中国激光, 2014, 41 (s1): s109009, 网络出版: 2014-07-03  

一种基于多次曝光的大动态范围图像融合方法

A High Dynamic Range Fusion Method Based on Multi-Exposure Imaging
作者单位
1 中国科学院上海技术物理研究所, 上海 200083
2 中国科学院上海技术物理研究所航天遥感信息获取重点实验室, 上海 200083
摘要
多次曝光图像融合的目的是结合不同曝光的各个图像的互补信息,组成一个新的图像以获得场景更为全面、准确、可靠的图像描述。目前常用的融合方法有辐照度重建法,金字塔变换法,小波变换法,HIS假彩色法等。这些方法有效地增强了可探测的目标亮度范围,但大都将大动态范围的场景信息映射到8 bit中,虽然增加了目标场景欠曝和过曝区域的亮度信息,但是牺牲了中间亮度的对比度。提出了一种多次曝光融合的方法,采用指数积分时间对固定场景进行4次采样,将获取的低动态范围图像序列乘以一个积分时间因子直接进行图像非饱和区域的匹配。然后,以图像各点的对比度为权重因子,对各幅图像高低端进行加权融合,最终获得一幅细节丰富,层次鲜明的11 bit图像。
Abstract
Multi-exposure image fusion aims to integrate complementary information of each exposure image to form a new image with the scene information more comprehensively, accurately and reliably. The main methods in image fusion are irradiance reconstruction method, pyramidal transform wavelet transform method, HIS feigned multicolor method and so on. These methods effectively enhance the detectable target brightness range, but among most of them, the large dynamic range of the scene information is mapped to 8 bit, which indeed increases the luminance information of the underexposed and overexposed scene areas, but at the expense of the contrast of the intermediate luminance. A multi-exposure fusion method is presented, using 4 exponential integral times in a fixed scene for image capture. The unsaturated zone is matched through multiplying the low dynamic range image sequence by a integration time factor images directly. All the images are fused with the contrast being the weighting factor, ultimately, a 11 bit image with rich details and distinct levels is captured.

黄思婕, 陈凡胜, 廖星星. 一种基于多次曝光的大动态范围图像融合方法[J]. 中国激光, 2014, 41(s1): s109009. Huang Sijie, Chen Fansheng, Liao Xingxing. A High Dynamic Range Fusion Method Based on Multi-Exposure Imaging[J]. Chinese Journal of Lasers, 2014, 41(s1): s109009.

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!