Automated people-counting by using low-resolution infrared and visual cameras

Abstract

Non-contact counting of people in a specified area has many applications for safety, security and commercial purposes. Visible sensors have inherent limitations for this task, being sensitive to variations in ambient lighting and colours in the scene. Infrared imaging can overcome many of these problems, but normally hardware costs are prohibitively expensive. A system for counting people in a scene using a combination of low cost, low-resolution visual and infrared cameras is presented in this paper. The aim of this research was to assess the potential accuracy and robustness of systems using low-resolution images. This approach results in considerable savings on hardware costs, enabling the development of systems which may be implemented in a wide range of applications. The results of 18 experiments show that the system can be accurate to within 3% over a wide range of lighting conditions.