Workpiece classification method, device and apparatus and medium

Abstract

The invention provides a workpiece classification method, device and apparatus and a medium, and the method comprises the steps: obtaining a workpiece image of a to-be-identified workpiece, and extracting a histogram of oriented gradients HOG feature of the workpiece image; based on the extracted HOG features and a pre-trained support vector machine (SVM) model matched with each workpiece category, determining a predicted classification probability that the to-be-identified workpiece belongs to each workpiece category; screening out a target prediction classification probability meeting a preset condition from the determined prediction classification probabilities; and determining the workpiece category corresponding to the target prediction classification probability as the category of the to-be-identified workpiece. By adopting the method, the workpiece classification accuracy can be improved.

Description

technical field [0001] The present application relates to the technical field of image processing, in particular, to a workpiece classification method, device, equipment and medium. Background technique [0002] With the rise of the intelligent manufacturing industry, the industrial manufacturing process is becoming more and more intelligent. For example, in the process of building a smart factory, the robotic arm needs to sort and grab various workpieces, which requires accurate identification and classification of workpieces. [0003] At present, when identifying and classifying workpieces, image features such as edge features of workpiece images are often extracted, and the extracted image features are matched with preset image features to determine the workpiece category. However, for some workpieces with relatively simple structures, such as flanges, the features extracted from the flange image are relatively simple. When predicting the category of such workpieces, the...

AI Technical Summary

Problems solved by technology

However, for some workpieces with relatively simple structures, such as flanges, the features extracted from the flange image are relatively simple. When predicting the category of such workpieces, the extracted image features are generally independent of each pixel. , does not take into account the correlation between each pixel, so there will be a problem of poor classification accuracy of artifacts

Images