Are there any good repo sources for code to demonstrate Argus to OpenCV pipeline? I want complete control over the camera settings that Argus can offer but have existing opencv code to process the frames. C++ and Python are of interest.
Thanks
Are there any good repo sources for code to demonstrate Argus to OpenCV pipeline? I want complete control over the camera settings that Argus can offer but have existing opencv code to process the frames. C++ and Python are of interest.
Thanks
chuckis277,
Currently there is no public sample for it. You could convert that buffer to cpu through NvBufferMemMap and NvBufferMemSyncForCpu.
Hi chuckis277,
we had a sample 11_camera_object_identification which demonstrates Argus → openCV → caffe. We encourage users to run TensorRT, so the sample is removed in r28.2.1.
Please refer to the sample in r28.1 and below patch of removing caffe:
[url]https://devtalk.nvidia.com/default/topic/1010111/jetson-tx1/nvmm-memory/post/5162049/#5162049[/url]
You may try to port it to r28.2.1 for your case.
Please follow below steps for running Argus → OpenCV:
1 Install OpenCV https://github.com/AastaNV/JEP/blob/master/script/install_opencv3.4.0.sh
2 Apply attached patch and rebuild 09_camera_jpeg_capture
3 Run
$ export DISPLAY=:0
09_camera_jpeg_capture$ ./camera_jpeg_capture --disable-jpg
09_camera_jpeg_capture.txt (5.66 KB)
Thanks, patching 09_camera_jpeg_capture was perfect. Much appreciated.
hey @daneLLL, I cannot find the sample that you are referring about argus to OpenCV
could you post a link ?
thanks in advance
Please install tegra_multimedia_api via Jetpack. You shall see 09_camera_jpeg_capture. The patch is at comment #4