Experiences with Video-Based Parking Space Surveillance
Klaus-Peter Döge, Jürgen Krimmling · 19th ITS World CongressERTICO - ITS EuropeEuropean CommissionITS AmericaITS Asia-Pacific · 2012
In general, diverse and sometimes severe environmental conditions are still a challenge for video-based traffic analysis. Additionally, most commercial image processing software depends on a particular camera model, and it must be calibrated by a professional. The research goal of the presented work was to resolve these problems with a novel in-house development. Due to the lack of motion information and light produced by the vehicles at night, video based parking space surveillance is a special challenge. The method developed does not disclose vehicle-specific characteristics; only stochastic and non-specific characteristics are used for image processing. The results showed a good robustness against weather and lighting conditions, while the measurement failure is sufficient to the geometric distribution.