【摘要】 详细分析了SAR图像道路自动提取中的各种不确定性因素,并给出相应参数选取准则与提取性能评价。首先根据SAR图像中道路的典型特征建立起局部结构模型,然后针对SAR影像的统计特性,对模型中路体区域与两侧区域的均值比变量函数进行概率分析,由变量函数的参数不确定性分析得出合理的参数选择准则,再依照准则选取合适参数,据K-S假设检验理论对图像中道路点进行判定提取。经过不同参数提取方案的实验及结果评价验证了不确定性理论。参数选取准则可用于指导SAR图像道路自动提取,以更好服务于匹配制导,制图等实际应用。更多还原

【Abstract】 The uncertain factors of automatic extracting roads from the Synthetic Aperture Radar (SAR) imagery were analyzed in detail; the parameters selection rules and performance evaluation were drawn following the analysis. The procedure was, at first, through analyzing the road typical features in the SAR image, the local model of road structure was built. Then the statistic characteristic of SAR image was utilized, and the ratio probability density function (pdf) of the adjacent region to the road region was obtained. Followed by analyzing the uncertainty of the parameters in the pdf, the adaptable rules on parameters setting were concluded. Subsequently, after selecting proper parameters by the rules drawn above, by the K-S hypothesis testing theory one could get the roads points in the image. The experiments through selected various parameters and quality indexes to evaluate the results have shown the consistency with theoretical uncertain analysis. The parameters selection rules can guide the roads extraction in SAR images in the applications such as matching guidance and mapping, etc.更多还原