基于多层感知器神经网络对遥感融合图像和TM影像进行土地覆盖分类的研究

Classification for RS Fused Image and TM Image Using Multi - Layer Perception Neural Network

８０年代以来，人工神经元网络（ＡＮＮ）技术的应用不断向广度和深度发展．用不同空间分辨率的ＴＭ和ＩＲＳ遥感图像进行融合，综合了不同传感器数据所提供的信息，增强了图像的清晰度，改善了解译效果。用多层感知器神经网络对遥感融合图像进行分类，分类精度达９５％，比用多层感知器神经网络对ＴＭ图像进行分类（分类精度达７３％）效果要好．

Artificial neural network technology was greatly developed on many aspects. The fused re- mote sensing image synthesized information from different remote sensor, and its characteristics were strengthened. The classification for fused RS image based on the multi - layer perception neural net- work is far better than that for TM image. The classification accuracy for fused RS image is up to 95%, higher than that for TM image, which is 73 %.