基于神经网络-证据理论的遥感图像数据融合与湖泊水质状况识别

为了进一步提高湖泊水质状况识别的准确性，提出了一种基于神经网络．证据理论的遥感图像数据融合处理方法，并以太湖水质监测数据为例进行了实证分析。该方法先对不同的遥感输入图像，采用各自相应的神经网络进行处理，然后对神经网络输出的结果做归一化处理，再利用D-S证据理论进行数据融合，最终给出水质的识别结果。该方法的优点为(1)可增加水质识别的容错性；(2)由于融合了多源水质遥感图像的数据，因而水质状况识别的可信度更高。

Remote Sensed Images Fusion and Lake Water Quality Identification Based on Neural Networks and Evidence Theory

In order to identify the lake water quality accurately, this paper presents a method for remote sensed image fusion based on neural networks and evidence theory. This method firstly employs a neural network for each remote sensed image and then normalizes the output of neural networks. After that, D S evidence theory is used to fuse with results from all the neural networks, resulting in the water quality evaluation. The proposed method is applied to the water quality of Taihu lake. The developed approach to water quality identification has the two features:(1) low fault tolerance; and (2) high reliability as multi source water quality data are fused.