smart-queue-system

Smart Queue System using Intel OpenVINO for inferencing. This uses queue regions to manage people count per queue.

Setup

1. Install OpenVINO Toolkit 2019 r3
2. Setup environment using Python 3.6

source /opt/intel/openvino/bin/setupvars.sh -pyver 3.6

3. Download the model for detecting a person: person-detection-retail-0013.

Running

There are 3 scenarios

• Manufacturing
• Retail
• Transportation

$ python3 main.py \
  --model person-detection-retail-0013 \
  --video videos/<scenario>.mp4 \
  --threshold 0.8 \
  --output_path results/<scenario> \
  --queue_param data/<scenario>.npy

Sample:

$ python3 main.py \
  --model person-detection-retail-0013 \
  --video videos/manufacturing.mp4 \
  --threshold 0.8 \
  --output_path results/manufacturing \
  --queue_param data/manufacturing.npy

Output

Manufacturing Scenario

Predict person location in frame to see if the production line is well-supported.

Video: Download

Retail Scenario

Detect number of customers in queue and be able to redirect customer/s to less congested queues.

Video: Download