基于视觉对比度分辨率的红外图像增强算法

针对传统红外图像增强方法存在增强后目标边缘模糊及背景噪声过增强的缺陷,结合人眼视觉特性,提出了基于视觉对比度分辨率的非线性变换算法。该算法根据人类视觉在不同背景灰度下分辨目标的能力不同,自适应调整灰度变换曲线,使目标映射到人眼分辨的敏感区域,同时使背景噪声映射到人眼分辨的不敏感区域。经测试表明:提出的算法与传统算法相比更易突出红外图像目标的细节信息及其边缘轮廓,峰值信噪比提高近1倍,对比度增益提高近0.5倍。

An Algorithm for Infrared Image Enhancement Based on Vision Contrast Resolution

Traditional methods for infrared image enhancement have some problems of fuzzy edge detail and excessive enhanced background noise. Combined with human visual properties, a novel algorithm that tailors the required amount of contrast enhancement based on human vision contrast resolution is proposed. According to the difference of human vision resolution in different background gray, the algorithm has self-adaptive characteristic, which makes the target mapped to the region of suiting the human eye to distinguish the background noise. The experimental results show that the proposed algorithm owns better performance in terms of highlighting infrared image target detail information and edge contour compared to traditional methods. The peak signal to noise ratio and contrast gain of the processed images increase by 100% and 50% respectivelty.