机载SAR图像目标识别中若干问题的研究

发布时间：2018-06-13 05:33

本文选题：SAR图像 + 去噪　；参考：《合肥工业大学》2014年硕士论文

【摘要】：合成孔径雷达(SAR, Synthetic Aperture Rader)是一种高分辨率成像雷达,具有全天时、全天候的侦查特点,在民用和军事领域都有着广阔的发展前景和实用价值,SAR图像处理技术也成为国内外的研究热点。因此,对SAR图像处理技术进行研究具有重要的理论和实际意义。本文主要对SAR图像目标识别系统中几个重要问题进行了研究,主要工作如下：(1)研究了SAR图像的相关去噪算法,并提出了一种增强型Shearlet域SAR图像去噪算法,实验结果验证了该算法在去噪的同时能够很好的保持图像的边缘细节信息。(2)研究了SAR图像分割算法,运用改进型PCNN网络模型实现高分辨率SAR图像分割,实验表明,该方法取得了有效的分割,并且在运算时间少的同时取得了很好地分割结果。(3)分析了SAR图像上一般目标的特征并分析了道路的特征模型,提出了一种SAR图像道路目标识别的方法,实验结果表明,本文方法能够有效的检测出道路边缘,实现道路目标的识别。
[Abstract]:Synthetic Aperture radar (SAR) is a kind of high-resolution imaging radar, which has the characteristics of all-day, all-weather investigation. The SAR image processing technology has become a hot research field at home and abroad, which has a broad development prospect and practical value in both civil and military fields. Therefore, the research of SAR image processing technology has important theoretical and practical significance. In this paper, several important problems in SAR image target recognition system are studied. The main work is as follows: 1) the correlation denoising algorithm of SAR image is studied, and an enhanced Shearlet domain SAR image denoising algorithm is proposed. The experimental results show that the proposed algorithm can keep the edge detail information of the image well while de-noising. (2) the SAR image segmentation algorithm is studied, and the improved PCNN network model is used to realize the high resolution SAR image segmentation. The experimental results show that the proposed algorithm can be used in SAR image segmentation. This method has achieved effective segmentation, and achieved a good segmentation result at the same time of less computation time.) the features of general targets in SAR images are analyzed and the road feature model is analyzed. A road target recognition method based on SAR images is proposed. The experimental results show that the proposed method can effectively detect road edges and realize road target recognition.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN957.52

【参考文献】