中文摘要：

从高分辨率遥感影像中提取道路有着非常重要的意义,但是受到遥感影像噪声、复杂的自然场景和已有算法的局限性的影响,道路提取有待于进一步研究.近些年来水平集方法被用于提取道路,但是初始水平演化曲线的确定却是一个大的难点.笔者提出一种自动的水平集分割方法,并将其用于道路检测中.首先,将卷积神经网络用于道路的粗分类.然后,利用形状特征和孔洞填充方法得到比较准确的道路区域.在此基础上,利用张量投票来提取道路的交叉口,并将其轮廓作为水平集演化的初始曲线进行水平集分割.最后,结合卷积神经网络分类和水平集分割的优势,得到比较完整的道路区域,并保持了道路的边缘.实验结果表明,该方法能自动地提取准确完整的道路区域.

英文摘要：

Road extraction from high-resolution satellite images is very important. Due to image noise, the natural scene complexity, and the extraction algorithms limitations, it still needs to be further researched. In recent years, level set evolution has been used to extract the road, but it is difficult to automatically generate initial level curves for the level set evolution （LSE）. In this paper, we propose an automatic approach to the generation of initial level curves and use it to extract the road. Firstly, the convolutionat neural network（CNN） is used to classify the road or nonroad, then shape features are adopted to filter nonlinear features to get the accurate road region. And on this basis, we exploit tensor voting to detect the road junctions and utilize them as initial level curves; finally we fuse the results obtained by the CNN and LSE. Experiments show that this algorithm can get an accurate and complete road.