D415 - Crop images / reduce FOV for closer range?
Hello,
I want to use the realsense D415 camera to scan objects in close range. To reduce minimumZ I can use disparity shift and/or reduce resolution of the depth imagers, as stated in the documentation.
In order to maintain maximum depth resolution per degree, is it possible to crop the boundaries of the images used for depth calculation (instead of simply reducing resolution but keeping the FOV of the image)?
This would reduce the field of view, but keep the resolution in my region of interest - I need to scan an area of only 200x100mm and want to achieve maximum resolution of this area.
(besides I am using the D415 via python 3.5)
Thank you for your help.
-
Another Python user recently asked the same question. It is somewhat difficult but may be possible by creating a 'bounding box' and then occluding data outside of that bounding box, as described in the link below.
https://github.com/IntelRealSense/librealsense/issues/2016#issuecomment-403804157
The linked example is for a point cloud and so may not be directly adaptable if you are using a 2D image. It is the best reference I know of though for demonstrating the viability of using a bounding box for cropping the image.
-
The bounding box discussed in the linked post crops the image not only in Z depth but in X and Y too. In that particular case, the user found though that they didn't need to crop the XY of the image after all and could manage with min and max Z depth-clamp values.
The link below has another possible method for cropping a RealSense image by manipulating the XYZ coordinates manually.
https://github.com/IntelRealSense/librealsense/issues/3894#issuecomment-491714494
To be honest, the difficulty of this task means that it may be better to stick with reducing the camera's minimum distance by the methods you are already using (disparity shift or reducing resolution).
Please sign in to leave a comment.
Comments
3 comments