Research On Key Technologies For Detection Of Heterogeneous Fiber Content In Cotton

In view of the low degree of automation in the process of evaluating the content of foreign fiber in cotton and weight ratio evaluation of foreign fiber content index is too single,a foreign fiber content detection system is designed and a model for evaluating the content of foreign fibers based on the grade of yarn faults was established.The mechanical structure of the foreign fiber content detection system is designed.The light source in the foreign fiber image acquisition part is modeled,the optimal light source brightness and model is solved and the optimum location of the light source is calculated.The control part completes the location of the foreign fiber,and the extraction program is written to realize the real-time collection of the foreign fiber images and the extraction of the foreign fiber.In order to obtain the area S,the length L of foreign fiber and the identification data of the different types of foreign fibers,the image processing algorithm of the foreign fiber is studied.The maximum value method is the optimal method for the graying of the foreign fiber images.The median and mean filter are used to denoise the gray image,and finally the median filter of 5 x 5 template is selected by comparison.Several threshold segmentation methods are compared,and large law is selected through screening and comparison.After a series of image processing methods,five clear kinds of threshold segmentation images of foreign fibers are finally obtained,which lay a good foundation for the accurate extraction of all kinds of foreign fiber features.Based on the images of five common foreign fibers(hemp,hair,wool,polypropylene filament and feather),the area and length parameters of the foreign fibers in the content parameter equation are obtained based on the edge detection operator.The color characteristics of the foreign fiber are selected and analyzed to effectively distinguish the color characteristics;Through the programming to obtain the different shape features of foreign fiber;The characteristics are analyzed and the effective features are screened out.As many as 21 kinds of characteristic parameters of the foreign fiber are extracted comprehensively,and the foreign fiber types are identified by using support vector machine and neural network.In this paper,according to the different area,length,quantity,type and different yarn faults of different foreign fiber,we get the yarn faults data of different foreign fiber through experiment,analyze of factors affecting the rating parameters of foreign fiber,calculate the multiple parameter regression equations of different foreign fiber,and according to the yarn faults equation,the parameter equation of the foreign fiber content is established.The evaluation of foreign fiber content gradually change from weight index to more reasonable species and its corresponding yarn faults series and area and other indicators.The evaluation criterion of cotton foreign fiber content is more scientific.It is proved by experiments that the evaluation of the grade of foreign fiber content in this paper is feasible,and has high theoretical value and practical value for the classification evaluation of many kinds of test objects in other related fields.