CV Tasks #1
Description
Project Description: Autonomous 3D Photogrammetry of Coral Restoration Area
Objective:
Develop a computer vision pipeline that autonomously captures images from the ROV’s onboard camera, reconstructs a 3D model of the coral restoration area using photogrammetry, and produces accurate measurements of its height (±5 cm).
Likely Tools Needed:
ROS for navigation and image capture
ROV’s: Intel RealSense D455 (for now but likely change) or similar depth camera.
Scripting/automation: Python + OpenCV for frame extraction & preprocessing.
9/27: Depth Measurement Tasks above water
First connect the Realsense camera to the Nvidia Jetsons
Make sure Realsense sdk is installed on the jetsons, read up intel documentation for this
https://github.com/IntelRealSense/librealsense
Install its python wrapper, or whatever language you are most comfortable with
First task: Measure the distance between two points above water
Will worry about this later:
Point cloud processing: CloudCompare or MeshLab.
Photogrammetry tools: Agisoft Metashape (best, but paid), or open-source COLMAP/OpenMVG.
https://20693798.fs1.hu11bspotusercontent-na1.net/hubfs/20693798/2024%20EXPLORER%20Manual%20FINAL_1_16_2024_with_Cover.pdf
Page 30 onwards (Task 3.3)