基于CFAR的高分PolSAR影像桥梁自动识别方法

发布时间：2018-08-05 20:05

【摘要】：桥梁的自动解译具有重要的应用价值,而在影像分辨率为分米级、桥梁场景复杂、桥梁目标较小的复杂情况下,准确地进行桥梁目标的自动识别比较困难。在分析高分辨率SAR(synthetic aperture radar)影像的统计特征和桥梁特征的基础上,提出了一种新的桥梁自动识别方法。首先采用基于Weibull分布的CFAR(constant false alarm rate)算法检测出潜在桥梁目标,然后基于Wishart-H-Alpha分类和形态学处理提取出桥梁场景区域,随后引入霍夫变换并利用桥梁的场景特征、几何特征和散射特征识别出桥梁目标。采用国产机载XSAR数据和美国AIRSAR数据进行验证,结果表明,该识别方法在复杂情况下能够取得令人满意的识别结果,具有较好的适应性。
[Abstract]:The automatic interpretation of the bridge has important application value, but it is difficult to identify the bridge target accurately when the image resolution is decimeter, the scene of the bridge is complex and the target of the bridge is small. Based on the analysis of the statistical features and bridge features of high-resolution SAR (synthetic aperture radar) images, a new automatic bridge recognition method is proposed. Firstly, the potential bridge targets are detected by CFAR (constant false alarm rate) algorithm based on Weibull distribution, then the bridge scene regions are extracted based on Wishart-H-Alpha classification and morphological processing, and then the Hough transform is introduced and the bridge scene features are used. Geometric features and scattering features identify bridge targets. The domestic airborne XSAR data and the American AIRSAR data are used to verify the proposed method. The results show that the method can obtain satisfactory recognition results under complex conditions and has good adaptability.
【作者单位】：武汉大学测绘遥感信息工程国家重点实验室;武汉大学遥感信息工程学院;首都师范大学资源环境与旅游学院;
【基金】：测绘公益项目(201412002) 国家自然科学基金(91438203,61371199) 中国海事局烟台溢油应急技术中心项目城市空间信息工程北京市重点实验室项目(2014204) 地理空间信息工程国家测绘地理信息局重点实验室项目(201406)~~
【分类号】：TP751;U446

【相似文献】