基于神经网络的黄河三角洲土地利用遥感获取技术研究

利用高光谱分析技术和神经网络图像分类方法对黄河三角洲1996、1998、2000和2002年LandsatTM5图像3、4和5波段假彩色合成图像进行信息提取，采用高斯分布布点的观测点进行试验地验证反馈改进算法精度，获取黄河三角洲地区土地利用状况。结果表明，黄河三角洲地区土地利用类型所占比例分别为耕地33％、林地13％、居民地5％、水域27％、盐碱地18％和未利用地4％，土地利用率为51％．土地垦殖率为46％。

Study on Application of Remote Sensing Image in Land Use of Yellow River Delta Based on Neural Network

The spectral analysis technology and neural network classification method were used in the false color synthesis image made by Landsat TM3,4,5 wave bands to get the type of the yellow river delta land utilization,then the algorithm accuracy was tested by the Gaussian distribution monitoring points to get optimum neural network classification,the types of the yellow river delta and utilization were got by statistics method of GIS.The results indicated that the percentage of land use types in yellow river delta were as follows: arable land 33%,woodland 13%,resident land 5%,waters 27%,saline land 18% and unutilized land 4%,land utilization ratio and land reclamation rate were 51% and 46%,respectively.