Hyperion Forecast Model Based On Fuzzy Reasoning For Rice Threatened By Lead Contamination

In the farmland ecosystem, heavy metal lead is an extremely typical pollutant full of potential harm. Lead will not only affect the quantities and quality of the products cultivated in the farmland, but also harm the human health mostly by the means of the enriched food chains. Therefore, the study of condition that crops threatened by lead pollution is vitally significant for the prevention of food contamination and protection of human health.In this paper, as a project supported by the National 863 Project and the National Natural Science Foundation, three independent rice fields, located in Dongqiao Town and Zhengyi Town, Suzhou in Jiangsu province, were the areas chosen for the experiments. Focused on the hyperion remote-sensing extracting algorithm of information that rice threatened by lead pollution and integrated with the fuzzy mathematics theory, hyperion fuzzy forecast model of rice threatened by lead contamination was constructed.The former part of this paper was about data acquisition and spectral data process and analyses. From three independent experimental rice fields of different lead contamination, necessary data of rice in every growing stage were collected, and then, on the use of continuum removal method, fractal analyzing method and also other methods, rice canopy spectral data were analyzed. The spectral featured parameter (Da) and the fractal dimension were finally acquired which is highly relevant to the lead content of rice.The latter part of this paper was about the construction of the hyperion fuzzy forecast model of the content of lead contamination in the rice. The model was developed in the following way: focused on the spectral featured parameter (Da), the fractal dimension and the pH value of the soil, using the fuzzy set theory in fuzzy mathematics, based on the fuzzy inferring processing order of the three featured parameters and the theoretical basis that the three parameters among the three fuzzy sets with three different levels of lead contamination have uncertainty and fuzzy characteristic, membership functions of three featured parameters were set in proper sequence and the fuzzy weights among parameters were determined, and eventually the model was constructed.Comparing the predicted value to the true value of lead content in rice, in the three different fuzzy sets of the content of lead pollution, the multiple correlation coefficients between the predicted value and the true value were 0.9565, 0.8578 and 0.9145,which was a high level and consistent with the expectation of the model establishing requirements. Therefore, the hyperion fuzzy forecast model of lead contamination in the rice developed in this paper was logical and reasonable, and the value predicted by the forecast model was relatively accurate and correct.