A Granary Quantity Detection Model Based On The Fusion Of ELM And SVR

Grain quantity security is closely related to grain security.In order to ensure our grain storage quantity security and improve grain macro-control ability,it is of great significance for China to research and develop a kind of national grain storage quantity monitoring technology which is convenient,fast and accurate online and networked.Based on systematically studying the existing methods of grain storage quantity detection and aiming at the theoretical and technical problems of pressure data preprocessing and detection modeling involved in the detection of grain storage quantity based on pressure sensor,this paper studies the theoretical and technical problems of data preprocessing and the construction of grain storage quantity detection model based on machine learning theory.In this paper,the data preprocessing method based on the improved ant colony algorithm is explored for the single ring pressure sensor layout in the bottom of the granary.Meanwhile,this paper studies the detection model construction of the granary storage quantity based on the extreme Learning Machine and support vector regression machine.The main research includes:1.According to the arrangement of the pressure sensor in the single ring at the bottom of the granary,this paper analyzes the output data characteristics of the pressure sensor,studies the relevant rules of data selection and proposes a data preprocessing method of the pressure sensor based on the improved ant colony algorithm.By finding the maximum linear independent group of the target matrix to reduce the dimension of the data,the selection optimization of the output value of the pressure sensor is realized.The practical application results show that the proposed method is feasible and effective.2.The method of dividing the sample set based on SPXY algorithm is introduced.The method of constructing the model of grain storage quantity detection based on support vector regression,BP neural network and extreme learning machine is given.The selection of modeling parameters is studied.The detection accuracy of three detection models is compared and analyzed.3.Based on extreme learning machine and the support vector regression machine,amethod building the model of grain storage quantity detection is proposed.The improved ant colony algorithm is used to optimize the output value of the single loop pressure sensor.The data set is reconstructed by the limit learning machine.The data regression is realized by the support vector regression machine.The comparative evaluation results of modeling results verify the rationality and effectiveness of the proposed method.