I implemented an app on cpu whose main payload is opencv remap operation,now I need to move the app to Jetson AGX Orin and do remaps on gpu.
I planned to do it by vpi because there just are interfaces like vpi.remap(),but I found it was not that easy as I thought.There is my oringinal code:
result = cv2.remap(
fisheye_img,
map_x,
map_y,
interpolation=cv2.INTER_LINEAR,
borderMode=cv2.BORDER_CONSTANT,
borderValue=0,
)
as you can see map_x and map_y is pre-caculated,now I want to remap the img by map_x and map_y on gpu to gain speed,what should I do?