Research On Real-time Detection System For Paddy And Brown Rice Flow Based On Machine Vision

Rice is one of the main edible crops in China,and it is also one of China’s major export crops,which plays a significant role in the development of China.In the processing of rice,almost every processing process has a flexible demand for material flow,and the flow rate needs to be adjusted according to the processing conditions of each step.At present,there is almost no mature machine vision detection method in this respect.It mainly relies on manual observation of the working conditions of each processing step,however,this is not only easy to cause worker fatigue,but also easy to cause material blockage and overflow in the previous process because the failure of a certain process is not detected in time.Not only affects the processing efficiency,but also easily causes the waste of human resources.In view of the above requirements,based on the machine vision technology,the flow detection in the grain separation is the main research content.According to the difference between the grain roughness and other objects,the identification and flow measurement are carried out,and the classification is trained by artificial neural network.The main research contents are as follows:(1)According to the working environment of the grain separator,a reasonable image acquisition system was built,including the choice of light source,illumination mode,industrial camera,optical lens and programming language.(2)Based on the HSV color space to identify the grain roughness,the canny algorithm is used for edge extraction,and the state of storage of the grain in the feed hopper and the feed trough is divided into three.(3)Two algorithms that directly perform flow measurement by function and indirectly complete flow measurement based on a given reference object are used.(4)A classifier for accommodating grain in the feed hopper and feed trough was trained by constructing an artificial neural network.