capture = cv2.VideoCapture(Gst-launch-1.0 nvarguscamerasrc ! “video/x-raw(memory:NVMM)” ! nvvidconv ! xvimagesink sync=false) like this？
What’s the problem here?
If just want to launch camera just enter those command to console for it.
Or reference to below link for OpenCv launch camera.
This is how I test, but there will be a 2s delay in the process of real-time object detection, how do I solve this problem? No tensorrt is used, only GPUs are used
Try high speed frame rate sensor mode. 2s delay is un-normally usually the delay is 120 ~ 150 ms for 30fps sensor output.
Please guide you on how to do it specifically, I don’t know much about this aspect, I use the imx219 camera.
Have a try below
... nvarguscamerasrc ! "video/x-raw(memory:NVMM), width=1280,height=720,framerate=60/1,format=NV12" ! nvvidconv ! xvimagesink sync=false
Hello, is this code running on nano’s terminal before running the real-time object detection program?
Or I changed this statement to the video stream input for object detection
Change it in your object detection code.
like this？
Yes
So why is this code different from the CSI camera input code I saw before, this needs to download an additional installation package into the environment?
Sorry I don’t understand your question. Please give much detail information would be great.
hi ,when i run the code ，it has some warning。
GStreamer warning: Cannot query video position: status=0, value=-1, duration=-1
Do you know which cause that？
it will have this probelm when i use “nvarguscamerasrc ! “video/x-raw(memory:NVMM), width=1280,height=720,framerate=60/1,format=NV12” ! nvvidconv ! xvimagesink sync=false”
Need change the xvimagesink to appsink for the OpenCv usage.
This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.