应用于车辆主动安全的轻量级图像匹配

Lightweight image applied to vehicle active safety match

基于立体视觉的路障探测是当前汽车主动安全技术的研究热点,其核心是图像匹配技术。本文引入二进制鲁棒独立基元特征(BRIEF),提出一种兼具高匹配精度和低运算复杂度的轻量级图像匹配方法,为利用车载低功耗处理器进行双目图像实时匹配提供了可能。

The obstacle detection based on stereo vision is a hot research topic of vehicle active safety, the core of which is image matching technology. In this paper, a low-weight image matching algorithm with high accuracy and low computational cost is proposed, utilizing the Binary Robust Independent Elementary Features（BRIEF）. This algorithm makes real-time matching on binocular images possible when being executed by a low-cost embedded processor with limited computational capabilities.