摘要
提出了一种新的基于信号奇异性分析的小目标检测方法,该方法将图像中的小目标看作是图像信号中的奇异点(突变点),利用小波变换模极大值点同信号奇异点之间的对应关系,结合信号奇异性与模极大值在多尺度上的变化规律,实现噪声抑制和信号增强,然后通过区域分析,检测图像小波变换模极大值,进而实现小目标检测。实验结果证明该方法是行之有效的。
A new method for detection of dim targets in noisy background was presented, and the detectionwas based on signal’s singularity analysis. Dim targets were considered as singularities in image, then weused the corresponding relationship between wavelet transform maxima and singularities, and combined withthe variation rule of singularities and maxima in multi-scales, to realize noise suppressing and signalenhancing, finally, we detect dim targets from background through area analysis and modulus maximadetection. Numeral experiments on real IR images show its effectiveness.
出处
《红外技术》
CSCD
北大核心
2005年第3期245-249,共5页
Infrared Technology
