基于人眼视觉特性的低照度图像增强算法研究

发布时间：2018-04-10 05:40

本文选题：图像增强　切入点：低照度图像　出处：《南京邮电大学》2017年硕士论文

【摘要】：在现实生活中,由系统采集设备所获取的图像和视频,在周围环境光照不足的情况下容易出现对比度下降、细节丢失、色彩失真等问题,这将严重影响到图像的后续处理与应用,因此有效地对低照度图像进行增强具有重要的意义。本文分析了不足光照环境下图像降质的原因及特性,研究常用的图像增强相关算法,并根据实际情况对现有的算法进行改进和完善。本文的具体研究内容如下:首先,本文针对采集的单幅低照度图像提出了一种基于对数图像处理模型的图像增强算法。为了提高反射分量的提取效果,本文提出利用LIP模型改进低通高斯滤波器来估计光照分量,更好地保持图像的细节特征;为了进一步提高图像的对比度,防止局部过度增强现象,本文结合韦伯-费希纳定律提出了基于局部背景亮度信息的光照分量增强算法。实验结果表明,与传统的增强算法相比,本文所提出的基于人眼视觉特性的增强算法,有效地提高了图像的整体亮度和对比度,保留了图像的细节信息,有较好的主观视觉感受。其次,本文针对采集的低照度视频提出了一种基于图像融合技术的增强方法。针对低照度视频图像融合增强过程中容易产生图像混淆、色彩漂移和前景运动区域增强效果不佳等问题,提出了一种改进的基于高质量视频帧信息补偿的融合增强算法。该算法首先对低照度视频中前景运动目标和静态背景进行分离,对背景区域采用改进的基于权重比例的融合策略来提高其质量,对前景物采用限制对比度的直方图均衡化对目标进行增强,最后将背景与运动目标融合得到最后的增强视频帧。实验结果表明,本文提出的改进融合策略的增强算法与传统算法相比更好地凸显图像细节信息,亮暗区域的对比度都得到较大的改善,且不会出现混淆和色彩漂移等现象。最后,本文构建了一种基于人眼视觉特性的图像增强质量评价方法。目前图像增强缺乏有效统一的质量评价方法,本文从人眼的主观视觉特性出发,提出了一种有效的低照度图像增强效果评价方法,并建立相应的评价函数。实验表明,该方法能够对不足光照环境下降质图像的增强效果给出较为准确的评价。
[Abstract]:In real life, the images and videos captured by the system acquisition equipment are prone to the problems of contrast decline, detail loss, color distortion and so on, when the surrounding environment is not fully illuminated.This will seriously affect the subsequent processing and application of the image, so it is of great significance to enhance the low illumination image effectively.In this paper, the causes and characteristics of image degradation under insufficient illumination are analyzed, and the commonly used image enhancement correlation algorithms are studied, and the existing algorithms are improved and improved according to the actual situation.The main contents of this paper are as follows: firstly, an image enhancement algorithm based on logarithmic image processing model is proposed.In order to improve the extraction effect of reflection component, this paper proposes to use LIP model to improve the low-pass Gao Si filter to estimate the illumination component, to better preserve the detailed features of the image, and to further improve the contrast of the image.In order to prevent the phenomenon of local over-enhancement, this paper presents an algorithm of illumination component enhancement based on local background luminance information in combination with Weber-Fischner 's law.The experimental results show that compared with the traditional enhancement algorithm, the proposed enhancement algorithm based on human visual characteristics can effectively improve the overall brightness and contrast of the image, and retain the details of the image.Have good subjective visual feeling.Secondly, this paper proposes an enhancement method based on image fusion for captured low illumination video.Aiming at the problems of image confusion, color drift and poor performance of foreground motion region enhancement in the process of low illumination video image fusion enhancement, an improved fusion enhancement algorithm based on high quality video frame information compensation is proposed.The algorithm firstly separates the foreground moving target from the static background in the low illumination video, and improves the quality of the background region by adopting an improved fusion strategy based on the weight ratio.The object is enhanced by histogram equalization with restricted contrast. Finally, the final enhanced video frame is obtained by fusion of background and moving object.The experimental results show that the improved fusion algorithm presented in this paper can better highlight the details of the image compared with the traditional algorithm, and the contrast of the bright and dark areas will be improved greatly, and there will be no confusion and color drift.Finally, an image enhancement quality evaluation method based on human visual characteristics is proposed.At present, there is a lack of effective and uniform quality evaluation method for image enhancement. Based on the subjective visual characteristics of human eyes, this paper presents an effective method for evaluating the effect of low illumination image enhancement, and establishes the corresponding evaluation function.The experimental results show that the method can accurately evaluate the enhancement effect of degraded images in insufficient light environment.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41

【参考文献】