The goal of the Greenum project is to use a camera and image analysis models to study the flow of movement in a given space. The results of this preliminary study are intended to improve the design of the space in question in view of a possible revitalization.

In this first project, the main objective was to explore the different analysis possibilities and to create a functional prototype capable of processing analysis data over a period of approximately one week. First, the installation of the equipment was carried out, which involved configuring a camera with a GSM router and placing the whole thing in a waterproof case for outdoor use. Then, the videos recorded by the camera were analyzed. Finally, in the third part of the project, the data produced was used to generate heat maps representing the different elements detected. It is possible to filter the results when creating these heat maps, based on criteria such as moving objects, object types or average displacement.

In short this project demonstred how to:

• Track public spaces usage through AI in an anonymous way
• Optimize public places (plant trees, add benches etc.) based on real usage
• Leverage existing AI technologies such as Yolo and norfair to perform long duration video analysis