DADDI@CRAS - Durius Aerial Dataset with Diversified Information

The Durius Aerial Dataset with Diversified Information (DADDI) was acquired using an Unmanned Aerial Vehicle (UAV), the RAVEN, equipped with an heterogeneous perception system, the TriOPS, during an inspection mission of the DURIUS offshore floating structure. DURIUS is a moored floating structure that consists of a decommissioned Catenary Anchor Leg Mooring (CALM) buoy designed to simulate the floating column of an offshore wind turbine. It boasts a diameter of 16 meters and a total height of 6 meters, with 2 meters submerged. DADDI provides valuable data from an offshore structure, facilitating comprehensive multimodal inspections and the creation of 3D model representations. It provides real-world data for developing and testing algorithms, models, and technologies related to inspection, monitoring, and maintenance of complex structures, in particular in the maritime environments. This dataset was acquired during a campaign of the project ATLANTIS on the ATLANTIS Coastal Testbed, in the harbor of Viana do Castelo, Portugal (location of the ATLANTIS Coastal Testbed: https://maps.app.goo.gl/2LdS3C9KSeB2XD4J7). In this mission, RAVEN was remotely operated while collecting synchronized multimodal data of the DURIUS structure from different points of view at 10~Hz, gathering visual images (from a HIKROBOT MV-CA016-10UC visual camera), thermal images (from a FLIR Boson 640 Radiometry thermal camera) and point clouds (from a Ouster OS1-64), as well as of the navigation sensors such as RTK GPS (UBLOX F9P with IST8310 compass) and IMUs (ICM20649, ICM20602 and ICM20948, with MS5611 barometer). The dataset comprises over 2500 images captured by both visual and thermal cameras, complemented by a corresponding number of Point Cloud Data (PCDs) generated by the 3D LiDAR. The correspondent intrinsic and extrinsic calibration of these sensors are also provided. The data collected from the GPS and IMU sensors have also been recorded in CSV files, with GPS data (latitude, longitude and altitude), IMU data (orientation, angular velocity and linear acceleration), and Odometry data (local pose: position and orientation, and local velocity: linear and angular). The timestamp indicating the collection moment for each data type is embedded in both the image and PCD names, as well as in each line of the CSV files. All this data is also provided through ROS topics saved into rosbag files."