Research On Recognition Algorithm Of Cotton Foreign Fibers Based On Image

Foreign fibers in cotton mainly refer to the non-cotton fibers,which are mixed in cotton in the process of picking,drying,packaging,processing and transporting,such as hair,all kinds of polypropylene fiber,plastic film,rags,color lines,also known as the "three wire".Although the content of foreign fibers in cotton is low,the harm caused by them is great.The harm is mainly manifested in three aspects: in the process of carding,the foreign fibers can easily be crushed,leading to yarn breakage after mixed in yarn;in the process of printing and dyeing,the different dyeing rate between foreign fibers and cotton fibers will cause uneven dyeing and a large number of printing and dyeing defects,thereby affecting the appearance and quality of cloth;the defective products caused by foreign fibers seriously affect the reputation of the business,the quality of the textile and the market competitiveness.Therefore,it is very important to study a cotton foreign fiber recognition algorithm which satisfies the high recognition rate and real-time requirement.In this paper,the recognition algorithm of foreign fibers in cotton is mainly studied.Firstly,cotton foreign fiber images are preprocessed,then recognition experiments are carried on foreign fiber images using iterative global threshold method and the improved two-dimensional maximum between class variance method(Otsu)respectively,and recognition effects of two algorithms are contrasted.Finally,the software design of foreign fiber recognition algorithm is compeleted by using Matlab,the specific work contents are as follows:(1)Cotton foreign fiber images are collected by high speed CCD camera on the cotton foreign fiber testing device.(2)RGB color space is selected for experiments.Firstly,the gray histogram of the image is analyzed,according to the different characteristics between dark foreign fibers and white foreign fibers,in the subsequent recognition,dark foreign fibers are analyzed by using the gray histogram of the image,but white fibers are analyzed byusing the blue component histogram of the image.And then the gray value of the background is transformed so that the gray value range of the background falls within the gray value range of cotton,reducing the influence of the background on the recognition of foreign fibers.Finally,median filter is used to eliminate the noise in the image.(3)Several commonly used image segmentation methods are analyzed,and the iterative global threshold segmentation method is used to recognize the images of foreign fibers,the experimental results of several images are analyzed,and the recognition rates of various kinds of foreign fibers are counted.(4)In this paper,the improved two-dimensional Otsu method is used to recognize the foreign fiber images.Based on the traditional two-dimensional Otsu algorithm,the block strategy of the target and background is adjusted to reduce the running time of the algorithm.In the experimental part,a comparative analysis is carried on the running time and the foreign fiber recognition rate between the method and the iterative threshold method.(5)In order to facilitate user operation and improve the interactivity,the foreign fiber recognition interface is designed by using the Matlab graphical user interface(GUI),and the main functions of the image processing are also realized.