Identification of Unusual Visitor Activity Using a Hybrid Gaussian Model in Picturesque Locations

Abstract

The management of tourists in picturesque locations is becoming increasingly crucial due to the rapid growth of the tourism industry. The subsequent issue, though, is the rise in strange visitor behaviours, like carelessly crossing hazardous locations and doing illicit graffiti, among other things. These strange behaviours harm tourist destinations' reputations while also raising the possibility of safety hazards. Because of this, this paper uses a regional feature analysis approach to detect abnormal pedestrian behaviour. Specifically, we begin by modelling the background, obtain the moving target region information through moving target detection, and use the minimum outer rectangular box as the regional feature. Lastly, we compute the aspect ratio of this rectangle, perform curve fitting and prediction, and finally conclude the abnormal behaviour detection process. The method's simplicity and speed are demonstrated by the experimental findings.