摘要
高分辨率POLSAR图像的机场感兴趣区域(Region of interest,ROI)的自动提取是自动目标识别(AutomaticTarget Recognition,ATR)系统的任务之一,也是准确识别分类飞机等小目标的基础。针对全极化合成孔径雷达(POLSAR)图像极化相干的特点,提出一种融合提取方法:先使用J.S.Lee Sigma filter滤波,再利用Shannon-Entropy理论提高ROI和背景对比度,采用基于CV模型的方法分割图像,然后对分割得到的图像进行形态学等图像处理,最终得到机场ROI。实验结果表明,该方法具有分割界限清晰、定位准确的优点。
Automatic detection of the airport as the region of interest（ROI） from the high resolution POLSAR image is one of tasks of Automatic Target Recognition （ATR） systems as well as the foundation of airplane detection. Considering the coherent polarimetric property of the POLSAR image,a new combined extraction algorithm is proposed. The image enhancement method based on Shannon- Entropy theory is firstly applied on the T3 matrix filtered by J. S. Lee sigma filter to improve the contrast between the ROI and the back- ground, and then the CV model is used to track the contour of the ROI ,finally the airport ROI is obtained after the morphologie processing. Experiment with actual data indicates the algorithm is effective in segmenting the image and obtaining the proper airport ROI.
出处
《无线电工程》
2012年第7期10-12,28,共4页
Radio Engineering
基金
中央高校基本科研业务费专项奖金资助项目:地理信息科学教育部重点实验室开放研究基金资助项目(KLGIS2011A12)
