Coarse-to-fine paper defect detection method

Abstract

The invention relates to the technical field of image processing and paper defect detection, in particular to a coarse-to-fine paper defect detection method which mainly comprises three modules and processes, namely a paper defect screening module, a coordination control module and a paper defect classification module. The method comprises the following steps that: 1, a paper defect screening module runs in an FPGA-based embedded system to ensure that a high-speed high-resolution paper image can be processed in real time, thus screening out an image possibly having a defect after the sampled paper image is primarily processed, directly discarding a normal image after the paper defect is screened, and transmitting an image with a suspected paper defect to the next step; secondly, the coordination control module mainly completes the functions of function calling, feedback control and the like between the paper defect screening module and the paper defect classification and grading module; thirdly, the paper defect classification and grading module builds a classification and grading network through a convolutional network, the classification network can complete the function of classifying paper defects, and the grading network can complete the function of grading the severity degree of the paper defects.

Description

technical field [0001] The invention relates to the technical fields of image processing and paper defect detection, in particular to a coarse-to-fine paper defect detection method. Background technique [0002] In the papermaking process, due to the influence of a series of factors such as raw materials, cutting tools, assembly line and production environment, various paper defects will appear, such as: holes, wrinkles, scratches, dirt, etc. The existence of these paper defects is reduced the quality of paper products. Early factories used the traditional manual inspection method for quality management, but the manual inspection method has low sampling rate, low accuracy, poor real-time performance, low efficiency, high labor intensity, and is greatly affected by manual experience and subjective factors. Repeated, noisy on-site, inspectors are prone to fatigue, resulting in unsatisfactory inspection results. Due to the higher and higher requirements for detection accuracy...

AI Technical Summary

Problems solved by technology

This method meets the requirements of real-time detection of paper defects, and partially meets the needs of applications, but its classification of paper defects is not fine enough, and the severity of paper defects cannot be distinguished under the same category, which is not conducive to detailed classification and grading of paper defects. According to these Rough detection and classification results are also not conducive to fine analysis of paper defects and improvement of production process, so as to better improve the quality of paper products

In addition, the above method also has problems such as poor adaptability to different paper types, low adaptability to ambient light changes, and inconvenient maintenance.

Images