Fast, unconstrained camera motion estimation from stereo without tracking and robust statistics

Abstract

Camera mohon estmation is useful for a range of apphcations. Usually, feature tracking is performed through the sequence of images to determine correspondences. Furthermore, robust statist~cal techniques are normally used to handle large number of outliers in correspondences. This paper proposes a new method that avoids both. Motion is calculated between two consecutive stereo images without any pre-knowledge or prediction about feature location or the possibly large camera movement. This permits a lower kame rate and almost arbitrary movements. Euclidean constraints are used to incrementally select inliers from a set of mitial correspondences, instead of using robust statistics that has to handle all inliers and outhers together. These consi"are so strong that the set of initial correspondences can contain several times more outliers than inliers.Experiments on a worst-case stereo sequence show that the method is robust accurate and can be used in real-time.

Toward a Fully Autonomous UAV: Research Platform for Indoor and Outdoor Urban Search and Rescue

“…Based on this 3-D information, the six-dimensional delta position and orientation between two points in time as well as the corresponding measurement covariance are calculated [15]. The algorithm supports a key frame buffer so that the delta measurement refers not just to the last acquired image but also to the image in the buffer that gives the delta measurement with the smallest absolute covariance.…”

Centre National d’études Spatiales

“…In this way, our work differs by focusing on large displacements. The proposed method is related to the work presented in [6]. Yet, it mainly differs on two points.…”

Stereo camera based navigation of mobile robots on rough terrain

“…Position estimation is based on visual odometry with 6 degrees of freedom computed from the left stereo image and the corresponding depth image [6]. A Harris Corner Detector is used to detect features in consecutive left camera images.…”