论文部分内容阅读
为了对遥感合成孔径雷达(SAR)图像中桥梁目标进行自动检测,在极化散射矩阵参数的基础上,提出了一种寻找河流地区对比参数的方法:将总功率、极化熵、相似性参数等参数进行加权组合,利用特征值分析方法求取最优加权系数,从而得到对河流特征敏感的参数组合表达式。进而利用该表达式对河流地区进行了分类实验,对分类结果应用图形形态学方法处理后形成水域模板。在此基础上将水域模板对敏感参数图像覆盖后,对河上桥梁目标进行了自动检测。采用PI-SAR型的雷达的全极化数据进行实验,获得了良好的检测效果,和只使用功率数据进行桥梁检测相比,原来检测不出来的桥梁目标都明显显现出来。实验结果表明,综合利用全极化SAR信息在遥感应用中可以更好的检测和识别目标。
In order to automatically detect the bridge targets in remote sensing synthetic aperture radar (SAR) images, a method to find the contrast parameters in river area is proposed based on the parameters of polarization scattering matrix. The total power, polarization entropy, similarity parameter And other parameters weighted combination, the use of eigenvalue analysis method to obtain the optimal weight coefficient, and thus get the characteristics of river characteristics of the combination of parameters. Then, the expression is used to classify the river area and apply the morphological method to the classification results to form a water template. Based on this, after the watershed template covered the sensitive parameter images, the bridge targets on the river were automatically detected. Experiments using fully polarimetric data from the PI-SAR radar have yielded good results. Compared with using only the power data to detect the bridge, the originally undetected bridge targets are all obvious. Experimental results show that the comprehensive utilization of fully-polarimetric SAR information can better detect and identify targets in remote sensing applications.