Develop obstacle_detector Action

[charlesmadjeri]

Description:

Implement the obstacle_detector Action to detect and localize obstacles using camera images, camera orientation and IMU data.

Tasks:

• Subscribe to /usv/hardware_interface/image_raw for visual input

• Subscribe to /usv/hardware_interface/imu/data for orientation data

• Subscribe to /usv/hardware_interface/camera_pos for camera orientation

• Implement obstacle detection and localization logic:

  • Make the camera moving 360° in loop to find obstacles as a radar.
  • If artifact found check if it is an obstacle, and find its type
  • Save image of the obstacle, gps position, imu data (ship orientation) and camera orientation
  • Calculate estimated gps position of the obstacle (estimating object distance from 2d image taken)

• Publish detected obstacle estimated position and type on usv/perception/obstacle_detected

• Ensure this analysis is only running and consuming ressources when the action is triggered, and stop it when terminated.