Method and system for selecting feature of cotton heterosexual fiber target image

Abstract

The invention discloses a method and system for selecting the feature of cotton heterosexual fiber target image. The method includes: conducting the gene encode for the to-be-selected color feature, shape feature and texture characteristic of the cotton heterosexual fiber target image by using the sectional type binary encoding project and generating the initial species group randomly; calculating the sufficiency of the individual in the initial species group; selecting the initial species group according to the sufficiency to obtain new species group; conducting the sectional intersect and variation for the individual in the new species group to produce new individual and substitute original individual, calculating the sufficiency of the new individual; judging whether the sufficiency conforms to the stopping condition; if the sufficiency conforms to the stopping condition, encoding the individual with highest sufficiency to obtain the optimum feature subclass of the cotton heterosexual fiber target image. In the invention, the overlay of different feature in the cotton heterosexual fiber target image is realized, the optimum feature subclass with the optimum classifying capability is selected, thus being beneficial for the subsequent realization of the quick right classification of the heterosexual fiber.

Description

technical field [0001] The invention relates to the fields of pattern recognition and machine vision, in particular to a method and system for feature selection of cotton foreign fiber target images. Background technique [0002] Cotton foreign fibers refer to non-cotton fibers and colored fibers that are mixed into cotton during the process of cotton picking, tanning, and purchase, which have a serious impact on the quality of cotton and its products, mainly including polypropylene yarn, hemp rope, and hair. Although the content of foreign fibers in lint is small, it has a serious impact on the quality of cotton textiles. Once mixed and involved in weaving, it will affect the strength of the yarn, and affect the coloring during dyeing, which will cause significant economic losses to the cotton textile industry. [0003] At present, it is possible to reduce the content of foreign fibers in cotton by picking foreign fibers, or take samples from cotton production and send them...

AI Technical Summary

Problems solved by technology

[0004] In the process of realizing the present invention, the inventor found that the feature selection method in the prior art has at least the following technical defects: the two key technologies in the feature selection method, "search strategy" and "evaluation function", have defects such as large amount of calculation

Among them, the genetic algorithm in the search strategy tries to avoid falling into the local optimal solution through randomness, and can obtain the global optimal solution with a high probability; but the current genetic algorithm still has some shortcomings, such as premature convergence problems, etc. In particular, it does not consider the rationality and necessity of localized management of features with completely different properties in the target image of cotton foreign fibers

If the feature selection method of the above-mentioned genetic algorithm is used, the efficiency of feature selection will be low, and foreign fibers cannot be quickly and correctly classified, so that it is impossible to quickly and accurately determine the foreign fiber content in seed cotton in the seed cotton purchase or trading process, and cannot pass the detection of seed cotton. The purchase price is affected by the content of foreign fibers in cotton, which cannot effectively cut off the way of mixing foreign fibers, which may cause major economic losses to the cotton textile industry.

Images