Video Observation And Automated Analysis To Evaluate Physical Activity At A Rural Linear Park

The environments in which people live influence their health behaviors, so researchers and policy makers are targeting the built environment to promote physical activity within communities. PURPOSE: The purpose of this study was to evaluate physical activity behaviors at a rural linear park and to further validate the use of an innovative video observation and automated data analysis method. METHODS: High-definition video footage (720p at 30 frames per second) was obtained at five sites along a linear park in southern United States between March 15, 2022 and March 26, 2022. Trail cameras were placed along the pathway between trail heads. A total of 9,613, 30-second video clips were analyzed using A.I. Whoo PARKVision software. The software uses proprietary machine learning techniques (i.e. neural rendering) which extract and analyze human skeletal positioning for physical activity behaviors. For this study, we analyzed usage (people counts), mode of use (frequency), total energy expenditure (METs), and peak usage. RESULTS: A total of 9,613, 30-second video clips were obtained (80.1 video hours) across the 11-day study, capturing 2,275 people counts, with walking (83%) as the primary form of physical activity observed. In total, we observed 1,455 METs used across all participants captured. Results from this study indicated three specific trends: (1) community members were more active during weekday evenings, (2) there was an increase in weekend PA behavior compared to the weekday, and (3) walking was the most commonly performed type of PA (compared to running and biking). CONCLUSION: The video observation and automated analysis provided a feasible, valid, and cost-efficient method to evaluate usage of a rural linear park. Park users primarily engaged in low-to-moderate intensity physical activity, and followed usage trends similar to other linear parks. Affordable, easy-to-use technology can serve a key role in evaluating the public health impact of rural linear parks. Supported by an internal institutional seed grant.