I just installed the cuda_aruco package on the nvidia jetson TX2 and run a test on simple image using the ./aruco_test in the "utils" folder.
Here is the results I am currently getting when processing an image of 2448x2048pixels
A simple Python implementation on the same Jetson TX2 and processing the same image is done in ~~150ms.
So my question to you: am I using correctly the "aruco_test.cpp" because I was expecting the cuda_aruco implementation of the detector to be much faster than the Python!
Hi @nbergst
I just installed the cuda_aruco package on the nvidia jetson TX2 and run a test on simple image using the ./aruco_test in the "utils" folder. Here is the results I am currently getting when processing an image of 2448x2048pixels A simple Python implementation on the same Jetson TX2 and processing the same image is done in ~~150ms.
So my question to you: am I using correctly the "aruco_test.cpp" because I was expecting the cuda_aruco implementation of the detector to be much faster than the Python!
Regards Merwan