Home>Posts>News>TrashVision – Using Deep Learning and Artificial Intelligence to Quantify Trash
IRVINE, CALIFORNIA – January 15, 2022

TrashVision™ — Using Deep Learning and Artificial Intelligence to Quantify Trash

New science provides techniques for collecting roadside trash data

Microplastics in the water. The Great Pacific Trash Island. The dangerous — and horrifying — consequences of trash in our water are clear. Cities across the country are seeing new environmental regulatory requirements demanding they reduce the trash washing into waterways during storms. California, a leader and bellwether for environmental regulation, is going a step further with its ‘Trash Amendments,’ requiring cities to quantify their efforts in trash observation and reduction strategies. Collecting roadside trash data is instrumental to creating a data-driven plan of action and measuring the effectiveness of those efforts.

Fortunately, actionable new science from Fuscoe Engineering and 2NDNATURE Software provides techniques for doing exactly that. In their recent peer-reviewed paper, Fuscoe and 2NDNATURE introduce the science behind TrashVision™, an innovative AI-driven approach to measuring roadside trash. Vehicle-mounted cameras combined with the power of artificial intelligence give results 57 times faster and twice as accurate compared to walking visual inspection methods, and at much lower cost. This data is the foundation for all analyses, planning, and trash-cleanup programs. This is how cities will comply with new regulations, and how we’ll keep the trash out of the water, creating a more sustainable and resilient world.

Read the Full Paper

The Fuscoe Engineering team with car-mounted camera. Left to right: Stephanie Castle Zinn, Taylor Hanson, Howard Wen

For additional information please contact:

Stephanie Castle Zinn
Project Manager – Water Resources
scastlezinn@fuscoe.com
949.474.1960

Back to News