sylwekqaz / ultraparkingmachinelearning

The purpose of the project is to classify in real time the marked parking places on video as free or occupied. This is the fork of my diploma thesis. This fork add functionality to preview in realtime stream from camera and show algorithm classification result. This project is presentation for event WULS-SGGW days at Faculty of Applied Informatics and Mathematics stand.

Intelligent algorithm to recognize free parking spaces based on images from a digital camera

Read this in Polish

The purpose of the project is to classify in real time the marked parking places on video as free or occupied. This is the fork of my diploma thesis. This fork add functionality to preview in realtime stream from camera and show algorithm classification result. This project is presentation for event WULS-SGGW days at Faculty of Applied Informatics and Mathematics stand.

Thesis summary

Subject of this diploma thesis is the algorithm to recognize free parking spaces based on images from a digital camera. Nowadays information about free parking spaces within specific area is an obvious advantage. The number of cameras in the urban areas is still increasing as well as the capabilities of computers and software. The presented algorithm uses elements of both, the computer image analysis and some artificial intelligence tools. The algorithm consists of several important steps. Those are initial image processing, colour transformation, parameters calculation and decision making (based on previous data knowledge) using the SVM method. A special computer program, which implements the proposed algorithm, was developed and series of tests were carried out. It turns out that in most test cases, the algorithm works very well.