Research On Optimization System And Parameter Calibration Of Foreign Fiber Sorter

Foreign fibers in cotton fabric affects the appearance,reduce product quality,serious damage to the reputation of the enterprise,bring huge economic losses to cotton spinning enterprises.At present,cotton textile enterprises mainly use fiber optic sorter based on CCD optical imaging to detect foreign fibers,but the detection results for many kinds of foreign fibers are not ideal.This paper aims at the 60AT4 type foreign fibers sorting machine produced by Beijing Daheng image Co.Ltd,in order to improve the detection rate of sorting machine for many kinds of foreign fibers,its main contents are as follows:In this paper,an optimal design method based on fuzzy clustering neural network is presented.By analyzing the relationship between CCD imaging and the incident light energy,the number of light sources derived from multi class fiber detection;established the CCD target surface exposure function,to determine the optimum position for the source and based on the analysis of the light distribution and the average gray level of the CCD background image,a fuzzy clustering neural network is established to optimize design of light source.The experimental results show that the detection rate of the optimized fiber reaches 94.79%.In foreign fibers detecting,adjust the white balance camera software not only requires a certain expertise,and the operation is complicated,need of exposure time and gain multiple combination to get the white balance adjustment of the final result.Therefore,this paper presents an automatic white balance method for CCD camera based on simulated annealing algorithm,by analyzing the relationship between the camera gain and exposure time and white balance value,the regression equation is established.Simulated annealing algorithm is used to realize automatic white balance adjustment.Foreign fiber sorting machine detection system for different colors and shapes of foreign fibers,according to the actual situation needs to be adjusted constantly,this process is time-consuming and laborious,and it is difficult to achieve the desired results.In this paper,the common size fibers is taken as the experimental object,the data samples are collected,and the optimal detection parameters of each fiber are predicted by neural network model.Under the condition of optimal detection parameters,the calibration accuracy of the fiber sorter is achieved.The results show that the optimal design of the light source and the automatic white balance method are feasible.The precision calibration of the fiber sorting machine has reached the requirements of the actual production.It is of high theoretical value and practical value to optimize the equipment in the field of machine vision.