Presentation at the 2004 Active Living Research Annual Conference
To develop an automated system for monitoring use of indoor and outdoor facilities for physical activity and to assess the validity and reliability of the system.
a) Develop, field test, and refine an automated monitoring system based on a complementary metal oxide semiconductor (CMOS) image acquisition sensor, that addresses the weaknesses of existing automated systems;
b) Assess reliability and validity of the automated system;
c) Provide documentation of all hardware and software involved, including user instructions, to be placed in the public domain.
Existing technology for automated monitoring of people engaged in various forms of physical activity include the inductive loop, infrared beam, seismic sensor, and videocamera but none have proven completely satisfactory. These methods have both strengths and weaknesses. Notable weaknesses include the inability to recognize and count people walking in groups; the inability to recognize and count individuals (as opposed to gross counts of passages past the monitored point); privacy concerns for videotaping; inaccuracies due to extraneous counts; vandalism and theft; calibration and maintenance concerns; and sensitivity/malfunction because of varying environmental conditions.
RESEARCH DESIGN AND METHODS:
We propose the development and testing of an automated counter system based on a complementary metal oxide semiconductor (CMOS) image acquisition sensor, which has not been applied previously to physical activity measurement through automated systems. The optimal system hardware will include one or more microprocessor boards, one CMOS image acquisition sensor module with an industry-standard lens mount, a standard lens assembly, and a battery, all contained in a single industry-standard custom-modified tamper- and weather-resistant enclosure. The absolute minimum of custom-designed hardware necessary to the investigation will be developed and documented to provide the requisite interconnection, communications, and power conditioning/control functionality; thus, while this system is "new", its components are commercially available and all documentation needed to replicate the system at the end of the study will become public
This system will produce user counts (of individuals), user passage counts (per-user counts of passages past the monitored site), and user passage volume (total counts of all passages by all users past the monitored site during the monitored period). The system will also record the velocity of movement of the users and will have the capability to recognize and count multiple types of travel (pedestrian, skateboard, bicycle, horseback, skates). It is expected to show good validity compared to human observers and good inter-instrument reliability.
Field testing under varying environmental conditions will confirm the system's robustness. The final product will be a valid and reliable system that can be built from off-the-shelf products using instructions in the public domain or purchased ready-made from the systems engineer for this project.