Machine Vision Protects Tree Biodiversity in Peru, Madagascar, and Indonesia

Illegal logging ranks third in transnational crime and costs billions of dollars in lost revenue annually. Increasing global demand for precious timber—used to produce items such as wind turbine blades and flooring—has led to a surge of illegal exploitation. The resulting deforestation accelerates climate change since the process releases carbon sequestered in trees into the atmosphere and supplies land for growing commodities that create more CO2.

Despite existing trade bans and international agreements, such as the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES), illegal harvesting continues due to high lumber demand and the difficulties in law enforcement and identification of stolen goods. Wood identification is one of the ways to combat illegal logging, but it requires officials to be familiar with the diverse characteristics of tree species—a challenge when wood products often look similar and no longer have readily identifiable parts such as leaves, flowers, and fruits.

PEER PI Dr. Bako Harisoa Ravaomanalina in her lab at the University of Antananarivo Madagascar. Credit: Dr. Harisoa Ravaomanalina

PEER PI Dr. Bako Harisoa Ravaomanalina in her lab at the University of Antananarivo Madagascar.
Credit: Dr. Harisoa Ravaomanalina

To address these challenges, three researchers in the USAID-supported Partnerships for Enhanced Engagement in Research (PEER) in PeruMadagascar, and Indonesia have each partnered with scientists from the U.S. Forest Service to develop tools to assist private and public organizations in wood species identification and help curb illegal logging. The Xylotron, a portable wood identification tool with machine-vision technology, is a U.S. Forest Service invention that is being piloted around the world. Machine vision is the combination of optics, electrical, and software engineering that uses light captured by a sensor to inform decisions. The three research teams broaden the Xylotron’s database by cataloging thousands of samples and images of native timber species in their respective countries. With machine vision, the Xylotron processes and catalogs these images until it recognizes the species independently. By equipping the Xylotron with data on important species and training monitoring officials on how to use it, researchers can increase local capacity to reduce lumber trafficking, leading to healthier ecosystems and safer communities. 

The USAID-funded sample sander in Lima, Peru.
The USAID-funded sample sander in Lima, Peru.
Credit: Melissa Trimble

Meet the PEER teams addressing illegal logging

Wood sample collection at the Indonesian Ministry of Environment and Forestry.

Peru: Dr. José Ugarte Oliva and his team at the Instituto Tecnológico de la Producción - CITEmadera worked with government authorities to select 15 species to contribute to the Xylotron database. They took 65,762 images of wood samples, collected from 103 trees in three regions that they sanded down individually. Once the Xylotron could identify the samples, the team trained government staff in charge of regulating wood trade, created user guides, shared their wood samples with other universities and institutions, and conducted a virtual training of the Xylotron system for researchers in Costa Rica.

Madagascar: Dr. Bako Harisoa Ravaomanalina from the University of Antananarivo collected wood samples in eight regions of Madagascar, resulting in a library of 4,500 wood specimens at its inauguration in 2020. The wood library serves as a reference for highly sought-after Malagasy rosewood, palissanders, and ebony species, increasing the identification accuracy of CITES-listed species for scientists, private companies, and government authorities. 

Indonesia: Dr. Ratih Damayanti’s team at Indonesia’s Forestry Research Development and Innovation Agency developed and launched the automatic wood identification system AIKO-KLHK Version 2, in both Bahasa Indonesian and English. It contains 1,180 wood species and 4,054 users, plus six wood industry organizations. The team filed for a patent for their WIDER (Wood Species Identification System Based on Resistance, Capacitance, and Inductance Properties) prototype — a new identification tool that Dr. Damayanti continues to develop using new funding she received as a result of her PEER work. 

With these portable, machine vision-driven tools, these PEER researchers are creating digital collections that can map the distribution of wood biodiversity in Peru, Madagascar, and Indonesia—empowering law enforcement agencies to identify timber samples in the field and reducing the ability of traffickers to destroy forests and harm ecosystems through illegal logging.