Research on Detection Method of Stored Grain Quantity in Granary Based on Index Relation and SVM

Abstract

In view of the urgent need and specific requirements of the national grain storage quantity online detection, according to the physical characteristics of grain bulk, combined with multiple regression, support vector machine and other technologies, the mathematical relationship between the bottom cloth pressure and the height of grain pile is studied. The feasibility of online detection of stored grain quantity by pressure sensor is feasible, and a granary based on exponential relationship is proposed In this paper, the detection model of stored grain quantity is constructed based on the combination of the detection value statistics of pressure sensor and support vector machine, and the specific modeling method is given. The method has been tested in real granaries of Tongzhou, Qihe and Xinan of Guangdong Province. The results show that the average maximum error is less than 0.07%. The proposed detection model has high precision and is suitable for various types of granary structures, and can meet the needs of remote on-line detection of stored grain quantity in granaries.