Food packaging defect sample generation method based on custom algorithm

Abstract

The invention belongs to the technical field of machine vision detection, and designs a method for generating a food package flaw sample based on a custom algorithm, which is used for detecting flaws existing in a plastic package. Firstly, a food package defect image sample data set is collected; sampling the gray scale and curvature of textures of all flaw image samples; the method comprises the following steps of: randomly generating curves according to sampling data, randomly generating the number and postures of the curves, and drawing flaws on a background image to generate an image with flaws on the surface; when real defect data is lacked, a large number of surface flaw defect images are randomly synthesized by simulating the characteristics of the surface flaw defects, and the blank of insufficient real surface flaw defect image data is filled up, so that a related flaw defect detection algorithm is assisted to improve the detection performance.

Description

technical field [0001] The invention belongs to the technical field of machine vision detection, and in particular relates to a method for generating defective samples of food packaging based on a custom algorithm. Background technique [0002] In the processing and production of instant food products, when plastic sealing of food, due to factors such as water vapor, air, food soup and other factors generated by high temperature during plastic sealing, food packaging bags are often unqualified for plastic sealing. Food packaging bags are easily contaminated by harmful bacteria, viruses, fungi and parasites in the air. Eating contaminated food may pose a hazard to the personal safety of consumers. [0003] The existence of food safety problems seriously endangers the personal health and legitimate rights and interests of consumers, and also causes heavy losses to the reputation and reputation of food processing and sales companies. At present, the detection method of food p...

AI Technical Summary

Problems solved by technology

[0006] For a new product, because the image data of its surface defects cannot be collected in a short period of time, the algorithm cannot adapt to the new product well, thus affecting the detection performance of the algorithm

[0007] In the fields of metal, wood, and hard plastic boards, a large number of defect generation algorithms have been produced and achieved good results and excellent practical application performance. However, the texture of plastic packaged food is complex and changeable, and it is not a simple straight line fitting. There is no research on the generation of texture defects in plastic-packaged foods, and the related needs are urgent

Images