Bridging the Illumination Gap: An Illumination-Robust Feature Extractor Enhanced by Relightable 3D Reconstruction

Visual perception serves as a crucial method for robots and mechatronic systems to sense their environment. Visual features, which encapsulate textural and semantic information from images, have consequently become essential components in visual mechatronic applications such as navigation and localization. However, the extraction of visual features is usually disturbed by the variation of illumination conditions, making it challenging for real-world applications. Previous works have addressed this issue by establishing datasets with variations in illumination conditions, but can be costly and time-consuming. This paper proposes a design procedure for an illumination-robust feature extractor, where the recently developed relightable 3D reconstruction techniques are adopted for rapid and direct data generation with varying illumination conditions. A self-supervised framework is proposed for extracting features with advantages in repeatability for key points and similarity for descriptors across good and bad illumination conditions. Experiments are conducted to demonstrate the effectiveness of the proposed method for robust feature extraction. Ablation studies also indicate the effectiveness of the self-supervised framework design.