首页|Recent Studies from University of Freiburg Add New Data to Robotics (Automatic T arget-less Camera-lidar Calibration From Motion and Deep Point Correspondences)
Recent Studies from University of Freiburg Add New Data to Robotics (Automatic T arget-less Camera-lidar Calibration From Motion and Deep Point Correspondences)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning DailyNews Daily News-Researchers detail new data in Robotic s. According to news reporting originatingfrom Freiburg, Germany, by NewsRx cor respondents, research stated, "Sensor setups of robotic platformscommonly inclu de both camera and LiDAR as they provide complementary information. However, fus ingthese two modalities typically requires a highly accurate calibration betwee n them."Financial supporters for this research include German Research Foundation (DFG), Nvidia Corporation.Our news editors obtained a quote from the research from the University of Freib urg, "In this letter,we propose MDPCalib which is a novel method for camera-LiD AR calibration that requires neither humansupervision nor any specific target o bjects. Instead, we utilize sensor motion estimates from visual andLiDAR odomet ry as well as deep learning-based 2D-pixel-to-3D-point correspondences that are obtainedwithout in-domain retraining. We represent camera-LiDAR calibration as an optimization problem andminimize the costs induced by constraints from senso r motion and point correspondences. In extensiveexperiments, we demonstrate tha t our approach yields highly accurate extrinsic calibration parameters andis ro bust to random initialization."
FreiburgGermanyEuropeEmerging Tech nologiesMachine LearningRoboticsRobotsUniversity of Freiburg
NETL
NSTL
