Affiliation:
1. School of Information Science and Technology North China University of Technology Beijing China
Abstract
Sensor calibration is an important step in SLAM, and lightweight, inexpensive sensors are often used in Augmented Reality applications. The accuracy of sensor calibration directly affects the accuracy and reliability of subsequent image processing, recognition, perception, and positioning. The inertial processing unit is robust to the sudden motion effect, blurring appears in the middle of camera motion, and the overlap area between two frames is too small, so it is difficult to match feature points. The camera itself can solve the problem of IMU drift in the middle of slow motion, and the two are complementary. VI‐SLAM, which integrates vision and IMU, has become a hot research topic at present. In this paper, the INDEMIND calibration method is analyzed and improved, the maximum likelihood method is used to improve the calibration accuracy, and its effectiveness is verified in the experiment.