针对车辆环视系统的光照补偿方法
Illumination compensation method for vehicle around view monitor system
赵成诚 1苗作华 2朱良建 1刘代文 1陈澳光1
作者信息
- 1. 武汉科技大学资源与环境工程学院,武汉 430081
- 2. 武汉科技大学资源与环境工程学院,武汉 430081;冶金矿产资源高效利用与造块湖北省重点实验室,武汉 430081
- 折叠
摘要
针对车辆环视系统车载鱼眼相机,因光照影响导致画面色彩异常和边缘亮度下降的问题,设计了一种针对车辆环视系统的光照补偿方法.首先在4路鱼眼图像色彩校正时,提出了基于灰世界模型多图联合色彩校正算法,实现了快速降低不同方位图像因光照产生的色彩差异;然后利用相机内外参进行仿射变换,实现畸变校正和鸟瞰图生成;最后在鸟瞰图重叠区域引入光照调节映射表进行边缘亮度调节,经图像融合得到全景2D鸟瞰图.采集多种光线场景下4路鱼眼相机摄取的实际行车视频进行试验,该文方法处理后图像自然图像评价指标降低了 40％,单帧运行时间仅27 ms,实验结果表明该方法可以在不同光照场景下稳定输出光照均匀的全景2D鸟瞰图.
Abstract
In order to solve the problem of poor visibility of the Vehicle Around View Monitor system affected by ambient light,this paper designs a joint color correction algorithm for multiple original images and a.transparency fu-sion algorithm for adjacent images.Firstly,in the color correction of 4-way fish-eye images,a multi-image joint color correction algorithm is proposed based on the gray world model,which can quickly reduce the color difference caused by illumination of different cameras;Then the affine transformation is realized by using the external matrix of the cam-era to generate the aerial view;Finally,the illumination adjustment mapping table is introduced into the overlapping area of the aerial view to adjust the edge brightness,and the panoramic 2D aerial view is obtained through image fu-sion.This paper collects the actual driving video taken by four fish-eye cameras under various light scenes for testing.After processing,the evaluation index of the NQIE of the image in this paper is reduced by 64％,and the running time of a single frame is only 27ms.The results show that the method can stably output panoramic 2D aerial view with uni-form illumination under different lighting scenes.
关键词
车辆环视系统/光照补偿/色彩校正/图像融合Key words
around view monitor system/light compensation/color correction/image fusion引用本文复制引用
基金项目
国家自然科学基金(41071242)
国家自然科学基金(41971237)
教育部产学合作协同育人项目(202102136008)
出版年
2024
NETL
NSTL
万方数据