The software was supposed to output information about parking occupancy given a camera feed.
Poland is drastically developing technologies in the area of urban efficiency and smart cities. The city of Katowice has one of the largest surveillance systems in eastern Europe and is developing complementary technologies to improve quality of life for its citizens. The entire region is now striving to become “smarter” and a part of that goal is being able to improve traffic efficiency and drivers’ comfort by real-time monitoring of parking spaces.
Current technologies involve expensive electromagnetic sensors that are installed under each parking space. Camera-based technology would theoretically provide much cheaper solution to the problem and could make use of existing surveillance systems.
Python, OpenCV, Google Cloud Storage
The software used a simple Python interface to mark parking spaces on the camera feed and extract high frequency information from these areas to decide if given space has been taken or not. This problem is a perfect case for Machine Learning of course but at the time I haven’t had that expertise yet, thus the use of classical Computer Vision.
The software has never been tested or sold, given that I left to study in Copenhagen and the company decided to use the electromagnetic sensors instead of camera-based solutions.
My contribution as Software Engineer
I developed the software myself.