基于数学形态金字塔双色融合的红外目标检测

为了提高红外目标检测性能,在分析了数学形态学基本运算的基础上,提出了基于数学形态金字塔的融合检测方法。首先把中波和长波红外图像分解成形态金字塔,建立融合准则,在相应的级上提取特征进行融合,然后重建原图像。在得到的融合图像中,适当地选取阚值,可以有效地把目标检测出来。试验结果表明,该算法优于单波段的目标检测,有较强的实用性。

Infrared Target Detection by Using Mathematic and Morphologic Pyramid Dualband Fusion

In order to improve the performance of target detection,a novel approach is proposed to de- tect targets by using mathematic and morphologic pyramid fusion dualband infrared images on the ba- sis of analyzing mathematic and morphologic operations. Firstly,mid-wave and long-wave infrared im- ages are decomposed into morphological pyramids respectively, and features are extracted in corre- sponding level. Then fusion rules and merge images are set up at different levels for reconstructing o- riginal image.In fused image,target can be detected by threshold. The experimental results show that the method is superior to that of single-wave target detection.