- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
Hi,
I'm using ncs v1 with a pi 3+. With one ncs I have 8fps, 2: 15 fps. But with 3, the fps number doesn't change, and with 4 ncs, the fps slows down to 7 .
- Tags:
- Raspberry Pi
Link Copied
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
@spi100
Are you using Python?
If you are using Python, the performance of MultiThread will be exacerbated by GIL (Global Interpreter Lock).
Performance improves when using MultiProcess.
Sample of MultiProcess + MultiThread + MultiStick + MultiGraph + etc…
https://github.com/PINTO0309/MobileNet-SSD-RealSense.git
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
@PINTO yes, I do. Thanks for the tip, I'm going to have a look.
Regards
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
@PINTO I've read your git hub page, a clear and pretty good job. Now I've a better understanding of what you explained in your previous post.
If I want to reach the 25 fps real time, at first, to avoid the GIL issue, I can write my code in C or C++.
Perphaps also using an aaeon card and a realsense camera is likely a better solution than using a simple pi 3+ with a piCam.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
@spi100
If I want to reach the 25 fps real time
I realized the performance of over 30 FPS over NCS2 x2 + RaspberryPi3.
Perphaps also using an aaeon card and a realsense camera is likely a better solution than using a simple pi 3+ with a piCam.
There are points to be noted.
The minimum resolution of RealSense is 640x480.
Inference by NCS or NCS 2 is very fast, but it is a very bottleneck because of the high load (640x480) on the preprocessing and post-processing CPUs.
My benchmark understands that pre-processing and post-processing abilities are overwhelmingly deficient in ARM CPUs.
Perhaps aaeon is a very good choice.
If you plan to introduce OpenVINO, I encourage you to refer to this benchmark.
When using Intel's CPU, OpenVINO has been optimized considerably.
https://ncsforum.movidius.com/discussion/1329/lattepanda-alpha-openvino-cpu-core-m3-vs-ncs1-vs-ncs2-performance-comparison
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
@PINTO Hi, I've tried to run your sample
python3 SingleStickSSDwithUSBCamera_OpenVINO_NCS2.py -mod 1 -wd 640 -ht
480
after installing openVino on my pi3+.
I haven't any error message displayed on the console but no window is displayed on the screen.
Have you got any idea ?
regards
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
The call to cv2.VideoCapture(0) is not ended. I use 'modprobe bcm2835_v4l2 ' to simulate an usb camera with a piCam. It could be the root cause of my issue ?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
@spi100
The following posts may be helpful.
https://raspberrypi.stackexchange.com/questions/32840/camera-ubuntu-modprobe-bcm2835-v4l2-error-could-not-insert-bcm2835-v4l2-op
Have you run raspi-config?
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page