摘要： 在卫星影像上，一些城区和森林区的公路由于障碍物和阴影的影响，常导致道路出现模糊路段或断裂现象，单一方法很难检测出这些模糊路段，该文针对这种特殊道路提取难题，提出了一种结合模糊C 均值（FCM）算法、反向传播神经网络（BPNN）和图像处理方法的混合智能模型。实验结果表明，混合智能算法可以得到比单一的神经网络方法和知识处理方法更好的道路提取结果。

关键词: 道路识别；卫星图像；模糊C 均值；神经网络；知识处理；图像处理

Abstract: Road extracted from satellite imagery has been used for many different purposes, e.g., military, map publishing, transportation, and car navigations, etc. Many methods, such as, neural network, knowledge-based, optimal search, snake model, semantic model, road operator model, etc. are researched to identify road from satellite image, but because of complicated characteristics of road and image itself, and automated road network extraction still remains a challenge problem, and no existing software is able to perform the task reliably. This paper presents a hybrid method which combines Fuzzy-C-Means with back-propagation neural network and knowledge processing technique to detect roads in SPOT image.The resultant image shows this hybrid identification method performs better than only using knowledge-based method or neural network techniques

Key words: Road recognition; Satellite image; Fuzzy-C-Means; Neural network; Knowledge processing; Image processing