Using the disparity space to compute occupancy grids from stereo-vision

Abstract

The occupancy grid is a popular tool for probabilistic robotics, used for a variety of applications. Such grids are typically based on data from range sensors (e.g. laser, ultrasound), and the computation process is well known [1]. The use of stereo-vision in this framework is less common, and typically treats the stereo sensor as a distance sensor, or fails to account for the uncertainties specific to vision. In this paper, we propose a novel approach to compute occupancy grids from stereo-vision, for the purpose of intelligent vehicles. Occupancy is initially computed directly in the stereoscopic sensor's disparity space, using the sensor's pixel-wise precision during the computation process and allowing the handling of occlusions in the observed area. It is also computationally efficient, since it uses the u-disparity approach to avoid processing a large point cloud. In a second stage, this disparity-space occupancy is transformed into a Cartesian space occupancy grid to be used by subsequent applications. In this paper, we present the method and show results obtained with real road data, comparing this approach with others.

Computing occupancy grids from multiple sensors using linear opinion pools

“…For computation of occupancy grids with stereo-vision, our method is based on the visibility approach proposed in [10]. The idea is that an occupancy grid is computed directly in the u-disparity plane -which is a horizontal plane in the disparity space associated to the stereo camera -while taking into account the geometrical visibility of each cell.…”

“…W vis stereo is then obtained by remapping on the Cartesian plane the visibility values estimated during the disparity grid computation (see [10] for more details about this computation). W dist stereo is a confidence value related to the precision of stereo-vision.…”

Using obstacles and road pixels in the disparity-space computation of stereo-vision based occupancy grids

“…By contrast, moving to a Cartesian space first would give a varying density of measurements. The basis of the method has been presented in [7]. Here we propose to improve the approach by taking into account the road pixels from the disparity image, in order to refine the estimation of the free space.…”

“…The occupancy grid computation will be briefly described here for clarity. Please refer to [7] for more details.…”

“…As described in [7], the method considers the visibility of different regions of the image to deal with partially occluded objects, and is computationally efficient. The work here further accounts for road pixels and obstacle pixels both being visible.…”