A Study On The Evaluation Of Cultivated Land Quality Based On BP Neural Network

Evaluating the quality of cultivated land has an important significance for the management of quality and quantity of cultivated land, the efficient use of land resources, and the development of sustainable agricultural. Most of theoretical circles choose "Agricultural Land Classification of order" to evaluate the quality of cultivated land, but the research of intelligence, information technology is relatively small. According to this, the study, based on the BP neural network theory, builds an appropriate model to calculate the mass fraction of the quality of cultivated land to explore the method of evaluating the quality of cultivated land based on BP neural network taking Gong’an county in Hubei Province as an example.These studies have shown that artificial neural network model is a new infor mation processing technology which could be applied to the arable land in the na tural quality evaluation. BP neural network could reduce the error resulting from t he improper selection of model function and subjective factors on the evaluation r esults. According to the nonlinear relationship between the score of the study sam ple automatically map and its influencing factors of the quality of arable land nat ural, the model could give a more scientific prediction of arable land natural qual ity and make the evaluation results more scientific and accurate. Generally speakin g, the natural quality of the cultivated land in Gong’an county is excellent, which has the rules District flake distribution. The total area of Level Ⅰ and Level Ⅱ1and accounted for65.63%, of which the second accounted for56.11%, which mai nly distributed on both sides of Yangtze River (tributary) to western part of the c ounty and main road of Yangtze River in eastern part of the county. The total ar ea of Level Ⅲ and Level Ⅳ are mainly distributed in the central and northern p arts of Gong’an County. The area of Level Ⅴ accounted only for6.85%of the to tal area, which mainly located in the south and southwest of Gong’an County. In addition, the study also found that differences in ability of the evaluation factors distinguish arable land natural quality is not always its factor of importance or i nfluence a positive correlation. Important role of an evaluation factor to the natur al quality of the evaluation unit of arable land does not necessarily means high s ensitivity. Taking Gong’an County for example, the order of sensitivity for natural quality results:irrigation guarantee rate> drainage conditions> element content, profile form, surface texture, degree of salinization, organic matter content of soil pollution, soil pH, effective soil layer thickness> potential productivity by radiation and temperature.The characteristics of this study lies in the fact that, in the evaluating the qu ality of cultivated land research field, the BP neural network model is introduced, and the system was analyzed and studied, including network model selection, n etwork structure design, construction, algorithm and other samples.