Adaptive Robust Low-Beam Feature Point Detection Algorithm Based on Convex Hull

The existing algorithms for detecting low-beam feature points are neither universally applicable nor effective. Additionally, such algorithms typically involve numerous parameters, making them difficult to tune and lacking in robustness. In response to these challenges, this paper proposes a novel algorithm for low-beam feature point detection that incorporates convex hull detection capabilities, enabling the automatic localization of the detection area. This algorithm achieves parameter adaptation through the analysis of the inherent characteristics of low-beam light distributions. It is designed to be applicable to the detection of feature points in “two-segment,” “three-segment,” and “five-segment” low-beam light patterns. Furthermore, two application modes have been developed: one for elbow point detection and another for the simultaneous detection of multiple feature points. The experimental results demonstrate that the proposed algorithm can accurately detect and differentiate all types of low-beam elbow points and other feature points and is applicable to both right-hand and left-hand traffic low-beam light patterns. It offers advantages such as high detection accuracy, ease of debugging, strong versatility, real-time performance, and high robustness, making it well suited for applications in automotive lamp optical inspection and alignment adjustment.