Research On Key Techniques Of Cotton Foreign Fiber Recognition Based On Image Vision

Foreign fiber in cotton(referred to as "foreign fiber")refers to non-cotton fiber and color-changing cotton fiber mixed in the whole process of production,transportation,processing,etc.Commonly there are mulch film,colored thread,various types of polypropylene yarn,hay Film etc.Although the foreign fiber content is small,it is extremely harmful,which not only brings huge economic losses to the country and enterprises,but also seriously damages the credibility of China's cotton processing products.Therefore,efficient and accurate detection of foreign fibers in cotton is essential..At present,automatic detection equipment based on images and vision is widely used in various fields due to its advantages of easy expansion,high efficiency,and high accuracy.Therefore,this paper presents the key technology research of cotton foreign fiber recognition based on image vision.Mainly complete the following aspects:1.Determine the technical requirements and main technical parameters of cotton foreign fiber detection through full investigation of domestic and foreign products;build a foreign fiber detection system based on the modular design principle;according to cotton flow rate,image resolution,image transmission rate The main technical parameters complete the optimization design of key parts and the selection of important parts;2.On the basis of fully researching the traditional one-dimensional Otsu algorithm and the two-dimensional Otsu algorithm,it is concluded that the characteristics of the traditional algorithm do not meet the requirements of automatic online detection technology;improve the blocking strategy of the two-dimensional Otsu algorithm,using upper and lower double boundaries Asymmetric histogram oblique division method,using the edge operator and the least square method to determine the upper and lower limit intercept value and adjust the detection target and detection background;perform noise replacement to improve the algorithm's noise resistance;3.Use texture feature images and use texture parameters as thresholds to identify foreign fibers with thresholds similar to cotton.By analyzing the influence of changes in step d,direction ? and gray level L on texture parameters,a reasonable value is determined;a method for unequal compression of gray levels of texture images is proposed and compared;The rationality of the image recognition for the threshold and the generation of texture images to select the appropriate texture parameter value as the segmentation threshold,based on the Otsu algorithm for image recognition and verification;4.On the basis of the Qt framework,the system software is designed and implemented with C++as the main programming language;the main functions are encapsulated,the instructions from the driver and user settings are processed in a thread-exclusive manner,and the login interface of the software is introduced.Main window interface,parameter setting interface,image analysis interface.