一种用于目标识别的图像融合算法

An Image Fusion Algorithm for Target Recognition

随着越来越多的装甲车辆装备了红外图像传感设备,使得将可增强战车侦察力的多传感器图像融合技术运用到火控系统的改进中成为一种趋势。针对这一情况,提出了一种基于IHS空间的红外与可见光图像的像素级融合算法。该算法基于IHS空间建立源图像灰度级与色调、饱和度和亮度等与人的感知颜色特征之间的函数关系,通过矩阵转换,得到融合图像的RGB值,可输出显示。实验结果显示:采用该融合算法得到的融合图像目标突出、色彩自然,具有较好的视觉效果,有助于伪装目标的识别。

With the equipping of more and more infrared image sensors,it can be done to fuse the multisensor images So a pixel-level image fusion algorithm for infrared and visible light image based on IHS space is proposed. Firstly, the individual functional relationship of hue, saturation and intensity with gray level is established. Then the corresponding R, G and B are obtained through the transformation from the perceptual space to RGB space, and are fed separately into the red,green and blue guns of the monitor and the result irmge is shown. It can be said by judging the fused image that the target is stand out and the color is natural. The algorithm can be used to improve the ability of target recognition of armored vehicles.