We plan to record datasets using the ZED 2i mounted on a robot. However, we currently can’t use the GPU for the ZED 2i on the robot. I was thinking of simply recording the stereo image pairs—which I believe is possible without a GPU—and then, offline on another computer equipped with a GPU, generating the depth maps using the ZED 2i depth NN to complete the dataset. Is that feasible?
Hi @martinakos
you can use ZED Explorer to record SVO files in JPEG or LOSSLESS mode if you do not have a GPU.
Next you can use the SVO file as input to the ZED SDK for offline processing.