Exploiting known unknowns: Scene induced cross-calibration of lidar-stereo systems

Abstract

We propose an automatic, targetless, data-driven, extrinsic calibration method to calibrate push-broom 2D lidars with a multi-camera system. The calibration problem is decoupled into alternating optimisers over two hierarchical levels, where both levels are linked with a penalty term. The lower-level optimises the six degrees-of-freedom (DoF) rigidbody transforms between the lidar and each camera of the multi-camera unit by minimising the Normalised Information Distance between intensity measurements obtained from both sensor modalities. The upper-level minimises a nonlinear least squares error between the lower-level solutions. We describe the theory, implement the method, and provide a detailed performance analysis with experiments on real-world data.

The Road is Enough! Extrinsic Calibration of Non-overlapping Stereo Camera and LiDAR using Road Information

“…Scott et al [14] leveraged the Normalized Information Distance (NID) metric to estimate optimal extrinsic parameter between 2D LiDAR and multiple cameras. In the optimization process, they utilized factory-calibrated transformation as constraint.…”

“…However, these approaches rely on an overlapping FOV between two sensors. In order to overcome the non-overlapping configuration issue, a local 3D pointcloud is accumulated using the odometry to ensure that the overlapping region appears on the camera [14,15,16]. Another line of study [17,18,19], referred to as handeye calibration, estimates the motion of each sensor and performs extrinsic calibration based on the estimated motion relationship.…”

Accurate laser scanner to camera calibration with application to range sensor evaluation

“…This approach requires multiple frames and varied scene geometry such that sufficient corresponding edges can be found and a smooth cost function can be obtained. Scott et al [23] presented an approach for the calibration of a laser scanner and cameras setup that also exploits reflectivity information. The method is suited for setups that move through an environment, e.g., in an autonomous driving scenario.…”

Autonomous golf cars for public trial of mobility-on-demand service

“…self-contained versus reliance on infrastructure), the different approaches vary primarily on their level of redundancy. In terms of localization, the following combinations have been used in conjunction with odometry: infrastructure magnets [8]; LIDAR only [10]; LIDAR and camera [11]; LIDAR, camera, and DGPS [12]. In terms of obstacle detection, the following combinations have been used in conjunction with LIDAR: LIDAR only [6]; ultrasonic [8], [13]; camera, radar, and infrared sensors [14].…”