基于WorldView-02高分影像的BP和RBF神经网络遥感水深反演

遥感水深反演是水深测量的一种重要技术和手段.以美济礁水深反演为例，选择WorldView-02高分影像为数据源，在辐射定标和大气校正的基础上，构建BP（Back Propagation）和RBF（Radial Basis Function）人工神经网络水深反演模型，以遥感影像8个波段为输入层，通过tansig、logsig、高斯函数和purelin函数变换实现从输入层到隐含层、隐含层到输出层的转换，以便反演水深.最后对反演水深与实测水深采用回归分析，求解决定系数（coefficient of determination，R2）、平均决定误差（Mean Absolute Error，MAE）、均方根误差（Root Mean Square Error，RMSE）等进行比较，评价2种模型的精度.结果表明，RBF神经网络模型结构更简单，对样本要求更低，反演精度达到0.995，更适合遥感水深反演. 

Inversion of Water Depth from WorldView-02 Satellite Imagery Based on BP and RBF Neural Network

The inversion of water depth from remote sensing imagery is an important technology of depth measurement. In this paper, on the basis of radiometric calibration and atmospheric correction, BP(back propagation)and RBF(radial basis function) neural networks were built to retrieve water depth from WorldView-02 high-resolution satellite imagery in Mischief reef. Band 1 to band 8 of satellite imagery were used as the input data of the neural networks. Then, they were converted from input layer to hidden layer and from the hidden layer to output layer with tansig, logsig, Gaussian and purelin functions. Finally, the accuracy of the two models was evaluated by R2 (coefficient of determination), MAE(mean absolute error), RMSE(root mean square error) and the regression analysis between retrieved water depth and ground measured water depth. The results show that RBF neural network has simpler model structure, and lower requirement of samples. Besides, its retrieval accuracy reaches 0.995. Therefore, RBF neural network is more suitable for the inversion of water depth.