One criteria for obtaining funding and resources for park trails is being able to determine the traffic coming through a particular area during any given period of time. In general, park usage has been determined by the number of vehicles that purchase a day use pass at a park entrance. However, local residents tend to ride/hike from home and enter the park through alternative entrances. There is a need to determine trail usage so that the proper allocation of resources can be committed to remedying issues on those trails and for future trail planning.

At productOps we have some avid mountain bikers on our staff of developers and Santa Cruz, our head office location, is an area renowned for its mountain biking culture and infrastructure. We saw an opportunity to improve on the technology currently being used to solve one of the biggest, most important problems faced by land managers. How to track and report usage of trail networks being built for the community?

Infrared solutions had been tried previously to capture trail usage information but were found lacking in distinguishing different types of traffic: bikers, hikers and horseback riders. We set about building an intelligent system that could not only count passing objects, but define them, using both machine learning and computer vision. Building upon the computer vision on the edge platform provided by our partners alwaysAI, productOps created what we dubbed CROSS COUNT, a compound object tracker which could successfully parse a variety of traffic that was using trails. We then leveraged our team’s data analytics knowledge to create a platform that would not only capture data but help organize, present and interpret it as well.

We now have the knowledge to apply what we’ve learned to other locations such as ski resorts, urban environments, anywhere where the frequency of people or objects passing through an area needs to be determined.