The Power of Aerial Images and AI: Unveiling Forest Biodiversity
A Hidden Biodiversity Haven
Aspen trees, or Populus tremula, are like nature's secret superheroes. Despite making up only a tiny fraction of Finnish forests, they support an astonishing array of life. Over a thousand species, from mammals to mosses, call these trees home. The unique characteristics of aspens, such as their rapidly decomposing litter and rich bark, create an ideal habitat for many species to thrive.
Unveiling the Secrets of Aspen Distribution
But here's where it gets controversial: accurately mapping the distribution of these keystone species has been a challenging and costly endeavor. That is, until now. A research team, led by Associate Professor Samuli Junttila from the University of Helsinki, has developed a groundbreaking method to identify aspens using open aerial image data. This method offers an affordable, scalable, and reliable solution to monitor forest biodiversity.
"Our research provides an innovative approach to forest conservation and management," Junttila explains. "By combining remote sensing and AI, we can obtain accurate information about the forest ecosystem."
The Global Ecosystem Health Observatory research group, led by Junttila, has been at the forefront of developing these cutting-edge methods. Their previous work has aided in monitoring forest health, from tracking the impact of bark beetles to identifying tree mortality.
A Reliable Neural Network Model
The lead author of this groundbreaking research, Anwarul Chowdhury, a Doctoral Researcher from the University of Eastern Finland, is thrilled with the results. The neural network model they created can accurately identify aspens, even without leaves, across all seasons.
"Our methods produce reliable data that can be applied to forest management and conservation efforts across Finland," Chowdhury emphasizes. The model is particularly effective at identifying fully grown, taller aspens, which are crucial for biodiversity. On average, these trees were identified with a 71% probability, and even higher accuracy when the trees were leafless.
Improving Identification of Juvenile Aspens
And this is the part most people miss: even dead trees play a vital role in biodiversity. Another study by researchers from Junttila's group focused on developing a more accurate technique for identifying standing dead trees from aerial images. Dead trees provide habitats for many specialized and threatened species, but their identification from aerial images has been challenging, especially under dense canopy structures.
By combining machine-learning algorithms and adaptive filters, the research group achieved better results in surveying dead trees compared to traditional forest remote sensing models. "Aspens and standing dead trees are key indicators of biodiversity. Automating their mapping from open datasets is a significant advancement in monitoring forest biodiversity," Junttila concludes.
The future of forest conservation looks bright, thanks to these innovative methods. The research group aims to further improve the accuracy of identifying juvenile aspens and combine open laser scanning data with aerial images to enhance the model's capabilities.
"The potential for these methods is immense," Chowdhury adds. "We're excited to continue pushing the boundaries of what's possible with AI and remote sensing in forest conservation."
References:
Rahman AU, Heinaro E, Ahishali M, Junttila S. Dual-task learning for dead tree detection and segmentation with hybrid self-attention U-Nets in aerial imagery. Int J Appl Earth Obs Geoinf. 2025;144:104851. doi:10.1016/j.jag.2025.104851
Chowdhury AI, Heinaro E, Tanhuanpää T, Rahman AU, Junttila S. Mapping large European aspens (Populus tremula L.) using national aerial imagery and a U-Net convolutional neural network. Remote Sens Appl Soc Environ. 2025;40:101755. doi:10.1016/j.rsase.2025.101755
