So I have a project that requires Computer Vision (video stream). But when I want to try my Pi Camera (not USB), there's two method that I can use. They are PiCamera.capture_continuous and cv2.VideoCapture.read().

Both of these methods do well... But I'm wondering is there any difference to use one over another? Is there some performance issue with opencv's implementation? I found that if I want to use opencv, someone said I must activate the driver with modprobe bcm2835-v4l2. But in my RPi, I don't need to do that... Is that a latest firmware autoloaded the driver? Or the OpenCV did autoload the driver?

I also found that someone said I can't use VideoCapture for this CSI camera, but that said 4 years ago. Here I can use it with no problem. Is this updated on the newest firmware or OpenCV?

In case if you are wondering, I'm using newest OpenCV 4.1.0, cross compiled myself.


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.