Granary Storage State Detection Method And System Design

It is required to carry out large-scale stockpile inventory at present in our country every year that spends a lot of manpower and material resources to ensure the safety of grain storage quantity. In order to reduce the resources waste of the traditional granary storage state detection method that the development and application of modern granary storage state detection technology will have important significance. This paper focuses on the research of the detection method what is about grain storage status that based on SVM and decision tree, through modeling, data analysis, accurately detection of granary storage state, effective supervision and management of grain storage state to ensure the food security in maximum degree.This paper presents the view that detection method what is about grain storage status that based on SVM and decision tree. Upon the basis of studying the fusion technology of SVM and decision tree, this paper mainly studies the attribute importance of the effective measurement method in decision tree learning, and proposes an improved algorithm for decision tree construction. The main research work is as follows:(1)Firstly, it proposed the state of granary storage model based on SVM, on the basis of the transformation of the three categories problem of grain storage status detection into two categories. It constructed two kinds of support vector classification which are grain input and grain output, and according to the features of selected feature vectors and the output value of support vector machine constructs the granary storage state classification rules. The experimental results show that the proposed, the two kinds of support vector classification and the granary storage state classification rules, is feasible and effective.(2)Secondly, aimed at the existence problems of complex classification model massive calculations of support vector machine, the fusion technology of SVM and decision tree is proposed based on the advantages of support vector machine and decision tree, through the effective measuring of the importance of attribute in learning decision tree to improve the effect of classification. The experimental results show that compared with the C4.5 algorithm, the proposed algorithm can effectively improve the detection accuracy of grain storage state.(3)Thirdly, with the visual studio 2010 C + + development environment and My SQL 5.6 database management system constructed granary storage state detection system which based on SVM and decision tree. Through calculatingand analyzing the obtained data from pressure sensor to realize the correct classification of the grain storage state.