Right now if you want to acquire from multiple cameras and then stitch these images together in an aggregate image for processing you must allocate individual buffers for each frame grabber and buffers for the aggregate. In addition, you must perform N image copies where N represents the number of frame grabbers in your system. This is just dead overhead.
What I propose is to change the driver so that each frame grabber can point to a portion of a larger aggregate image. Then as each image comes in the aggregate image is populated directly. This save N memory copies and N x M image buffers (where M is the number of buffers in a ring acquisition). This would be a tremendous performance improvement. The figure below shows the idea.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Any idea that has received less than 8 kudos within 8 years after posting will be automatically declined.