H43F-1024:
Remotely Measuring Trash Fluxes in the Flood Canals of Megacities with Time Lapse Cameras and Computer Vision Algorithms – a Case Study from Jakarta, Indonesia.

Thursday, 18 December 2014
Frank Sedlar1, Etienne Turpin2 and Branko Kerkez1, (1)University of Michigan Ann Arbor, Ann Arbor, MI, United States, (2)University of Wollongong, Wollongong, Australia
Abstract:
As megacities around the world continue to develop at breakneck speeds, future development, investment, and social wellbeing are threatened by a number of environmental and social factors. Chief among these is frequent, persistent, and unpredictable urban flooding. Jakarta, Indonesia with a population of 28 million, is a prime example of a city plagued by such flooding. Yet although Jakarta has ample hydraulic infrastructure already in place with more being constructed, the increasingly severity of the flooding it experiences is not from a lack of hydraulic infrastructure but rather a failure of existing infrastructure. As was demonstrated during the most recent floods in Jakarta, the infrastructure failure is often the result of excessive amounts of trash in the flood canals. This trash clogs pumps and reduces the overall system capacity. Despite this critical weakness of flood control in Jakarta, no data exists on the overall amount of trash in the flood canals, much less on how it varies temporally and spatially. The recent availability of low cost photography provides a means to obtain such data. Time lapse photography postprocessed with computer vision algorithms yields a low cost, remote, and automatic solution to measuring the trash fluxes. When combined with the measurement of key hydrological parameters, a thorough understanding of the relationship between trash fluxes and the hydrology of massive urban areas becomes possible. This work examines algorithm development, quantifying trash parameters, and hydrological measurements followed by data assimilation into existing hydraulic and hydrological models of Jakarta. The insights afforded from such an approach allows for more efficient operating of hydraulic infrastructure, knowledge of when and where critical levels of trash originate from, and the opportunity for community outreach – which is ultimately needed to reduce the trash in the flood canals of Jakarta and megacities around the world.