Box measurement: speed and precision when using one or multiple cameras #13623
Replies: 1 comment
-
Hi @NgyAnthony Thanks very much for your questions! I am not aware of any negative performance consequences from having more than one camera with the box_dimensioner_multicam program. However, adding more than one camera to the same computer does consume additional computer resources for each camera that is attached. For a 4-camera setup, a computer with an Intel Core i7 or i9 processor or equivalent capability is recommendable. For 2 cameras, a computer with a lesser specification can be used. When you are using a single camera, the positioning of the camera around the chessboard does not matter as much as long as the camera is around 0.5 meters from the board (the optimal 'sweet spot' distance for box_dimensioner_multicam) and is positioned somewhere at the edge of the board facing diagonally downwards from above. The camera should not be placed directly above the center of the board facing straight downwards. When using multiple cameras, putting each camera at the corners of the chessboard is the optimal way for box_dimensioner_multicam to automatically calculate each camera's position relative to the other camera positions. As an alternative to box_dimensioner_multicam, you could investigate the pre-made RealSense powered commercial box dimensioning tool provided by MobileDemand. https://www.mobiledemand.com/industry-insights/intel-realsense-xdim-nrf-2024 |
Beta Was this translation helpful? Give feedback.
-
Hi everyone,
I'm in the market for depth cameras for my simple use case: measuring lots and lots of boxes.
I'm happy to tell you that software-wise RealSense has a lower barrier to entry than your competitors and it's very nice to see @MartyG-RealSense properly handling support on all issues that are raised which is a big + in my opinion.
For my use case, I have stumbled across this tutorial: Box Measurement and Multi-camera Calibration
and this video:
Measuring Width, Height and Length of object using single RealSense D435 camera
My question is the following:
In a controlled environment such as a case with:
What are the differences in speed and precision when using one camera (as seen in the video) as opposed to two or more cameras ?
Follow-up question: in the tutorial, why did the author make the choice of positioning the two cameras on the same horizontal axis as opposed to a parallel axis ? (opposed from one another)
Thank you !
Beta Was this translation helpful? Give feedback.
All reactions