摘要
在分析几种常用背景抑制方法的基础上,利用二代曲线波变换具有的优良特性,提出一种基于二代曲线波变换的红外弱小目标背景抑制方法。首先利用曲线波变换对图像进行分解提取图像的多尺度细节特征;然后对分解后的低、高频子带分别采用变分和模糊非线性变换进行处理来调整目标特征的强度,重构子带获得预测的背景图像;最终将其与原图相减得到背景抑制后的图像。实验结果表明,与几种经典方法相比,该方法在主观视觉和客观评价指标两方面均表现出良好的效果。
By utilizing the excellent characteristics of the second generation curvelet transform, a method of background suppression is presented for infrared dim target detection based on the second generation curvelet transform after discussing several classical methods. Firstly, the image is decomposed by the curvelet transform to extract the multi-scale detail characteristics. Then, the variational principle and the fuzzy nonlinear transform are used in the low frequency and high frequency components to change the＇intensity of the target; and the predicted background image is obtained by curvelet reconstruction. Finally, the original image is subtracted by the predicted background image and the result image is obtained. Experimental results demonstrate that the new method has better effects both in the subjective vision and objective evaluation compared with the several classical methods.
出处
《南京航空航天大学学报》
EI
CAS
CSCD
北大核心
2009年第6期757-761,共5页
Journal of Nanjing University of Aeronautics & Astronautics
基金
教育部科学技术研究重点(108114)资助项目
关键词
红外图像
背景抑制
曲线波变换
变分原理
模糊逻辑
infrared image
background suppression
curvelet transform
variational principle
fuzzy logic
