A workpiece identification method, device, electronic equipment, and storage medium

Abstract

The present application discloses a workpiece identification method, device, device and medium. The method includes: performing rough positioning on an image to be identified, and determining a candidate region of interest in the image to be identified; feature points; determine the feature points of the first level as the feature points to be matched, and use the preset template data to match the feature points to be matched to obtain the target matching degree; if the target matching degree is less than the first preset threshold, the subsequent matching is prohibited. process; if the target matching degree is not less than the first preset threshold, the feature points of the next level are determined as the feature points to be matched, and the step of matching the feature points to be matched is returned, until all the feature points of all levels are matched, and the workpiece is obtained. pose information. The present application narrows the range of feature matching through coarse positioning; if the matching degree of feature points at any level is less than the threshold, the subsequent matching process is not performed, which avoids useless work and improves work efficiency.

Description

technical field [0001] The present application relates to the technical field of image recognition, and more specifically, to a workpiece recognition method and device, an electronic device, and a computer-readable storage medium. Background technique [0002] With the advancement of science and technology, more and more industrial robots are used in the production field to replace manual labor for repetitive production activities. Industrial robots are multi-joint robotic arms or multi-degree-of-freedom machine devices for industrial manufacturing, which can automatically perform related tasks according to encoded programs and artificial intelligence technology. Practice has shown that industrial robots can significantly improve industrial production efficiency, and are the core part of industrial automation production processes, with very broad development prospects. [0003] In the current industrial production line, the detection and identification of workpieces is an i...

AI Technical Summary

Problems solved by technology

[0004] At present, there are two main methods for intelligent identification of workpieces. One is the deep learning method, which requires a lot of data to be manually marked for training, and the detection process is also very time-consuming, resulting in high labor and time costs; the other is the deep learning method. The first is the feature extraction and matching method. This method usually scans all the target workpiece images to extract features, and then matches all features. The recognition takes a long time and is inefficient.

Images