首页|基于显著性特征检测的低光照图像增强算法

基于显著性特征检测的低光照图像增强算法

扫码查看
计算机视觉技术在公共安全、智能交通和工业生产等领域有着广泛的应用,如人群分析、密度估计以及目标跟踪、识别、分割等.但是实际成像环境复杂,受雨、雾和低光照等因素干扰,室外环境下拍摄的图像往往存在色彩失真、细节缺失、成像质量差等问题,严重影响了后续视觉任务.为了降低光照和雨雾天气的影响,提高成像质量,改善视觉效果,提出了 一种基于显著性特征检测的图像增强方法.首先,针对图像颜色失真问题,提出了一种基于多通道融合和显著性亮度调节的颜色恢复方法.其次,为了增强图像细节,采用了基于显著性特征保留的方法实现图像细节增强.实验结果表明,该方法在客观评价指标和主观视觉效果方面均优于算法.
Low light image enhancement algorithm based on significance feature detection
Computer vision technology has a wide range of applications in public safety,intelligent transportation,and industrial production,such as crowd analysis,density estimation,and object tracking,recognition,and seg-mentation.However,the actual imaging environment is complex,and due to factors such as rain,fog and low light,the images taken in the outdoor environment often have problems such as color distortion,lack of detail,and poor imaging quality,which seriously affect the subsequent visual tasks.In order to reduce the influence of lighting and rain and fog,improve the imaging quality and improve the visual effect,an image enhancement method based on saliency feature detection is proposed.Firstly,aiming at the problem of image color distortion,a color recovery method based on multi-channel fusion and significance brightness adjustment is proposed.Secondly,in order to enhance the image details,the method based on saliency feature retention is adopted to achieve image detail en-hancement.Experimental results show that the proposed method is superior to the recent algorithm in terms of ob-jective evaluation index and subjective visual effect.

low-light imagesimage enhancementsaliency feature detectioncolor recoveryDetail enhance-ment

王腾龙、古玉立、阚健斐、张斯斯

展开 >

中国移动通信集团云南有限公司呈贡分公司,云南昆明 650500

低光照图像 图像增强 显著性特征检测 颜色恢复 细节增强

2024

云南民族大学学报(自然科学版)
云南民族大学

云南民族大学学报(自然科学版)

CSTPCD
影响因子:0.381
ISSN:1672-8513
年,卷(期):2024.33(4)