Hi,
We’ve been working on a project to develop a better way of measuring a baby in an incubator. To do this, we create a 3D scan using 4 Astra stereo S U3 cameras, and from this 3D scan we measure the length of the baby, the head circumference and the cranial volume.
We would like to use the highest depth resolution (1280x800) in order to preserve the details abound the ears and nose better. This is important, since we must be able to recognise specific parts on the ear and nose to determine the head circumference and cranial volume.
We are currently running into a problem with the cameras.
Around the edges of the baby there are large holes in the depth image. We do understand this is happening because of the use of stereo vision however, we noticed these holes are significantly smaller when using a lower resolution (640x400).
We would like to know if there is a reason this is happening, and if there is something we can do to improve the fill rate on a higher resolution?
I’ve attached two depth images below using a resolution of 1280x800 and 640x400.
