To address the challenges of difficult feature extraction and reduced positioning accuracy caused by image quality degradation for robot visual localization in indoor low-light environments

a multi-state constraint Kalman filter(MSCKF)visual inertial odometry(VIO)method that integrates image enhancement and gyroscope zero-bias correction is proposed.This method fuses data from an inertial measurement unit (IMU) and a binocular vision sensor

utilizes an image enhancement module to improve low-light image quality to assist feature matching

and compensates for gyroscope zero-bias through a correction mechanism

thereby enhancing the system’s robustness in low-light conditions. Experimental results on the EuRoC dataset demonstrate that compared with similar algorithms

the proposed method effectively increases the number of extracted feature points

avoids tracking loss

and significantly improves positioning accuracy under low-light conditions.It verifies the effectiveness and superiority of this method for indoor low-light positioning problems.