Hi ROS community,
I’m new to robotics and automation and so grateful for all the work you’ve all put into ROS. I just built my first project last week that uses OpenCV with pretrained YOLOv3 weights to identify apples and target an Arduino Braccio arm to pick them up.
It pulls images from a standard 2D webcam (an old cell phone running DroidCam) and passes the detected coordinates of the apple’s bounding box to a MoveIt python script. I’ve written a bare bones transform to convert from the image frame to the robot frame of reference for pick up.
For more details and a link to the source code, you can check out my post.
I plan to expand on this project to get the arm to target veggies, put them through a food processor and assemble salads/burrito bowls. Just want to say again what a huge help the ROS community has been in jump-starting this project.