A method for identifying wetted areas of multi-color fabrics based on hyperspectral image processing

Abstract

The invention discloses a method for recognizing a wetted area of multicolor fabric based on hyperspectral image processing. An EVA foam-modified gripper is used for providing a precise boundary for the extraction of an image of a test area; due to relatively large near-infrared reflectivity difference between the wetted area and an unwetted area, through the adoption of near-infrared band imaging, not only is the image contrast increased to the maximum extent, but also interference of the color as well as noises generated by fabric texture during wetted area recognition are effectively overcome; through MNF transformation and 2% linear stretching of the image, not only are most abnormal values removed, but also the image contrast is increased, so that the occurrence of mistaken segmentation can be avoided better; through the adoption of iterative threshold segmentation, the adaptability and automation capability during region segmentation are improved; through the selective filling ofthe wetted area, the reflective wetted area is effectively corrected; by adopting the method, the wetted area of fabric with any color, especially the multicolor fabric, can be accurately recognized,and an accurate basis is provided for automatic detection of the water dipping level of the fabric.

Description

【Technical field】 [0001] The invention relates to the technical field of textile and garment performance testing, in particular to a method for identifying wet areas of multi-color fabrics based on hyperspectral image processing. 【Background technique】 [0002] The identification of the wetted area of ​​the fabric in the process of traditional artificial fabric water staining grade assessment is easily affected by the assessor's physiology and psychology, lighting environment, fabric color, etc., and the experimental results have large errors and poor consistency. In order to solve these problems, an image processing-based identification method of fabric wetted area appeared for the spraying method of AATCC standard. [0003] The difficulty of the fabric wetting area recognition method based on image processing is that the contrast between the wetting area and the non-wetting area to be segmented is small, and the segmentation results are greatly affected by uneven illuminat...

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is to overcome the deficiencies of the above-mentioned prior art, and to provide a multi-color fabric wet area recognition method based on hyperspectral image processing, which aims to solve the problem of fabric wetness level detection process based on image processing algorithms in the prior art. Interfering influences such as colour, fabric texture, uneven lighting, light variations, reflections, low contrast between wetted and non-wetted areas and test areas prone to missegmentation etc. technical issues

Images