摘要
为了解决图像采集或视频监控应用中遇到夜视或背光等低照度的情况,文中从对数图像处理模型出发,结合色彩恒常性理论,提出一种基于人眼视觉特性的低照度图像增强算法。该算法首先将RGB彩色图像转换到HSV色彩空间,利用对数图像处理模型(LIP)算子改进普通空域滤波器对亮度通道进行反射分量的提取,以更好地保持图像高频信息,然后根据局部的背景亮度信息计算放大系数,在LIP模型下对每个像素的光照分量进行增强,以提高图像的亮度和对比度。实验结果表明,新算法基于人眼视觉特性,有效地提高了图像的整体对比度,保留图像的细节信息,有较好的主观视觉感受。
To solve the problems of low illumination for the images obtained by the situation of the poor light in image acquisition or video surveillance applications, a new image enhancement algorithm is proposed based on human visual properties by combining the color constancy theory with the logarithmic image processing model. Firstly, RGB color images are converted into HSV color space by the algorithm. Secondly ,the improved filter based logarithmic image processing（LIP） model is used to obtain the reflection image,keeping high frequency components better. Then,each pixel of scene illumination is enhanced according to the coefficient calculated by local background brightness, which can improve the brightness and contrast of image. The experimental results show that the new algorithm based on human visual properties can effectively improve the overall contrast of the image and keep the details with better subjective visual perception.
出处
《南京邮电大学学报（自然科学版）》
北大核心
2017年第2期98-106,共9页
Journal of Nanjing University of Posts and Telecommunications：Natural Science Edition
基金
国家自然科学基金(61471201
61501260)
江苏省自然科学基金(BK20130867)
江苏省高校自然科学重大项目(13KJA51004)
江苏省科协首席专家(工程师)基金
南京邮电大学自然科学基金(NY214039)资助项目
关键词
图像增强
对数图像处理模型
RETINEX理论
低照度图像
image enhancement
logarithmic image processing （LIP） model
Retinex theory
low illumination image
