3ZG-04

交差反復処理による深度カメラと地面のなす角度の推定

This paper proposes a principle statistical approach for GP (ground plane) detection by using depth camera. Current GP detection algorithms have focused on calculating the minimum distance to estimating the ground plane, while less of them have considered the probability distribution and the angle of the camera to improve the performance.

In this paper, we employ the KDE (Kernel Density Estimator) to obtain the probability density of the height distribution, with which the GP can be estimated. By using the estimated GP we calculate the angle of camera, and then feedback it to update GP until a certain condition has been reached. We evaluate our algorithm named as AGP-CIC(Angle-GP cross iterative convergence) on many challenging scenes and compare it with other distance-based algorithms.

In this paper, we employ the KDE (Kernel Density Estimator) to obtain the probability density of the height distribution, with which the GP can be estimated. By using the estimated GP we calculate the angle of camera, and then feedback it to update GP until a certain condition has been reached. We evaluate our algorithm named as AGP-CIC(Angle-GP cross iterative convergence) on many challenging scenes and compare it with other distance-based algorithms.