14/7/2024
Decided to start working on the codes for object detection using Intel RealSense D435 and YOLOv8.
3 objects have been chosen for the object detection:
- a stapler
- a can of drink
- a ball
* It must be kept in mind that, the depth sensor measures the depth from depth point start (from sensor tip/depth origin point) which is located -4.2mm inside the D435i camera’s front cover glass, therefore, to get a precise depth measurement, +4.2 mm must be added to the depth value.
I tried using the pre-trained model to detect the 3 objects, but the only one successfully detected is the ball. The rest were recognized wrongly (suitcase and mouse).
Attempted to identify the objects via an image:
It appears that a custom dataset would have to be prepared.
15/7/2024
I went to university to test the camera (and other cameras) but encountered the same issues. It was then discovered that the issue was the USB cable lead (it needs to us USB3.0). After obtaining the right lead, the camera worked.
After further discussion with the supervisor, we agreed to integrate my earlier context with the object / aim, which is to have a robotic arm act as an assistant to the dentist / surgeon by passing surgical tools to said person. Therefore, the objects to be picked and place have been changed to tools that resemble surgical tools such as polymer clay tools. Procurement has been made for the tools.
Another suggestion was made to change the fingertips of the RG2 gripper to make it easier to grip the "surgical tools". One option was to 3D print the fingertips using carbon fiber filled nylon. A 3D printed fingertip more suitable for gripping small cylindrical object was found and sent for printing:
20/7/2024
My tools have arrived and is placed in the lab for collection. I will collect the rolls tomorrow and commence taking pictures to create the necessary dataset for object recognition.
No comments:
Post a Comment