The Research On Compress Sensing Technology For Grain Information

Food security is of great significance for economic development and social. The improper grain storage will lead to huge losses."Food Technology’ Twelfth Five-Year’ development plan" points out that temperature sensor, humidity sensor, quality sensor, quantity sensor, gas sensor, mold sensor, pest image sensor should be R&D, which can gradually realize the goal that the food stocks information intelligent control is given priority to development of technical field. Compressed sensing is a new sampling technique which breaks the restraining of original Nyquist sampling theorem, what’s more, sparse representation theory extended from compressed sensing has better recognition efficiency compared with BP, SVM. This paper mainly do research on grain information processing based on compressed sensing, which contains grain temperature compressing and grain pest recognition. The paper is divided into six chapters, the specific work and innovation is summarized as follows:This paper firstly summarizes the research background and significance, and briefly introduces the compressed sensing, temperature and humidity monitoring, grain pests online monitoring; In the second section, this paper introduces compressed sensing basic theory, research focus, application examples, and in the last section, combining with some examples, some application was given. Third chapter describes the usage of compressed sensing in grain storage temperature information processing, which contains obtaining temperature information, the specific strategies used in the compressing grain temperature information, the temperature sparse effect in different sparse domain, in final, a new method is presented called " grain temperature monitoring error control method based on limited feedback", which was given some theory analysis combining with simulation. In the forth and five chapter, the paper discuss grain pest recognition based on sparse representation classify. The four chapter introduce pest image processing, step of feature extraction, and then sparse representation is introduced, new grain pest classify model is presented, some proof of theory is given based on gaussian matrix, the simulation show that new model can get good result. The last section introduce the theory model of convex optimization and greedy algorithm Base on those theory, a new combinational algorithm is presented. In the last section, A summary of the article is given, and the work of the future is made a simple expectations.