基于ANN分类的农田遥感动态监测模型研究

保护基本农田是我国农业可持续发展的基础和前提,动态监测基本农田在时间和空间上的变化能够为农业开发政策的制定,农业经济发展的规划与管理提供有效的辅助决策手段。论文利用人工神经网络的BP算法实现了对两个时期的遥感影像进行基本农田类型的分类提取,在保证精度的前提下,探索了一条把单要素监测和多要素监测相结合的遥感动态监测模型,并详细描述了模型实现的算法与步骤。最后利用该模型对实验区进行了监测,并对监测结果进行了分析,结果表明模型很好地评估了研究区基本农田的数量和发展潜力,定性、定量、定位地揭示了研究区基本农田类型在时空上的变化规律。

Studies on Remote Sensing Dynamic Detection Model of Cropland Based on the Classification of Artificial Neural Network

To protect basic cropland is the precondition of the agricultural sustainable development in China,dynamically detecting the change of cropland on a spatio-temporal scale can help to make out the agricultural development plan and managing the agricultural economic development.This article discussed the classification method of the remote sensing image using the BP artificial neural network, and on the condition of the precision of the classification, it explored a remote sensing dynamic detection model comprised of single component detection and multicomponent detection,and described the algorithm about it in detail. Finally, an application of this model has been shown in an experimental area,and the result of the application has been analysed with the real data. It is shown that this model has excellently evaluated the amount and potential of the basic cropland in this experimental area, and opened out the change regular of the basic cropland in the experimental area on a spatio-temporal scale in a qualitative,quantitative and orientative way.