| Food has always been the basis for human survival.Grain,as one of the important members of food,can be stored effectively for a long time only if its internal moisture is lower than the national safety standard.In the process of grain processing and transportation,the internal moisture must be strictly controlled to avoid serious economic losses and mildew waste.Therefore,it is of great value for grain storage to detect the moisture content of grain quickly and accurately.At present,the grain moisture content detection instruments on the market generally have the problems of high cost,time-consuming detection or too large to carry.Therefore,this paper designs a portable detection device for grain moisture content based on microwave transmission method and microstrip microwave sensor.The specific research contents are as follows:(1)A real-time detection device for grain moisture content was designed based on the microwave transmission method.The whole detection device is composed of three parts:high-frequency microwave circuit module,low-frequency DC circuit module and real-time measurement module.The signal receiving and processing module designed with STM32F103RBT6 chip as the core realizes the functions of control signal A/D conversion,microwave attenuation and phase shift calculation,grain moisture content calculation,temperature acquisition and display,screen control and serial communication.In order to improve the problem that the detection equipment is too large and difficult to move,this paper proposes to use microstrip antenna instead of horn antenna,which makes the whole equipment very small in size and light in weight.The size of the complete unit is 42 cm × 32 cm × 16 cm,weight 4.05 kg,greatly meeting the portable requirements of grain moisture content detection equipment.(2)Aiming at the problem that the measurement of different grains will lead to unstable and inaccurate detection results,this paper analyzes and processes the data obtained from different kinds of grains by learning eight different machine learning algorithms,namely KNN,ridge regression,SVM,decision tree,random forest,GBDT,XGBoost and Light GBM.Microwave attenuation,phase shift,and temperature compensation were selected as the input variables to the prediction models,which have achieved high accuracy.To find the corresponding optimal model parameters and embed them into the developed equipment for real-time measurement to meet the requirements of real-time and accurate measurement of the detection device.(3)In order to meet the reusability of the R&D equipment,experimental studies on rice,soybean and wheat were carried out respectively.Compared with the drying method,the maximum MAE(mean absolute error)of the measurement results of this device is 0.55%,the minimum is0.17%,the maximum SD(standard deviation)is 0.41%,and the minimum is 0.08% in the rice experiment study;In the study of soybean,the maximum MAE of the measurement results of this device is 0.52%,the minimum is 0.37%,the maximum SD is 0.23%,the minimum is 0.11%;In the experimental study of wheat,the maximum MAE of the measurement results of this device is 0.33%,the minimum is 0.19%,the maximum SD is 0.25%,and the minimum is 0.08%.The results show that the device can be well applied to the moisture content detection of rice,soybean and wheat.The portable detection device for grain moisture content is small in size,light in weight,and the detection results are fast and accurate.It provides an important reference for the real-time measurement of agricultural products,the development of intelligent agricultural equipment and industrial applications. |
PDF Full Text Request