Construction method of fabric defect recognition system based on lightweight convolutional neural network

Abstract

The invention provides a construction method of a fabric defect recognition system based on a lightweight convolutional neural network. The method comprises the following steps: firstly, configuring an operation environment of a fabric defect identification system; obtaining a lightweight convolutional neural network according to factorization convolution; then, collecting fabric image sample data; standardizing the fabric image sample data; dividing the standardized fabric image sample data into a training image set and a test image set; inputting the training image set into a lightweight convolutional neural network for training by using an asynchronous gradient descent training strategy to obtain an LZFNet-Fast model, and finally inputting the test image set into the LZFNet-Fast modelfor testing to verify the performance of the LZFNet-Fast model. According to the method, a standard convolution layer is replaced by a factorized convolution structure, the colored fabric with complextextures is effectively identified, the number of parameters and the calculated amount of the model are reduced, and the identification efficiency is greatly improved.

Description

Technical field [0001] The invention relates to the technical field of textile image recognition, in particular to a method for building a fabric defect recognition system based on a lightweight convolutional neural network. Background technique [0002] At present, the production cost of finished textile products is often affected by the original grey fabric. Most of the quality problems in the clothing industry are related to fabric defects, which is one of the main problems facing the textile industry. Fabric defects, often referred to as fabric defects, refer to defects in the appearance of products caused by various unfavorable factors in the process of fabric weaving. From fiber raw materials to finished fabrics, it generally needs to go through multiple processes such as spinning, weaving, printing and dyeing, and defects may occur in each processing link. The cost of manual inspection of fabric defects is too high, and it is easy to miss inspection. Therefore, textile m...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the technical problems of complex structure and large amount of calculation in the traditional deep convolutional neural network, the present invention proposes a fabric defect recognition method based on a lightweight convolutional neural network, and constructs a dedicated convolution module in the field of fabric recognition. This module incorporates the advanced factorable decomposed convolution structure. Firstly, the fabric image is extended to 32 dimensions through three-dimensional convolution operation as the input of the convolution module, and then the factorized convolutional layer is used for spatial filtering, and then the feature The graph is compressed into a low-dimensional space by a global average pooling layer on top of the network

Images