| Rice is one of the important food crops for human survival,occupying an important position in China’s agricultural production and national economy.Weedy rice,as a wild rice,seriously affects the growth of cultivated rice and greatly reduces the quality of rice.Preventing weed rice seeds from entering the sowing process is of great significance for improving rice yield.Due to the extremely similar morphological characteristics between weedy rice and cultivated rice,the efficiency and accuracy of manual recognition are low.Therefore,this paper proposes an efficient and accurate method for sorting weedy rice seeds based on machine vision and pneumatic conveying technology,and develops an intelligent weedy rice seed sorting device to improve the recognition and sorting accuracy of weedy rice seeds.The main research work of this article is as follows:(1)Develop an intelligent weed rice seed sorting equipment.The equipment mainly consists of a Delta robotic arm,an image acquisition module,a transmission speed control module,a removal and conveying module,and a control system module.During the operation,various components coordinate and cooperate to suck weed rice seeds into the collection box,achieving intelligent sorting of weed rice seeds.Process and test the prototype according to the structural design.A calibration experiment was conducted on the camera and Delta robotic arm,with randomly selected weed rice seeds uniformly distributed on the conveyor belt as the calibration target.The calibration results showed that the Delta robotic arm met the requirements for locating weed rice seeds;And conduct identification,speed regulation,positioning,removal,and conveying performance tests on the prototype.Finally,conduct a complete machine test on the prototype,mixing four types of rice seeds and weed rice seeds as samples to test the sorting effect;The results showed that the sorting accuracy was above 95%,meeting the sorting requirements for weed rice seeds.(2)Propose a weed rice seed sorting method based on image processing technology.Firstly,four rice varieties,Linhan 1,Danhan 53,Yuangan Rice,and Zhenghan 10,as well as weed rice seeds,were used as samples to measure the material characteristics of 1000 grains of each of the above varieties,and the average pixel values of the HSV and RGB channels of the rice seeds were measured;Secondly,the aspect ratio of weed rice seeds is used as the shape feature,and the R channel pixel proportion is used as the color feature.Two major features are used as the judgment targets for weed rice seed recognition.For adhesive seeds,the concave point segmentation method is used to segment the adhesive rice seeds in the image,completing the recognition of weed rice seeds in the collected image and accurately obtaining the position information of weed rice seeds.(3)Research on parameter optimization of conveying system based on gas-solid coupling.Establish a three-dimensional model of the conveying hose and collection box through Solidworks,and simulate the motion trajectory of weed rice seeds under airflow using EDEM FLUENT coupling;Based on pressure cloud map,velocity cloud map,and particle trajectory map,determine the optimal pipe diameter and flow velocity range;A three factor and three level orthogonal experiment was conducted with the minimum values of pipe diameter,flow rate,and seed particle spacing as factors,and the suction rate and number of missuction particles as evaluation indicators.The results showed that when the pipe diameter was 20 mm,the flow rate was 20 m/s,and the seed particle spacing was greater than or equal to 15 mm,the optimal transportation parameter was obtained. |
PDF Full Text Request