A monocular distance estimation method used in video sequence

摘要

A new monocular distance estimation method used in video sequence is proposed. The distance estimation in single frame is implemented by integrating the pin-hole model with edge points. Subsequently, the Hough Transform restriction based on Kalman filtering is employed in real-time distance estimation. The algorithm is capable of restricting the range of the object identification and tracking the target, so it improves the accuracy and robustness of distance estimation in video sequence. The experimental results show that the proposed distance estimation algorithm outperforms other methods which integrate the distance estimation methods used in single frame directly.