摘要
合成孔径雷达(SAR)图像包含丰富的信息,对SAR图像边缘特征的提取对战场侦察、打击效果评估等具有重要的意义,但SAR严重的相干斑噪声,给SAR图像边缘检测增加了极大的难度。该文在分析常用图像边缘检测算子性能的基础上,考虑各微分检测算子检测特性的优缺点,从信息融合角度出发,提出了一种针对SAR图像保留边缘的融合方法,在计算量不显著增加的前提下,提高了边缘检测的质量,实验结果证明此方法既保留了图像的边缘特征,又一定程度地抑制了图像的噪声,效果较好。
SAR Image includes much useful information.The edge character abstraction of SAR image is very important for battlefield spy and attacking effect evaluation,but the serious speckle of SAR image adds great difficulty to edge detection of SAR image.In the basic of analyzing characteristics of usual edge detection operators,this paper achieves advantages and disadvantages of each differential coefficient detection operators.From the view of information fusion,this paper brings forward a fusion method used in SAR image edge detection.This method improves the quality of edge detection and reduces computing capacity and the result of experiment testifies that this method holds image's edge character and restrains the speckle of SAR image.
出处
《计算机工程与应用》
CSCD
北大核心
2003年第27期15-17,共3页
Computer Engineering and Applications
基金
国家自然科学基金资助项目(批准号:60272022
69931040)
