xwqwenqi / counting-cars

Counting cars with Movidius NCS2 and Raspberry Pi

counting-cars

Counting cars with Movidius NCS2 and Raspberry Pi

Counting cars with OpenVINO running on Raspberry Pi 3 Model B (RPi) and Intel® Core™ i7-4790K CPU @ 4.00GHz × 8 (desktop)

Requirements

• Raspberry Pi 3 Model B
• Movidius Neural Compute Stick 2 (not tested yet on NCS)
• Webcam
• Display (Optional)

Performance (RPi)

No Mask, without resizing

python3 realtime_objectdetection_and_tracking.py

[INFO] starting video stream...

[INFO] frame size (W x H): 1920 x 1080

[INFO] elapsed time: 26.12

[INFO] approx. FPS: 46.51

Mask, resize

python3 realtime_objectdetection_and_tracking.py --mask 200,350,650,550 --resize 1024

[INFO] starting video stream...

[INFO] frame size (W x H): 1024 x 576

Observation mask (top left, bottom right): [(200, 350), (650, 550)]

[INFO] elapsed time: 28.05

[INFO] approx. FPS: 10.70

Mask

python3 realtime_objectdetection_and_tracking.py --mask 200,350,650,550

[INFO] starting video stream...

[INFO] frame size (W x H): 1920 x 1080

Observation mask (top left, bottom right): [(200, 350), (650, 550)]

[INFO] elapsed time: 21.76

[INFO] approx. FPS: 43.42

Desktop

No mask, without resizing

python3 realtime_objectdetection_and_tracking.py

[INFO] starting video stream...

[INFO] frame size (W x H): 1920 x 1080

[INFO] elapsed time: 23.19

[INFO] approx. FPS: 172.81

Mask

python3 realtime_objectdetection_and_tracking.py --mask 200,350,650,550

[INFO] starting video stream...

[INFO] frame size (W x H): 1920 x 1080

Observation mask (top left, bottom right): [(200, 350), (650, 550)]

[INFO] elapsed time: 20.83

[INFO] approx. FPS: 171.83