No-reference quality evaluation system and method for industrial two-dimensional code

Abstract

The invention provides a no-reference quality evaluation system for an industrial two-dimensional code image, and the system comprises an industrial two-dimensional code database module which providesa to-be-evaluated industrial two-dimensional code test sample and an industrial two-dimensional code training sample for training, and inputs the to-be-evaluated industrial two-dimensional code testsample and the industrial two-dimensional code training sample into a multi-task MTL convolutional neural network module; the multi-task MTL convolutional neural network module which is used for training a multi-task MTL convolutional neural network on the industrial two-dimensional code training sample and completing a quality evaluation task on the industrial two-dimensional code to be evaluatedby adopting the trained multi-task MTL convolutional neural network. After the industrial two-dimensional code is predicted by the shallow multi-task convolutional neural network and the deep multi-task convolutional neural network, on one hand, various distortion types of the image can be judged, and on the other hand, the quality grade of the image can be predicted. According to the invention,the time cost and the calculation cost for preprocessing and decoding the industrial two-dimensional code can be greatly reduced, the decoding efficiency of the industrial two-dimensional code is improved, and the throughput of an industrial assembly line is increased.

Description

technical field [0001] The present invention relates to the technical field of image quality assessment of industrial two-dimensional codes, in particular to a reference-free quality assessment system and method for industrial two-dimensional codes. Background technique [0002] Industrial two-dimensional (2D) codes are ubiquitous throughout automated assembly lines. QR codes have been widely used in automatic identification tasks such as semiconductor wafer marking, industrial component review and fault diagnosis. Compared with 1D barcodes, 2D codes have more advantages, such as large data capacity, small size and built-in error correction / correction mechanism. With the development of smartphones, the high-quality QR codes most commonly seen in daily life can be decoded quickly and accurately. However, how to quickly and reliably decode low-quality industrial 2D codes remains a challenging task and has long been an important topic in the field of machine vision. Most ind...

AI Technical Summary

Problems solved by technology

However, traditional handcrafted feature-based no-reference quality assessment methods only support prediction of two single distortion types, namely defocus blur and lighting artifacts, and cannot provide fine-grained distortion type classification and quality level estimation.

Images