XPRIZE Rainforest

  • Since 2022 I have been developing computer vision tools for Limelight: Rainforest, the first place winners of the XPRIZE rainforest competition. XPRIZE Rainforest was a 10 million dollar competition to develop automated technology capable of rapidly surveying biodiversity in tropical forests, concluded in November of 2024.
  • An automated blacklight trap sampling insects in the canopy of the Brazilian Amazon
      An automated blacklight trap sampling insects in the canopy of the Brazilian Amazon
    The overarching goal of XPRIZE Rainforest was to develop technology to allow rapid, automated surveys of biodiversity in highly complex tropical systems. Our solution is a canopy raft that takes data from a large number of separate streams, including image, audio, and DNA data, with a focus on insects. We also seek to avoid the colonialism tacitly present in many efforts to survey biodiversity by directly involving local indigenous communities in data collection and by incorporating indigenous knowledge of rainforest systems into our measures of rainforest communities.
    In June of 2023 Limelight: Rainforest competed in the XPRIZE Rainforest semifinals competition in Singapore. In a 24 hour period we were able to photograph more than 17,000 insects from 47 different families. With the addition of acoustic monitoring, DNA barcoding, and eDNA, we were able to identify 21 named species, 46 unique genera, 62 unique families, and 21 orders of plants and animals from the sampling area. Limelight: Rainforest was one of 6 finalist teams selected from the 13 competitors.

    In July of 2024 Limelight: Rainforest took first place in the XPRIZE Rainforest competition by surveying in excess of 250,000 insects, 50,000 bird and bat calls, and 20,000 plants in 24 hours in the Brazilian Amazon, totaling more than 600 unique taxa.

    Although the competition is over, we are continuing to research and deploy systems for rapid biodiversity surveys in the tropiccs. Our current work focuses on developing survey strategies for using camera traps to rapidly detect differences in bidoiversity between habitats, and changes in biodiversity over time.

    Read more