Crowd Flow Measurement System
Micro Hypercube Cloud Crowd Flow Measurement System
- This is how we measure crowd flow using micro Hypercube Cloud and IP cameras.
- Hypercubes servers have been added with cameras and now able to log images to Hypercube Server over Internet and server software to curate pictures into hours, minutes and seconds.
- These images are then amenable to processing by OpenCV and Python programs
- For example, we build simple QR code reader. When it sees the QR code it will shout out a message as a demo of the working system that is using OpenCV
- With OpenCV, we can use it to count the number of people in the image and log it to a database.
- From this data we can work out how many people flow into an area and flow out to other areas by linking together a lot of cameras and writing custom software.
- We can extend the system by writing the appropriate software to recognize other items such as wheel chairs, baggage, prams, number of children, people carrying shopping etc.
- We build a web interface in Python, PHP and MySQL to access the images from the servers:
- In this example application, we can pull up a picture every hour and see the road conditions for demonstration, and thus work out road usage by parked cars.
- We can write openCV programs to detect people in the road every second and plot a graph of people flow every second.
- This can be tied to other cameras to work out the number of people walking into any area and leaving the area.
- In a civic context, typically these systems are used to measure traffic flow conditions in stations, how tourists percolate throughout a venue or how tourists percolate throughout a complete city during special events.
- Micro Hypercube Cloud is more than adequate for such a large task. The data is physically secure. The hardware is relatively inexpensive. The energy costs 10x less than a 19" rack based data center. Most of the expenditure is spent on people and offices to customize the software which from a civic point of view creates lots of jobs.