Using machine learning to identify and contextualize residential rainwater downspouts and guide stormwater outreach programming

Municipalities are seeking ways to reduce the effects of uncontrolled rainwater and one of the primary contributors to the problem is residential downspouts. Downspout disconnections and redirections are low-cost ways to reduce the overall imperviousness of urban and suburban landscapes and ameliorate stormwater issues. This project will use machine learning to analyze Google street view data to find all visible downspouts and contextualize the areas where they drain as permeable or impermeable. The analysis is relevant to almost every municipality in Canada, because stormwater management is a responsibility of the municipal level of government under the Municipal Act of Ontario (2001) and other legislation elsewhere in Canada.

Faculty Supervisor:

Jeffrey Wilson

Student:

Lukas Weber

Partner:

Synergy Sustainability Development Group

Discipline:

Engineering

Sector:

Other

University:

University of Waterloo