Velocity Estimation for Detected Objects in Wide Angle Monocular Vision of Multi-Shift Stereo Camera

Abstract

We developed a multi-shift stereo camera and a monocular vision detection (MSMVD) employing top view image subtraction method. Far object detection performance can be preserved while obtaining a wider FOV without the use of an expensive image sensor. We also propose two methods in velocity estimation for MSMVD. One is using single ranging result and view angle. Another is a combination method with Kalman filter. We confirmed the improvements in accuracy and responsiveness of estimated velocity and found that the collision judgement in the case of crossing bicycles within only 3 frames which is approximately half of the conventional method.