Vehicle Ego-Localization by Matching In-Vehicle Camera Images to an Aerial Image

Abstract

Obtaining an accurate vehicle position is important for intelligent vehicles in supporting driver safety and comfort. This paper proposes an accurate ego-localization method by matching in-vehicle camera images to an aerial image. There are two major problems in performing an accurate matching: (1) image difference between the aerial image and the in-vehicle camera image due to view-point and illumination conditions, and (2) occlusions in the in-vehicle camera image. To solve the first problem, we use the SURF image descriptor, which achieves robust feature-point matching for the various image differences. Additionally, we extract appropriate feature-points from each road-marking region on the road plane in both images. For the second problem, we utilize sequential multiple in-vehicle camera frames in the matching. The experimental results demonstrate that the proposed method improves both ego-localization accuracy and stability.

Bundle adjustment using aerial images with two-stage geometric verification

“…On the other hand, some methods estimate camera poses directly from aerial images [13][14][15][16][17][18]. There are two types of aerial images: perspective and orthographic.…”

“…These methods can be classified into learning- [14] and feature-matching-based [15][16][17][18]. Lin et al [14] proposed a method based on the relationship of the appearance between ground-view and aerial images learned through community photos with position information.…”

“…The existing methods using aerial images [12][13][14][15][16][17][18] are based on feature matching between given aerial images and ground-view images taken by standard cameras. Unfortunately, existing methods can handle only a short image sequence that does not include difficult situations.…”

Satellite image based precise robot localization on sidewalks

“…Over the last decades, researchers have been proposing different outdoor localization methods that use various sensors. Among these, global positioning system (GPS) and inertial navigation system (INS) based methods [5], [6], landmark based simultaneous localization and mapping (SLAM) methods [7], [8] and map-matching methods [9][10][11][12] are the most common ones.…”

“…The study in [10], has used iterative corresponding point (ICP) algorithm to match lane markings on aerial maps with features on camera images. A similar work in [9] used the homography between camera and satellite maps in order to warp camera images to their top view equivalents and to match on-road features, where a feature map of the environment is prebuilt. The work in [12], obtains accurate top view equivalents of onboard stereo images by projecting the 3D point cloud of the scene observed by the camera.…”