薄雾环境下基于场景结构的图像匹配算法
Image Matching Algorithm Based on Scene Structure in Hazy Environments
蔡浩龙 1鲁鹏1
作者信息
- 1. 北京邮电大学人工智能学院,北京 100876
- 折叠
摘要
随着自动驾驶、无人机视觉导航等智能化应用的快速发展,图像匹配技术已成为计算机视觉领域的关键基础。然而,在薄雾环境下,图像质量严重下降,导致局部特征模糊、对比度衰减、颜色失真,传统图像匹配方法的性能急剧恶化。尽管宏观场景结构在薄雾中仍相对稳定,但现有方法难以有效挖掘和利用此类结构先验,限制了其在复杂环境下的应用。本文针对薄雾环境下基于场景结构的图像匹配算法展开研究,构建了基于离散化语义编码的场景结构先验学习框架,设计了结构先验驱动的薄雾图像增强算法和基于场景结构特征的图像匹配算法。实验结果表明,该算法在薄雾环境下的匹配准确率可达到 84.98\%,为智能视觉系统在复杂气象条件下的可靠运行提供了技术支撑。
Abstract
With the rapid development of intelligent applications such as autonomous driving and UAV visual navigation, image matching technology has become a critical foundation in the field of computer vision. However, in hazy environments, image quality degrades severely, resulting in blurred local features, reduced contrast, and color distortion, which drastically deteriorates the performance of traditional image matching methods. Although the macroscopic scene structure remains relatively stable under haze, existing methods fail to effectively exploit and utilize such structural priors, limiting their applications in complex environments. This paper investigates scene structure-based image matching algorithms for hazy conditions, constructs a scene structural prior learning framework based on discretized semantic encoding, and designs a structural prior-driven hazy image enhancement algorithm as well as a scene structure feature-based image matching algorithm. Experimental results demonstrate that the proposed algorithm achieves a matching accuracy of 84.98\% in hazy environments, providing technical support for the reliable operation of intelligent vision systems under complex meteorological conditions.关键词
人工智能/图像匹配/深度学习/特征提取Key words
Artificial Intelligence/Image matching/Deep learning/Feature extraction引用本文复制引用
蔡浩龙,鲁鹏.薄雾环境下基于场景结构的图像匹配算法[EB/OL].(2026-03-30)[2026-03-31].http://www.paper.edu.cn/releasepaper/content/202603-289.学科分类
计算技术、计算机技术
评论