Intelligent detector and method of copper strip surface quality on basis of vision bionics

Abstract

The invention discloses an intelligent detector and a method of copper strip surface quality on the basis of vision bionics. An industrial camera and the like are adopted to obtain a copper strip surface image; the obtained image is strengthened, denoised and debounced; Gaussian pyramid decomposition and Gabor pyramid decomposition are used to extract three characteristics of luminance, color and orientation; central peripheral difference of a central fine scale and a periphery rough scale is calculated to the extracted characteristics to obtain a characteristic saliency map; a saliency map is formed with a local iteration method; the distinguishing characteristic value, i.e. what information of each zone is directly obtained from an interesting map, and a where information flow is obtained according to the focal point transferring sequence; a discrete observable Markovian model is used to correctly classify the target so as to improve the identification and classification capability of the copper strip surface quality detection system to defects as well as satisfy the requirements of instantaneity and accuracy of on-line production, so that the invention can be widely applied to surface defect quality detection of other materials.

Description

technical field [0001] The invention relates to visual bionic technology and metal surface quality detection technology, in particular to the detection of copper strip surface quality based on visual bionic mechanism. Background technique [0002] With the wide application of copper in electrical, light industry, machinery manufacturing and construction industries, the quality of copper materials directly affects the quality and performance of downstream products (such as: communication cables, pipes, etc.). Therefore, how to solve the problem of copper Efficient and accurate detection of surface quality is of paramount importance. Copper strip surface quality inspection is a simple and repetitive task with low confidence. Traditional manual visual inspection and stroboscopic light detection methods have disadvantages such as poor real-time performance, low sampling rate, low detection confidence and harsh detection environment. The current surface quality inspection is ma...

AI Technical Summary

Problems solved by technology

At this time, prior knowledge needs to be used. Prior knowledge may be a template in memory, but the actual target is not exactly the same as the template, deformation may occur, or the degree of background interference is different.

Images