An Intelligent Image Processing-based Approach to Optimize Vehicle Headlamp Aiming

摘要

In this paper, a novel image processing-based approach is proposed to optimize vehicle headlamps aiming. Currently, most aiming devices rely on numerical derivative oriented methods to find the essential focal features, also called reference points, to perform the aiming process. However, these approaches are not robust, and minor changes in isocurves' smoothness may result in finding inaccurate aiming focal features. To address the associated robustness issue, a statistical signal processing-based approach named penalized contrast for changepoint detection is proposed. Experimental results indicate a high accuracy level of the proposed method concerning the master-tuned ground truth case studies while suggesting a robust mathematical process. Furthermore, it provides an extra level of flexibility for defining alternative focal features by calculating several points along isocurves, each of which can be considered as an axillary reference point for enhancing the aiming process.