一种基于引导滤波的多焦红外图像增强算法

A Fast Guided Filter-Based Multi Focus Infrared Image E nhancement Algorithm

红外成像已经广泛应用到军事和空间当中，但由于其对比度较低，往往需要增加图像增强算法以改善视觉效果。在视频检测、跟踪等系统中，为了同时获得不同距离场景的图像，需要多焦图像融合技术。然而，基于红外的视觉系统中，红外图像增强和多焦融合分别实现意味着需要更高的时间和空间成本。基于此，本文提出了一种基于引导滤波的多焦红外图像增强算法。该算法首先通过引导滤波算法将图像分为细节层和背景层，接着基于细节层图像建立局部清晰度评价函数并选择出每个图像中的聚焦像素，再将聚焦像素背景层和细节层按细节增强方法重新绑定。实验结果表明，提出的算法能实现多焦红外图像的快速融合增强，具有广泛的应用前景。

Infrared imaging system has been widely used in civil and space imaging applications. However, an image enhancement algorithm is often needed to improve the visual effects because of its low contrast. In the target detection and tracking system, multifocal image fusion technology is needed to obtain the images of different distance scenes simultaneously. However, infrared image enhancement and multifocal fusion are realized respectively means higher time and memory costs. Based on the reasons mentioned above, this paper proposed a multi focus infrared image enhancement algorithm based on guided filtering. Firstly, original image was divided into detail and background layers by guided filtering. And then an image clarity evaluation function was established based on the detail layers, which used to choose the focus pixels. Finally, a new fused enhanced image was obtained by binding the focused detail layer and background layer. Experimental results showed that the proposed algorithm is able to achieve fast fusion and enhancement of multi focal infrared images, and has a wide range of applications.