부산대학교 도서관

For robotics and autonomous driving, accurate perception is very crucial for safety and security. The tasks of robotics and autonomous driving come in mission-critical tasks. Such applications expect timely and accurate object detection as a desideratum. For a safer trajectory towards critical systems, depending on more than one sensor is recommended. Humans rely upon multiple senses to perform a certain action, and as an inspiration, an autonomous robot may require multiple sensors for the entire perception in varying environmental conditions. In such a situation, a cheaper and robust sensor can be joined to attain the information synchronously. Problems occur while defining a correspondence that implicitly depends on the transformation between different coordinate systems. This transformation is part of the extrinsic calibration between the sensors. This study will present a simple and efficient way to find the extrinsic calibration between radar and a camera. We have tested and implemented this approach on our autonomous vehicle.