超宽带SAR图像道路提取算法适应性研究

Study on Adaptability of Road Detector in UWB SAR Images

从道路特征提取的角度,对四种经典道路特征提取方法进行了分析和对比,探讨了它们在超宽带SAR图像道路提取技术中的适应性。超宽带SAR图像道路提取技术是在继承传统方法的基础上,结合在道路的灰度特征或其在自然界中特有的几何特征,形成合适的道路提取算法。在理论分析超宽带SAR图像特点和道路提取经典方法基础上,研究了四种道路提取方法在超宽带SAR图像中的适应性,开展了超宽带SAR图像的道路提取实验,系统总结了各方法针对超宽带SAR图像的优点和不足,形成了超宽带SAR图像道路提取技术的研究思路和道路边缘提取流程,相比经典方法性能得到提高。

To realize road feature selection, this paper analyzed and compared four classical techniques and investigated the adaptability in ultra-wide-band SAR（UWB SAR） images. As an in-depth study of line extraction from UWB SAR images, road extraction follows original ways of detecting edges. First we figured out that the key point is to build a right road extraction algorithm according to the gray scale contrast or geometric features of road. Then the chosen algorithms were applied to UWB SAR images to check if they can work well enough. At last, a flow chart and an analytical thought on road extraction methods for UWB SAR images were formed and better performance was achieved.