Printing image defect detection method based on area combination feature

Abstract

The invention relates to a printing image defect detection method based on an area combination feature, and belongs to the technical field of image processing. The printing image defect detection method comprises a learning process and a training process. The learning process comprises the following specific steps: (1) partitioning areas specific to a printing sample image; (2) judging an area to which a small printing image belongs in a training sample set; (3) performing feature extraction specific to a training sample; (4) building a WLDA (Wireless Loss Differentiation Algorithm) classifier; (5) training the WLDA classifier. The training process specifically comprises the following steps: (6) judging an area to which a testing sample belongs in the printing sample image; (7) performing feature extraction specific to the testing sample; (8) according to the area to which the testing sample belongs, judging whether the testing sample contains a defect image or a defect-free image by using the WLDA classifier corresponding to the area. The printing image defect detection method based on the area combination feature provided by the invention has the advantages that the pre-extraction workload of a template is small; the influence of the difference between different batches and a pre-extraction template on a detection result is small; the detection efficiency is high.

Description

technical field [0001] The invention relates to a printing image defect detection method based on area combination features, in particular to a printing image defect detection method based on area combination features and weighted linear discriminant analysis, belonging to the technical field of image processing. Background technique [0002] With the development of the printing industry, people have higher and higher requirements for the appearance, color tone and screen design of printed products. However, due to factors such as technology and machinery, such as missing printing, ink stains, Defects such as black dots, blurred text, wrinkling, scratches, pinholes, color distortion, overprint misalignment, etc. The traditional manual detection method has been difficult to complete the detection task with quality and quantity. With the rapid development of computer software and hardware, it has become feasible to use machine vision to automatically detect the image quality ...

AI Technical Summary

Problems solved by technology

The main disadvantages of the template method are: (1) The pre-extraction of the template is a lot of work

Different batches of printed products have differences between the images collected by the camera and the pre-extracted templates due to factors such as environment, lighting, and angles, which affect the detection results; (3) The detection workload is large

Due to the above reasons, different batches need to re-extract templates, or perform color correction and angle distortion correction based on pre-extracted templates, which increases the detection workload and lowers the detection efficiency.

Images