摘要
为更好地保留原始红外图像的目标信息,挖掘更多微光图像的细节信息,提出一种结合分割红外图像与增强微光图像的融合方法。首先对原始红外图像进行二维最大熵法分割以提取红外目标信息,对原始微光图像进行Zadeh变换以增强其细节信息,依据红外分割图,将增强后的微光图像与原始红外图像进行一次融合;然后,在非下采样contourlet变换(NSCT)域,对原始红外图像、微光图像和一次融合后的图像进行二次融合,得到最终融合图像。通过试验对比,所提出的二次融合方法得到的融合图像视觉效果明显优于其它方法。最后,利用多个指标进行客观评价。结果显示,该方法可有效突出目标信息,提高图像清晰度与对比度,挖掘更多细节信息。
In order to preserve the original object information of the infrared image and mine more detail information from LLL image, this paper puts forward a fusion method which combines to segment original infrared image and to enhance original LLL images. Firstly, segmenting the infrared image with 2D maximum entropy method to extract the infrared target information, enchancing the LLL image by Zadeh transform for getting more image details, on the basis of the segmentation map to fuse the enhanced LLL image and original infrared image; then two original images and the first fused image are used to realize fusion in non-subsapled contourlet transform（NSCT） domain, we get the second fused image. By contrast of experiments, the fusion image of the second fused method＇s visual effect is better than other methods. Finally, a multi-index evaluation is used to evaulate the fused images＇ quality, its results show that the proposed method can high- light the target information, improve the image resolution and contrast and mine more details.
出处
《软件导刊》
2017年第1期170-173,共4页
Software Guide
基金
重庆市教委科学技术研究项目(KJ132104)
第二批重庆市高等学校青年骨干教师资助项目(2014)
重庆工业职业技术学院科研项目(GZY201517-YK)
