Deepen AI has launched an advanced targetless calibration platform designed to streamline sensor calibration for autonomous vehicles, automotive ADAS, and robotics. This innovative solution accommodates various configurations, including GNSS receivers, multiple lidars, radars, cameras, and IMUs, processing inputs in one pass through a single ROS bag dataset. Traditional calibration methods can be bottlenecks due to their manual, iterative nature; Deepen AI’s platform automates this process, achieving up to 0.05° angular and 0.7 cm positional accuracy without needing physical targets.
The system can perform simultaneous calibration of all sensors during standard driving in everyday environments like parking lots or quiet streets, using just 30 seconds of continuous driving data. This enhances efficiency for both research and large-scale production deployments. The unified approach simplifies operations while delivering superior accuracy over traditional methods, making it ideal for teams needing to integrate complex sensor configurations seamlessly.
Source link
