates of the target object. The method of the invention has the advantages of high recognition rate, high calculation precision, fast calculation speed and little environmental impact.

Description

technical field [0001] The invention relates to the field of image and video data processing, and relates to the identification and positioning of a specific target in an image, in particular to a visual positioning method, system and storage medium of a clamping machine. Background technique [0002] Target recognition and localization are important issues in computer vision. Its purpose is to measure the three-dimensional space coordinates of the target object and provide relative position information of the measured object. Target recognition and positioning are widely used in industrial automation, robotics, unmanned driving and other fields. It is an important research field in computer vision, especially in image understanding. [0003] Target recognition and positioning can effectively extract the depth information of targets in complex scenes, providing a solid guarantee for machine intelligence. There are many methods for target recognition and localization, and m...

AI Technical Summary

Problems solved by technology

2. Sensitivity to noise

Factors such as working environment and illumination changes affect the quality of the captured image. The quality of the image and the interference structure in the scene will directly lead to misrecognition and missed recognition of the target object.

Images