A Multi-exposure Fusion Method to Improve the Dynamic Range of Visible Light Images
蔡鸿志①② CAI Hong-zhi；冯书兴① FENG Shu-xing；张梁② ZHANG Liang
（①航天工程大学，北京 100000；②63811部队，文昌 571300）
（①Aerospace Engineering University，Beijing 100000，China；②Unit 63811，Wenchang 571300，China）
摘要：靶场光学测量设备需要对火箭进行实况景象记录，可见光相机分辨率高且细节信息丰富，但其动态范围有限，难以满足靶场高动态范围成像要求。针对靶场可见光图像高动态范围成像问题，本文提出一种基于卷积神经网络的方法。该方法为解决传统图像融合方法缺陷提供了思路，并在靶场首区任务中进行了实验。实验表明，本文提出的方法可立足可见光图像本身，采用图像处理的方法获取更高动态范围的成像，具有一定的参考价值。
Abstract: The optical measurement equipment of the shooting range needs to record the live scene of the rocket. The visible light camera has high resolution and rich detailed information, but its dynamic range is limited, and it is difficult to meet the high dynamic range imaging requirements of the shooting range. Aiming at the problem of high dynamic range imaging of visible light images in the shooting range, this paper proposes a method based on convolutional neural network. This method provides ideas for solving the shortcomings of traditional image fusion methods, and has been tested in the first area of the shooting range. Experiments show that the method proposed in this paper can be based on the visible light image itself, and the image processing method can be used to obtain imaging with a higher dynamic range, which has a certain reference value.
关键词：动态范围；图像融合；卷积神经网络
Key words: dynamic range；image fusion；convolutional neural network
中图分类号：TP391.4 文献标识码：A 文章编号：1006-4311（2022）23-083-03
DOI:10.3969/j.issn.1006-4311.2022.23.027.
文章出处：蔡鸿志,冯书兴,张梁. 一种提升可见光图像动态范围的多曝光融合方法[J]. 价值工程,2022,41(23):83-85.
|
