Computer Engineering and Applications ›› 2022, Vol. 58 ›› Issue (19): 64-75.DOI: 10.3778/j.issn.1002-8331.2203-0600

• Research Hotspots and Reviews • Previous Articles     Next Articles

Overview of Infrared and Visible Image Fusion Algorithms for Automotive Driving Assistance System

AN Xiaodong, LI Yali, WANG Fang   

  1. School of Aerospace Engineering, Zhengzhou University of Aeronautics, Zhengzhou 450046, China
  • Online:2022-10-01 Published:2022-10-01



  1. 郑州航空工业管理学院 航空宇航学院,郑州 450046

Abstract: Infrared and visible fusion image fusion is one of the core functions of the automobile advanced driving assistance system. It is able to better understand the external environmental objectives when the light conditions are poor, and it plays an important role in the recognition environment for the driverless vehicles and the intelligent vehicles. The neural network algorithms based on the deep learning have significant advantages in the extraction and classification of the image feature. Therefore, this paper summarizes the infrared and visible image fusion algorithms in the automotive field. Firstly, the technology requirements of the vehicle fusion image are analyzed. Then, the traditional algorithms and the latest developments about the infrared and visible image fusion are summarized. Then, the infrared and visible image fusion algorithm based on the deep learning algorithm are discussed. Finally, the development trends of the infrared and visible image fusion technology are discussed for the vehicle.

Key words: neural network, infrared image, visible image, image fusion, feature extraction, deep learning

摘要: 红外与可见光融合图像融合是汽车高级驾驶辅助系统核心功能之一,能够较好地理解光线条件较差时车辆外部环境目标,对无人驾驶车辆和智能车辆识别环境具有重要作用,其中基于深度学习的神经网络算法在图像特征提取和分类中优势显著。针对汽车领域红外与可见光图像融合算法进行综述,分析了现代车辆对图像融合技术的需求;总结了基于数学方法框架的红外与可见光图像融合算法和最新发展;概述了基于神经网络结构的红外与可见光图像融合算法;最后讨论了车载红外与可见光图像融合技术的发展趋势。

关键词: 神经网络, 红外图像, 可见光图像, 图像融合, 特征提取, 深度学习