Flexible material detection method and storage medium thereof

Abstract

The invention discloses a flexible material detection method and a storage medium thereof. The method comprises the following steps: S1, acquiring image data of a flexible material; S2, detecting a plurality of pieces of surface defect information in the image data, judging the obvious degree of the surface defect information, and classifying each piece of surface defect information according to the obvious degree; S3, selecting a feature matching identification step or a neural network identification step according to the classification condition; s4, performing a feature matching identification step, and analyzing the surface defect information based on a feature matching identification model to obtain first defect type information; s5, performing a neural network identification step, and analyzing the surface defect information based on a neural network identification model to obtain second defect type information; s6, integrating the first defect type information and / or the seconddefect type information and outputting a detection result. According to the design, reasonable classification is carried out, the accuracy of an overall result is ensured, and the overall detection rate is improved through reasonable distribution.

Description

technical field [0001] The invention relates to the field of flexible material quality inspection, in particular to a flexible material inspection method and a storage medium thereof. Background technique [0002] In the process of textile production, the detection of surface defects of flexible materials such as cloth is a particularly critical part, and the detection requirements become more stringent with the continuous improvement of market requirements for product quality, and with the continuous increase of textile production, cloth inspection The speed and precision requirements of the process are constantly increasing. At present, the detection method of most textile enterprises is still manual detection, which is easily affected by the subjective factors of the inspectors, and the manual detection speed is slow, the cost is high, the false detection rate is high, and the domestic labor cost is getting higher and higher, so Manual inspection is increasingly unable t...

AI Technical Summary

Problems solved by technology

[0003] With the rapid development of science and technology, image processing technology is becoming more and more mature. Under the existing image processing technology, there are generally many image processing methods. However, the feature matching and recognition image processing steps are not very compatible and have poor scalability. When applied to the detection of flexible materials such as cloth, it cannot meet the detection requirements of multiple types of cloth, and for complex textures The cloth in the background cannot be processed well; the other is the deep learning detection method based on the neural network. The features extracted by this method have stronger generalization performance than the artificially set features, and can be well adapted to many The detection of different types of cloth has wide applicability, but the large amount of parameters in the neural network and the huge amount of calculation make the training very slow, the storage is very inconvenient, the complexity is high, the development cycle is long, and the hardware requirements are high, so the cost is expensive

Images