Vision-IMU Based Obstacle Detection Method

摘要

Obstacles' accurate classification is the first step in traditional obstacle detection methods, and the step causes the problem of high time and space complexity. In this paper, an obstacle detection method based on the principle of pinhole imaging is proposed to solve the problem. The monocular camera and inertial measurement unit are used as the basic sensing units in proposed method. The obstacle detection steps and indoor experiments are shown to expound the detection process of the Vision-IMU based obstacle detection method. The Vision-IMU based obstacle detection method and Adaboost cascade detection method are used to detect obstacles in indoor experiments, and the Producer's Accuracy, the User's Accuracy, the Overall Accuracy, and κ are used as evaluating indicators to compare test results, and the results show that the Vision-IMU based obstacle detection method has higher accuracy. The processing time of the Vision-IMU based obstacle detection method and Adaboost cascade detection method are compared, and it is shown that the Vision-IMU based obstacle detection method has faster processing speed.