Embedding range information in omnidirectional images through laser range finder

Abstract

Robot map navigation and localization are challenging tasks that require the solving of the data association problem for local and global features. Data fusion allows the advantages of two or more sensors to be combined, and complementary cooperation can be obtained. This paper presents two methods to embed depth information in omnidirectional images using the extrinsic calibration of a 2D laser range finder and a central catadioptric camera. The methods presented do not require a visible laser beam, but they assume the planar checkerboard patterns are visible for both the catadioptric camera and the 2D laser range finder. Unlike other approaches, the methods proposed used an invisible laser trace, and they are evaluated at pixel error level using ground truth data from the calibration patterns projected in the omnidirectional image. Results include a mean square error analysis of all calibration poses, and laser point projection on indoor omnidirectional images. We think that embedding range information in omnidirectional images is an interesting tool for data fusion approaches, which can be used in robot map building and localization.

Vertical edge‐based mapping using range‐augmented omnidirectional vision sensor

“…A Levenberg–Marquardt algorithm was used to minimise (3), which need an initial guess. In [13], three different ways to obtain it were presented; however, the best results were obtained using the linear least squares version of (2) applied to two different calibration planes. Afterwards, the initial guess gave a rank‐2 rotation matrix, since the resulting matrix is not a proper rotation matrix, because it does not satisfy RR T = I .…”

“…The URG‐04LX 2D LRF used in this work was previously calibrated in order to decrease the range error following the procedure described in [13]. The raw 2D laser scan data are previously processed using a median filter to discard spurious readings.…”

“…The problem focuses on finding R and T , the rotation and translation matrices of the camera with respect to the LRF, so that laser points P L can be represented in the camera coordinate system, and then projected onto the omnidirectional image. According to [13], the laser points ( P L ) that belong to the calibration plane satisfy a geometric constraint based on the distance between the camera and the calibration plane. This constraint can be expressed by (1) …”

“…In this paper, we present a sensor model based on the extrinsic calibration between a LRF and an omnidirectional camera [13], in order to robustly extract the 3D position of vertical edges in indoor environments and use them to solve the SLAM problem. Data association is very important for probabilistic frameworks that deal with mapping and localisation.…”

Indoor SLAM using a range-augmented omnidirectional vision

“…In this paper we present an implementation of the FastSLAM algorithm using a sensor model based on the extrinsic calibration between a LRF and an omnidirectional camera [12]. The main goal of our approach is to extract the 3D position of vertical lines in indoor environments and use them to solve the SLAM problem, instead of using the bearing information only as is done in [8] and [9].…”