Discussion

Left ArrowBack to discussions page
ZH2112ZH2112 Posts: 22 Apprentice
Or does it already work that way?

Comments

  • David_GouffeDavid_Gouffe Posts: 56Vacuum Beta tester Handy
    edited July 19
    Hi, 

     You are right and the short answer is YES: It already works that way.  

    Field of views on camera can be represented as a pyramid from the lens to the plane where we search for objects. 

    As it is a pyramid, the higher you are from the object, the bigger is the field of view ( FoV).

    In the Robotiq Camera manual, the field of view is described inside the section: " 7.3. Vision System Specifications". 
    Depending on the UR size (3 kg, 5 kg or 10 kg)  you use, work distance get greater with the robot reach. It is described in " Table 7 - 4: Camera field of view depending on the robot's reachable area "

    The minimum is always 10 cm x 7.5 cm as 7 cm of distance between camera and object is the minimum focus range. 

    With this minimum, minimum object detail detection is 1/10 of the field of view. At a minimum distance, the smallest detail reliable for detection is 1/10 of 10 cm. Meaning that 1 cm circle feature can be detected. This is a 10 mm circle. Going smaller than that can sometimes be detected but it might not be reliable, particularly edge and corner of the field of view. 

    The field of view is about 39 degrees from the vertical axis of the field of view, the same axis as the 30 degrees from tool flange Z axis, and 50 degrees in the lateral axis of the field of view.

    Let's recap all of this: the higher you get, the bigger is the field of view. The nearer you go, the smaller the features you can use for detection are. 

    If what you search is more than 6/10 of the field of view, it gets harder to detect. In this case, you are better to move away the camera to have a bigger field of view and reduce the perceived size of the object. 

    Let's add up a small bonus: 
    B1) Field of view matching the calibration plane is the best way to be accurate. 

    B2) It is still possible to create a snapshot position translated from the calibration position (X, Y, Z) (left, right, up, down, higher or lower) 

    B3) It is also possible to make a new calibration that matches the new tilted work plan. 

    You can find most of this information in the Robotiq Camera manual and the Robotiq URCap Camera Wizard inside the camera node. 
    https://robotiq.com/support (Select: Universal-Robots, Wrist Camera, Wrist Camera Instruction Manual )
    You can also learn how to use the camera with the E-Learning area of Robotiq webpage. 
Sign In or Register to comment.
Left ArrowBack to discussions page