I am building an application for Real-time Video analysis using Deep learning and so far am successfully able to,

  1. Capture set of frame [with PiCamera@4fps/ 10 Sec Video] using picamer.capture_continuous.
  2. Feature extraction using ConvNet[pre-trained Model- InceptionV3] using Keras.
  3. Predicting the action in video.

I also want to simultaneously live stream with picamera. Issue here is using picamera.capture_continuous, keeps picamera engaged all the time and no other process can gain access to picamera.

My question here is there any work-around or better solution?

What I have tried?

  1. Am using opencv(3.3.0) in conda environment with RPi-Jessie and not able to do cv2.imshow due to GTK+2x error.
  2. Recording video to sd card every time seems impractical and useless.
  3. Am not sure about using multiprocessing for this problem.
  • Welcome to Raspberry Pi :-) But your question is too broad and opinion based. Such questions are flagged and may be closed. This site isn't made for brainstorming and discuss general issues. It is made to do one specific question that can be answered detailed. You may visit the Help Center to get an idea how things work here. – Ingo Apr 15 at 9:34
  • @Ingo Thanks for the pointer. will take the question to RPi forum or some related subreddit. – Amar P. Apr 15 at 12:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.