<p>Visual-Inertial Navigation Systems (VINS) are advancing rapidly, with increasing demands for robust performance under complex lighting conditions. Automatic camera exposure often causes significant brightness variations, which degrade the accuracy of visual odometry. This paper proposes an adaptive gamma correction algorithm to mitigate brightness changes introduced by automatic exposure. By incorporating it into the VINS pipeline, we significantly reduce brightness discrepancies between consecutive frames and stereo images. Additionally, a local feature region strategy restricts feature extraction to edge-detected areas, enhancing matching accuracy and robustness. To validate the practical effectiveness of our method, we integrate the enhanced VINS as an external odometry source for RTAB-Map, forming a tightly coupled multi-sensor SLAM system with improved front-end and back-end performance.This paper proposes an adaptive gamma correction algorithm to mitigate brightness changes introduced by automatic exposure. By incorporating it into the VINS pipeline, we significantly reduce brightness discrepancies between consecutive frames and stereo images. Additionally, a local feature region strategy restricts feature extraction to edge- detected areas, enhancing matching accuracy and robustness. Experiments on the public EuRoC dataset and real-world indoor scenes show that our method improves localization accuracy by approximately 26% compared to classical approaches. Overall, the proposed method enhances VINS robustness under challenging lighting conditions.</p>