TRACKING THE URBAN CHAMELEON – TOWARDS A HYBRID CHANGE DETECTION OF GRAFFITI

Abstract. Colourful and ever-changing: Graffiti can be considered the urban chameleon skin. At Donaukanal (Eng. Danube Channel), Vienna's central waterway one of largest most active graffiti-scapes worldwide, this metaphor applies like hardly anywhere else. Every day a multitude graffiti is destroyed by creation new works. Recently, efforts have been made to mitigate constant loss cultural heritage along systematically documenting graffiti, mainly using photography photogrammetry. However, keeping track newly added works very time-consuming often finding needles in haystack, considering large extent high volatility monitored area. Thus, an automated change detection would significantly reduce effort avoid overlooking graffiti.This contribution outlines main challenges image-based for proposes hybrid method. The investigated method exploits combines established pixel-based algorithm, Iteratively Multivariate Alteration Detection, with novel descriptor-based latter relies on image features, rather than pixels as analysis unit robustly filter false alarms from high-performing but noise-prone approach. Overall, results indicate that proposed largely automate graffiti-scapes. It uncover graffiti-related changes distinguish them other differences such shadows tends overlook small-scale indicating need further fine-tuning.