An IoT edge-fog-cloud architecture for vision based planogram integrity

A planogram is a detailed visual map of the products in the shelves of a retail store. The planogram wants to provide the best location of products on the shelves, with the aim to improve the customer experience and satisfaction, to increase sales and profits and to better manage the products on the shelves. However, an important aspect is not only to design the best planogram, but mostly to maintain the right position, orientation and, quantity of the products in the shelves according to the accepted planogram. For this purpose, we propose a fog computing architecture consisted of edge nodes which are low-cost cameras able to transmit in wireless mode shelf photos to local fog nodes in the same store. These last examine the images coming from the edges and sends results to the cloud for further data aggregation and analysis. The experimental results derive from a real environment considering a two-month observation period.