摘要
针对红外图像序列中运动目标的检测问题,提出了一种提高红外图像信噪比的改进双重Top-Hat形态学预处理算法,并通过自适应形态学阈值获得了二值化图像,最后利用序列帧间信息进行管道滤波去除二值化图像的噪声和伪目标,完成多个红外运动目标的自动检测。实验结果表明,提出的方法有效地保留了运动目标的灰度信息,便于后续的目标识别处理,而且算法在实时性和信噪比方面的综合效果优于其他几种常用滤波算法。
A novel method for improving the SNR of the infrared image and detecting moving targets based on morphological double Top-Hat operator was presented.Then the morphology adaptive threshold was used to obtain the binary image.Finally,the pipe filter method was proposed to identify the real moving targets.Experimental results not only show that this method is practical and easy to perform,but also confirm that the enhanced morphology algorithm obviously better than other traditional filter algorithm if considering the general effects of the processing time and the SNR of the image after processing.
出处
《红外与激光工程》
EI
CSCD
北大核心
2007年第z2期141-145,共5页
Infrared and Laser Engineering
基金
航天科技创新基金"复杂背景下红外探测与识别关键技术研究"