Self calibration of multiple LIDARs and cameras on autonomous vehicles

Pereira, Marcelo Silva; Schwartz, David A.; Santos, Vítor; Dias, Paulo

doi:10.1016/j.robot.2016.05.010

Cited by 70 publications

(42 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The previous solution with a Sick LMS151 sensor in [18] stated the absolute mean error and the standard deviation of the Euclidean distance between CPs. Table 2 shows a comparison between the accuracy of our solution and that of the previous solution.…”

Section: Calibration and Analysismentioning

confidence: 99%

“…According to [8], the evaluation of point clouds can be roughly divided into three approaches: the control point approach, the subset approach and the full point cloud approach. Calculating the residuals of the CPs is a classical method to assess the calibration accuracy [18], which belongs to the control point approach.…”

Section: Robustness and Accuracy Validationmentioning

confidence: 99%

“…In contrast to the utilization of special targets, another approach was proposed based on geometric constraints consisting of the coplanarity and perpendicularity of different laser line segments [11]; however, despite being an automatic calibration method, this approach lacks a criterion for selecting corresponding line segments. A more recent automatic solution is the use of a ball target to estimate rigid body transforms between different sensors, including 2D/3D LIDARs and cameras [18]; however, the result between 2D LIDARs exhibits a low accuracy insomuch that the orientation accuracy after calibration is 68 mm, which is five times greater than the sensor intrinsic error of 12 mm. In summary, the drawback to the abovementioned existing methods is that they rely on manual procedures or otherwise suffer from low accuracies, and thus, automatic methods with a reliable accuracy are still missing.…”

Section: Introductionmentioning

confidence: 99%

“…In the proposed method, a mobile sphere is used as a calibration target that allows the extrapolation of the center of the sphere and the matching of corresponding points (CPs), which is automatic similar to [18]. However, in contrast to the method in [18], our approach can match CPs under the restriction constraint of the scan circle radius and achieve the goal of a high-accuracy calibration. Based on the relationship between the CP accuracy and the scan circle radius, a CP selection machine with a low error is established under the conditional constraint of the scan circle radius, thereby improving the orientation accuracy.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Extrinsic Calibration of 2D Laser Rangefinders Based on a Mobile Sphere

Chen

Liu

et al. 2018

Remote Sensing

View full text Add to dashboard Cite

Abstract:In the fields of autonomous vehicles, virtual reality and three-dimensional (3D) reconstruction, 2D laser rangefinders have been widely employed for different purposes, such as localization, mapping, and simultaneous location and mapping. However, the extrinsic calibration of multiple 2D laser rangefinders is a fundamental prerequisite for guaranteeing their performance. In contrast to existing calibration methods that rely on manual procedures or suffer from low accuracy, an automatic and high-accuracy solution is proposed in this paper for the extrinsic calibration of 2D laser rangefinders. In the proposed method, a mobile sphere is used as a calibration target, thereby allowing the automatic extrapolation of a spherical center and the automatic matching of corresponding points. Based on the error analysis, a matching machine of corresponding points with a low error is established with the restriction constraint of the scan circle radius, thereby achieving the goal of high-accuracy calibration. Experiments using the Hokuyo UTM-30LX sensor show that the method can increase the extrinsic orientation accuracy to a sensor intrinsic accuracy of 10 mm without requiring manual measurements or manual correspondence among sensor data. Therefore, the calibration method in this paper is automatic, highly accurate, and highly effective, and it meets the requirements of practical applications.

show abstract

Section: Calibration and Analysismentioning

confidence: 99%

Section: Robustness and Accuracy Validationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Extrinsic Calibration of 2D Laser Rangefinders Based on a Mobile Sphere

Chen

Liu

et al. 2018

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…To overcome this issue, [14] used a glass made calibration board to ensure nearperfect flatness and rigidity. On the other hand, Park et al [11] and Pereira et al [12] employed objects in special shapes to estimate the point-to-point correspondences. However, their works were also sensible to the lack of rigidity.…”

Section: Introductionmentioning

confidence: 99%

An Interactive LiDAR to Camera Calibration

Lyu

Bai

Elhousni

et al. 2019

2019 IEEE High Performance Extreme Computing Conference (HPEC)

View full text Add to dashboard Cite

Recent progress in the automated driving system (ADS) and advanced driver assistant system (ADAS) has shown that the combined use of 3D light detection and ranging (LiDAR) and the camera is essential for an intelligent vehicle to perceive and understand its surroundings. LiDAR-camera fusion requires precise intrinsic and extrinsic calibrations between the sensors. However, due to the limitation of the calibration equipment and susceptibility to noise, algorithms in existing methods tend to fail in finding LiDAR-camera correspondences in long-range. In this paper, we introduced an interactive LiDAR to camera calibration toolbox to estimate the intrinsic and extrinsic transforms. This toolbox automatically detects the corner of a planer board from a sequence of LiDAR frames and provides a convenient user interface for annotating the corresponding pixels on camera frames. Since the toolbox only detects the top corner of the board, there is no need to prepare a precise polygon planar board or a checkerboard with different reflectivity areas as in the existing methods. Furthermore, the toolbox uses genetic algorithms to estimate the transforms and supports multiple camera models such as the pinhole camera model and the fisheye camera model. Experiments using Velodyne VLP-16 LiDAR and Point Grey Chameleon 3 camera show robust results.

show abstract

ATLASCAR: A Sample of the Quests and Concerns for Autonomous Cars

Santos

2019

Informatics in Control, Automation and Robotics

View full text Add to dashboard Cite

Self calibration of multiple LIDARs and cameras on autonomous vehicles

Cited by 70 publications

References 20 publications

Extrinsic Calibration of 2D Laser Rangefinders Based on a Mobile Sphere

Extrinsic Calibration of 2D Laser Rangefinders Based on a Mobile Sphere

An Interactive LiDAR to Camera Calibration

ATLASCAR: A Sample of the Quests and Concerns for Autonomous Cars

Contact Info

Product

Resources

About