The WISC-XyloTron: Combatting the illegal logging trade, one image at a time

Dr. Alex Wiedenhoeft (left) and Dr. Prabu Ravindran (right) in the Forest Products Laboratory

Each year, the Baldwin Wisconsin Idea Endowment offers competitive grants to projects which embrace the values of and seek to advance The Wisconsin Idea. Among the recipients for this year were Dr. Prabu Ravindran, Assistant Scientist in the UW-Madison Department of Botany, and Dr. Alex Wiedenhoeft, Research Botanist and Team Leader at the Wood Anatomy Research Center of the Forest Products Laboratory and Adjunct Associate Professor in the UW-Madison Department of Botany, for their work on “Wood Identification and Screening for Timbers of Mozambique.” Last week, the African Studies Program sat down with Dr. Ravindran and Dr. Wiedenhoeft to discuss their project – how it came about, what has been accomplished thus far, and what they anticipate moving forward.

The impetus behind this particular project is the scourge of the illegal logging trade, which depresses the market value of United states forest products by more than one billion dollars each year, costs the global market some 130 billion dollars or more per year, and contributes to deforestation and environmental degradation all over the world. Not only this, but the profits from illegal logging regularly support organized crime and other forms of transnational trafficking, including wildlife, illegal drugs, weapons, and even human beings.

This is a widespread, profitable, and booming industry, and professionals guarding international borders are often nearly powerless to stop it. After all, the ability to identify trafficked wood requires a very specific form of expertise or the use of high-tech, expensive equipment, not to mention that it would be difficult to justify the incredible time and profit loss incurred from attempting to separate legal and illegal wood trade on the fly.

Dr. Ravindran and Dr. Wiedenhoeft hope to change all of that. They have developed the machine learning based Wood Identification and Screening Xylotron (WISC-XyloTron) system, a combination of an open-source imaging device and image-based wood species classifier. Development of the WISC-XyloTron initially began in 2010 at the Forest Products Laboratory, largely in Dr. Wiedenhoeft’s laboratory. He currently oversees the wood sample preparation, data curation, field deployment, and forensic wood anatomy aspects of this project. Dr. Ravindran, an expert in machine learning for wood forensics, contributes the computer vision and machine learning aspects of the project.

The WISC-XyloTron operates under a machine learning model, or one in which a computing system ‘learns’ to make certain decisions based on identified patterns in a wealth of user-provided data. In this particular case, the data provided to the system are images of the end grain of different wood species, especially those of importance to biodiversity or those at particular risk of being trafficked. The WISC-XyloTron had a lot to learn; after all, the Forest Product Laboratory here in Madison is the national research laboratory for the United States Forest Service and is home to the largest collection of wood samples in the world.

When paired with a laptop, the WISC-Xylotron can be used to identify the end grain of a wood in just a few seconds with a high degree of accuracy. When used at an international border, this technology would be able to identify those woods which are at a higher risk of being trafficked, and therefore which may need to be separated out and studied further. This process will allow legal trade to continue while dramatically limiting the prevalence of illegal wood in trade.

Dr. Ravindran and Dr. Wiedenhoeft have an ongoing project in Ghana, and on the day of our interview had just shipped fourteen XyloTrons to various branches of the national Timber Industry Development Division (TIDD) throughout the country. They will expand their Africa-based collaborations to Mozambique through funds provided by the Baldwin Wisconsin Idea Endowment. By partnering with Dr. Ernesto Uetimane, a wood anatomist from Eduardo Mondlane University in Maptuo, they are using this Baldwin project to extend the reach of work done in Madison to a new place where it has the potential to make a meaningful contribution. Dr. Ravindran and Dr. Wiedenhoeft are planning for a future where they can partner with colleagues to do the same for countries across the globe in years to come.

International partners are absolutely essential to this entire operation, of course; in the area of research and data collection and in ensuring funding, implementation, and regional spread of not only the XyloTron technology but also legal, political, and economic frameworks which can actively combat illegal logging using this technology. All of the technology is open source, and Dr. Wiedenhoeft and Dr. Ravindran have brought numerous academic and government officials to Madison in order to train them in best practices for using the technology. All of these decisions were made with the intention of enabling domestic capacity in not only the use of this technology but, in the future, the adjustment of said technology in making it a better fit to the needs of individual countries and communities.

We will circle back once again to the words of Dr. Dawd Siraj, Professor of Medicine in the Division of Infectious Diseases at UW-Madison, from last week: “We can’t just stay in our part of the world… our research needs to go beyond our limited horizon. There are many lessons that we will only be able to learn in different places.” Here, he echoes the WISC-Xylotron project, the Baldwin Grant, and the Wisconsin Idea, in arguing for what is possible when one pushes past the boundaries of academia and applies the knowledge gained there to real-world problems, obstacles, and questions.

Written by Rebecca Hanks.