| As a necessary equipment in rice processing,rice milling machine plays an irreplaceable role in improving grain utilization rate.The traditional adjustment method of rice processing precision is that the operators observe the skin retention and broken rice rate of the rice by naked eyes,and then adjust the position or weight of the weights to change the whitening pressure and the opening degree of the rice outlet.Artificial observation is not accurate and the weight adjustment is not timely will cause bad quality of the rice and economical loss of the rice milling factory during operation.Therefore,taking Tongtai MNJ180 rice milling machine as the object,a set of rice processing quality detection and feedback control system according to machine vision was designed,in order to achieve automatic control of processing precision.The specific research work is as follows:(1)Design of on-line detection device for rice.The physical property parameters of rice were determined.According to the measured parameters,a sampling and dispersing device was designed to make a small amount of rice fall into the decimeter bucket through the sampling hole.During the sliding process,the rice was dispersed by the decimeter block,the decimeter guide groove and the disperse guide groove without accumulation and a large number of adhesion.After comparison of tests,the conveyor belt was selected to horizontally transfer rice to the image acquisition area,and industrial brush was used to dust the conveyor belt.According to the requirements of image acquisition,the models of industrial camera and lens were determined.After lighting test,the detection light source is determined to be white ring light source,and the lighting mode is front irradiation type.(2)Extraction of rice processing quality characteristics.By comparison of experiments,Gaussian filtering,OSTU threshold segmentation and improved watershed algorithm were used to denoise the image,segment the background and separate the adhesive rice particles,respectively,to realize the region extraction of single grain rice.After calibration,the Pixel accuracy in the image is about 29.3 Pixels/mm.Through comparison and analysis,the methods of image subtraction,fixed threshold segmentation,minimum enclosing rectangle and LOG transform algorithm were used to extract rice skin,chalkiness,broken rice and burst waist area,respectively,to realize the feature extraction of rice processing quality.(3)Quality inspection of rice processing.Five images were collected continuously and feature extraction was carried out to achieve the processing quality detection results of rice.It was showed that the accuracy rate of processing accuracy,chalkiness,broken rice and burst waist defect was 91.63%,93.75%,94.76% and 72.97%,respectively.The average detection time of a single image was 754 ms.VS2015 was used to write the rice processing quality detection software to judge the processing quality of rice.(4)Design of feedback control system.A set of control device for the opening of the rice outlet was designed,which converted the rotating motion of the ball screw motor into the horizontal movement of the discharge door through the inner sliding sleeve,and realized the control of the opening of the rice outlet.A rice milling machine control system was built,When the rice milling machine is in normal operation,the working state of the rice milling machine was adjusted according to the results of rice processing quality discrimination.When the rice milling machine was abnormal,the piston rod of the cylinder contracted,the discharging motor stopped running,and the rice pushed the sliding sleeve to complete the rapid pressure relief. |
PDF Full Text Request