In the production process of cloth, like cloth quality detection this highly repetitive and intelligence work can only be done by manual detection, in the modern assembly line workers often can be seen behind the detection of many to perform this procedure, the enormous increase of labor cost and management cost at the same time to the enterprise, but still can not guarantee 100% inspection pass rate (i.e., "zero defect"). The detection of the quality of the cloth is repeated work and low efficiency, error prone.

Pipeline automation transformation, make cloth production line into a fast, real-time, accurate, efficient pipelining. On the assembly line, all the cloth color, and the number will automatically confirm (hereinafter referred to as the "fabric"). By using automatic recognition technology of machine vision done previously by manual to complete the work. In a bulk fabric inspection, using artificial to check the quality of products with low efficiency and low accuracy, detection methods using machine vision can greatly improve the production efficiency and production automation.

Feature extraction identification

Common cloth detection (automatic identification) first use the high definition, high speed camera shot standard image, based on certain standards; then the captured image is detected, then the contrast. But in the cloth quality detection in complex engineering:

1 the content of image is not a single image, number, each measured impurity region the existence of size, color, position is not necessarily consistent.

2 impurity shape it is difficult to determine in advance.

3 due to the rapid movement of light cloth produce reflection, there may be a lot of noise in the image.

4 on the assembly line, to detect the cloth, have real-time requirements.

Because of the above reasons, the image recognition processing should take corresponding algorithm, feature extraction of impurities, pattern recognition, the realization of intelligent analysis.