References

Publications

Triaging Deforestation Alerts – Clustering alerts for review 

Goodman, Chris. Triaging Deforestation Alerts – Clustering alerts for reviewIEEE Global Humanitarian Technology Conference (GHTC), Seattle, October 2016.

 

 

 

© 2016 IEEE

http://ieeexplore.ieee.org/document/7857264/

Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.

Abstract

Advances in the automatic detection of deforestation and forest disturbance now provides pan-tropical data down to 30x30m pixels on an 8 to 16-day update schedule. Global Land Analysis & Discovery (GLAD) Alerts, available from Global Forest Watch provide valuable new insights into near real-time forest canopy change. This introduces a new challenge for agencies to deal with the large volume of alerts. A technique is presented to triage deforestation alerts. Initial processing uses clustering algorithms from data mining and pattern recognition applications to group the alerts. By combining many adjacent and nearby disturbances into clusters leaving fewer cases to investigate. The cases are forwarded to a crowd of volunteers for a quick visual check before being sent to locals. Both the original alert coordinates and their cluster perimeter may be encoded in the same GeoJson feature collection. This aids both visualization and coordinating and recording the response. A prototype is demonstrated based on Google Earth Engine and real-time data from Global Forest Watch.

 

 

 

 

Note: At the time of publication Earth Engine did not have support for clustering algorithms. This paper describes a technique that used vector-to-raster conversion to achieve clustering indirectly. Earth Engine now directly supports clustering algorithms including DBSCAN. Using DBSCAN would simplify the computation.

Bunjil  A social network for proactive monitoring of tropical rainforests

Goodman, Chris (2010). Bunjil  A social network for proactive monitoring of tropical rainforests.Telecommunications Journal of Australia. 60 (1): pp. 4.1 to 4.16.

Bunjil Forest Watch – a Community-Based Forest Monitoring Service, Deforestation Around the World

Goodman, Chris (2012). Bunjil Forest Watch a Community-Based Forest Monitoring Service, Deforestation Around the World, Dr. Paulo Moutinho (Ed.), ISBN: 978-953-51-0417-9,
InTech, DOI: 10.5772/34808.
See also: Presentations

Presentations

Clustering Deforestation Alerts

Notes for Presentation to IEEE Global Humanitarian Conference, Oct 14th 2016.
Earth Engine Clustering code – requires Earth Engine access to view.

Crowd Classification of Clusters

Prototype Cluster Classification User Interface – (Swinburne Project)

How to create a globally networked forest monitoring tool without a budget

Notes for Presentation to Land and Poverty Conference, World Bank, Washington D.C. March 18th, 2016