By default the wrist camera process detected objects from high to low detection score. However in some application like "tray picking" you may prefer to control picking order.This post present a program which process detected objects line by line. Here is the methodology:Detect tray position with the camera and save it in a variable. From this position, it is possible to locate all tray placements.Detect with the camera the position of all objects available in the tray and save it in a list.Use the distance between object position and tray placements to process objects in the desired order. The program use parametric model to detect tray and objects but it is also possible to make a similar program with automatic model.
As this example on how to set picking order with the camera was not well understood by everyone, I had some comments about the program.The 2 camlocates used in this program use parametric mode.The first camlocates search for a scare of the size of the tray.The second camlocate search for a circle of the size of the object in tray placement.I hope it will help.
I attach some additional information that might help understand parameters and logic.I hope it will help.
By default the wrist camera process detected objects from high to low detection score. However in some application like "tray picking" you may prefer to control picking order.
This post present a program which process detected objects line by line.
Here is the methodology:
- Detect tray position with the camera and save it in a variable. From this position, it is possible to locate all tray placements.
- Detect with the camera the position of all objects available in the tray and save it in a list.
- Use the distance between object position and tray placements to process objects in the desired order.
The program use parametric model to detect tray and objects but it is also possible to make a similar program with automatic model.