ScanSAR图像舰船目标快速检测方法  

Fast ship detection method based on ScanSAR image

作　　者：林旭[1] 洪峻[2] 孙显[1] 鄢懿[3] 

机构地区：[1]中国科学院电子学研究所空间信息处理与应用系统技术重点实验室,北京100190 [2]中国科学院电子学研究所微波成像技术国家重点实验室,北京100190 [3]中国科学院大学电子电气与通信工程学院,北京100049

出　　处：《中国科学院大学学报（中英文）》2013年第6期793-799,共7页Journal of University of Chinese Academy of Sciences

基　　金：国家自然科学基金(41001285)资助

摘　　要：提出一种ScanSAR图像舰船目标快速检测方法.该方法通过周期性局部亮度变换技术降低了Scalloping效应对舰船检测的影响,在ScanSAR图像拼接前进行基于OpenMP的改进的双参数CFAR舰船目标检测,不仅避免了拼缝对舰船检测造成的巨大影响,而且极大地缩短了舰船检测的时间,提高了舰船检测的时效性,最后通过图像后处理从拼接后的ScanSAR图像中得到最终的检测结果.对15幅ScanSAR图像进行了实验,结果表明,该方法在提高检测性能的同时,检测的时效性也有明显改善.In this paper, a fast ship detection method based on ScanSAR image is presented. Firstly, the influence of scalloping noise on ship detection is reduced by seasonal local brightness transform. Secondly, an improved two-parameter CFAR ship detection based on OpenMP before ScanSAR images mosaicking is adopted, and it not only avoids the influence of splicing apertures on ship detection but also reduces the time of ship detection greatly. At last, the final detection targets are derived from mosaicking ScanSAR images. Experimental results on 15 different SeanSAR images show that the proposed system has good detection performance and prominent timeliness.

关 键 词：ScanSAR图像 Scalloping 拼缝 OPENMP 舰船快速检测 

分 类 号：TN957.52[电子电信—信号与信息处理]