An automatic analysis of shoppers behaviour using a distributed RGB-D cameras system

The aim of this work is to propose an integrated system consisted of an RGB-D camera and software able to monitor shoppers in intelligent retail environments. We want to propose an innovative low cost intelligent system that can evaluate not only the shopper behaviour, but also detect their interactions with the products in the shelves, by developing automatic RGB-D techniques for video analysis. The system of cameras, located in strategic locations within the store, detects the presence of a person by identifying the blob and the centre of mass. The camera detects the person as an object in moving. Through the video frames, the system detects the interactions of the shoppers with the products on the shelf and establishes also the type of interaction: if a product is picked up, if the product is taken and then repositioned and finally, if there is not contact with the products. To understand the shopper behaviour is very important for the marketing strategies of a retail store. The proposed architecture monitors this aspect, that is low cost, easy to install and able to ensure very satisfactory results also in real environments.