Detection Model And System Implementation Of Granary Storage Quantity Based On SVR

A necessary prerequisite for ensuring the safety of grain quantity in China is to grasp the distribution of grain storage amount accurately and timely in various regions.At present,the way of counting the granary storage situation in various regions is mainly weighing method.Weighing method is not only time-consuming and laborious,but also the annual inventory work will cause different levels of food losses.Therefore,in order to solve the drawbacks of the traditional statistical methods,it is of great significance to develop a fast and accurate grain storage quantity detection system.This paper introduces the technology of Internet of Things and intelligent decision support and so on,and establishes the detection model of grain storage quantity based on SVR and develops the corresponding detection system.This paper realizes the goal that the detection error of grain storage quantity is less than 3%,and satisfies the requirement of acquiring granary storage quantity in real time and dynamically.This paper focuses on the application of SVR algorithm in detecting the weight of granary storage.The system uses the pressure sensor network to collect the bottom pressure of the granary and uses the SVR algorithm to do the regression prediction of the grain weight.The main contents of this paper are as follows:(1)The basic theory of SVR is summarized in mathematical language based on the basic concept of statistical learning theory.This paper analyzes the advantages and disadvantages of three commonly used regression algorithms,and discusses the characteristics of these three algorithms and SVR in combination with the requirements of this topic.Finally,the choose basis of SVR method is given by comparison and summary.(2)This paper studies the relationship between the granary grain weight and the distribution of grain pressure,and establishes the pressure sensor detection model based on the bottom and side of the granary,and gives the layout model of the pressure sensor.Aiming at the problem that the output mean value of the sensor fluctuates greatly,a SVR detection model based on the pressure mean value of the sensor is proposed.Aiming at the problem of large prediction error,a SVR detection model based on sensor output sequence is proposed.(3)This paper designs the hardware and software modules of the detection system andrealizes the overall function.The system consists of pressure sensor,data concentrator,gateway,SVR detection model and detection software.This system can detect granary storage quantity in real time,which effectively improve the work efficiency of the national grain regulatory authorities.