Cotton defect detection and recognition method based on multi-spectral technology

Abstract

A cotton defect detection and recognition method based on multi-spectral technology, the steps of which are: (1) image acquisition; (2) image preprocessing; (3) preliminary detection of defects: using the improved morphology gradient algorithm for edge detection to obtain each defect Integrate the strength of edge detection to obtain the contrast-enhanced image of each defect; (4) Accurate defect positioning: convert the image into a connected domain image by improving the iterative threshold method and morphological processing, and accurately locate the defect by selecting the main defect algorithm, and extract it by marking Defect area; (5) Ginning quality judgment: Count the number of each defect, calculate the total number of cotton defects, and obtain the ginning quality grade. The invention has the advantages of simple principle, fast detection speed and high detection precision.

Description

technical field [0001] The invention belongs to the technical field of optical measurement, in particular to an automatic classification detection and identification method for cotton gin quality. Background technique [0002] Lint defects refer to two categories: impurities with fibers and fibers that hinder weaving, including seven types of broken seeds, sterile seeds, wires, soft seed skins, stiff flakes, fibrous seed crumbs, and neps. Since the defects are difficult to discharge during the weaving process, the residual impurities are wrapped in the yarn sliver or attached to the surface of the yarn, resulting in deterioration of the evenness, increase of neps and impurities in the yarn, and increase of the hairiness of the yarn. The surface of the gray cloth made of yarn with a large number of fiber seed crumbs shows a large number of neps and impurities. After dyeing, the cloth surface will appear infertile color spots and hair grains, and the hand feels stiff and rough...

AI Technical Summary

Problems solved by technology

[0004] (1) Existing technologies and methods generally use visible light as the light source, which can only effectively identify some defects, and the quality discrimination accuracy is low

[0005] (2) There are many image interference factors

Due to the fluctuation characteristics of the light source and various interferences such as shadows, light reflection and absorption in the imaging process under industrial conditions, even pure white cotton may have uneven brightness and darkness in the imaging results, making general detection methods vulnerable to interference factors Impact

[0006] (3) There are seven types of cotton defects with complex shapes, which increases the difficulty of detection

Images