节点文献

基于机器视觉的汽车前照近光灯检测算法实现

Realization of an Algorithm for Vehicle Low-beam Headlamp Detection Based on Computer Vision

  • 推荐 CAJ下载
  • PDF下载
  • 不支持迅雷等下载工具,请取消加速工具后下载。

【作者】 胥钧曾德斌林大超王文鑫

【Author】 XU Jun;Zeng Debin;LIN Dachao;WANG Wenxin;North China Institute of Science and Technology;Mirong (Beijing) Transportation Technology Co, Ltd;

【机构】 华北科技学院弥荣(北京)交通科技有限公司

【摘要】 近光灯明暗截止线及拐点检测是汽车前照灯检测中极为重要的技术项目。本文依据热点、水平明暗截止线及倾斜明暗截止线三者之间的相对位置关系,提出了一种基于热点的自适应感兴趣区域定位算法,并对其进行了较为详细的阐述。最后,通过使用标准光源,从视觉和检测偏转角两个方面对算法进行验证。验证结果表明,本文的算法具有较好的鲁棒性、稳定性及实用性。

【Abstract】 Detection of the cut-off line and corner of low-beam is an important item in the examination of vehicle headlamps. This paper proposes an adaptive localization algorithm of the region of interest that is realized on the basis of the relationship among the hotspot, the horizontal cut-off line and the inclined cut-off line, and presents an analysis and discussion of the algorithm in detail. The algorithm is then examined from two aspects of the vision and deflection angle measurements, by using the standard low-beam source. The result shows that the proposed algorithm has good robustness, stability and possibility for practice.

【基金】 中央高校基本科研业务费(3142020046)
  • 【文献出处】 照明工程学报 ,China Illuminating Engineering Journal , 编辑部邮箱 ,2021年02期
  • 【分类号】TP391.41;U463.65
  • 【被引频次】2
  • 【下载频次】89
节点文献中: 

本文链接的文献网络图示:

本文的引文网络