Intelligent Land Evaluation Research Based on Matlab and GIS

Applying the neural network to the land evaluation, we can break through the limitations that the traditional approaches are impacted by the human factors. Back propagation neural network (BP neural network) was used to evaluate the land suitability of the Changling town of the Guangshui city, Hubei province, China. We first establish evaluation index system, these indexes include the soil contamination degree, the irrigation guaranteed rate, the drainage condition, the pH value, the organic matter content. Then we establish the BP neural network and use MatLab to write the code forming the network. The evaluation criteria were input the network to train it. Then the network performance was test until the network meets the requirements. The evaluation data of the Changling town was input as the vectors to the appropriate network which calculates to get output vectors. And the output vectors were transformed the evaluation levels that can be imported the ArcGIS software to create the land suitability assessment figure. We can draw the conclusion that the suitability for the paddy field of the unused land and the arable land is very high and the ChangLin town is suitable for the development of paddy field agriculture.