Jetson Nano OpenCV OpenCL/CUDA SURF

Hello,

Is someone can share is information about SURF algorithm performance with those chips
with OpenCV OpenCL/CUDA environment ?

Hi,

Sorry that we don’t have benchmark data of OpenCV module.
In our latest JetPack, there is a OpenCV v4.1.1 pre-installed, you can test it directly.

Thanks.

Hi,

We didn’t buy it yet I need any information vs Snap Dragon 820 of Qualcomm

we didn’t decide yet.

Hi,

We can share some performance from our device directly.
Although it won’be an official report, it still can give you some idea about the performance.

Is this sufficient for you?

Thanks.

Yes please

Please share link or code for surf orb sift

Hi,

Sorry for the late.

We try to profile the SURF feature on Jetson but found the algorithm is patented.
So we may not able to get the performance data for you.

terminate called after throwing an instance of 'cv::Exception'
  what():  OpenCV(4.1.1) /home/nvidia/topic_1068663/opencv_cuda/opencv_contrib-4.1.1/modules/xfeatures2d/src/surf.cuda.cpp:55: error: (-213:The function/feature is not implemented) This algorithm is patented and is excluded in this configuration; Set OPENCV_ENABLE_NONFREE CMake option and rebuild the library in function 'SURF_CUDA'

Aborted (core dumped)

Thanks.

Hi,

We rebuilt the OpenCV and test the SURF sample shared from this page for you:
https://answers.opencv.org/question/87165/surf-on-cuda-every-execution-produces-different-weird-results-for-some-images/

The experiment is tested on the Jetson Xavier.
For a 800x800 image, it takes 0.003724s to find 513 key-points.

Thanks.

Thanks, but I need information about the nano.

The Xavier is expensive card