摘要
红外复杂背景抑制是红外告警等系统发现远距离弱小目标的难题之一。提出了一种将奇异值分解与对偶树复小波变换(DTCWT)相结合的多尺度截断复杂背景抑制新方法。首先采用DTCWT对图像进行正变换,获得图像的多尺度和方向细节特征;然后根据目标和背景杂波信号系数在不同尺度之间的差异,对各子带采用奇异值分解进行处理,并利用最大的特征值重构子带;最后将系数调整后的各子带逆变换到图像域,从而将弱小目标和背景杂波分离,达到抑制背景的目的。实验结果表明,该算法可以在很大程度上抑制结构化背景,保存并增强目标信号。
Complex background suppression of infrared dim and small target detection is a key problem for finding long-distance target in infrared warning system.A complex background suppression algorithm based on multiscale truncation,which combines the singular value decomposition with the dual tree complex wavelet transform（DTCWT）,is presented.Firstly,DTCWT is adopted to decompose the input infrared image,which extracts multi-scale detail features of images.Then,according to difference between targets and background clutter signals,the singular value decomposition is introduced to process sub-bands,and maximum eigenvalues are utilized to compose the sub-bands.Finally,the image is synthesized by the modified sub-bands,then targets and background details are separated,by which background suppression is realized.Experimental results validate that the presented method could suppress the structured background in some degree,and preserve and enhance the target signal.
出处
《光学学报》
EI
CAS
CSCD
北大核心
2010年第10期2812-2816,共5页
Acta Optica Sinica
基金
国家部委科技项目(7130721
41101050104)
教育部科学技术研究重点项目(108114)
国家自然科学基金(60902080)
中央高校基本科研业务费专项资金(72104810)资助课题
关键词
图像处理
背景抑制
对偶树复小波变换
奇异值分解
目标检测
image processing
background suppression
dual-tree complex wavelet transform
singular value decomposition
target detection
