摘要
:研究了用人工神经网络对航空影像进行土地类型分类的方法 ,用选定窗口的平均值 ,标准差和傅利叶频谱平均值作为神经网络的输入 ,土地类型的标准值为神经网络的输出。将计算出来的未知土地类型标准值与已知土地类型的标准值相比较 ,可以确定出未知的土地类型。
A method for land cover classification of aircraft image using an artificial neural network was developed.Inputs to the neural network are means,standard deviations,and average values of Fouriet spectra calculated from a selected pixel window.Outputs from the network ate indices which represent unique land cover types.Knowing these indices,unknown land covers can be located by comparing an index of unknown land cover to that of the known land cover.
出处
《航空计算技术》
2000年第1期22-24,共3页
Aeronautical Computing Technique
关键词
人工神经网络
航空影像
土地类型分类
artificial neural network
image
mean
standard deviation
fourier spectra
