Abstract
The spectral and spatial resolutions of modern optical Earth observation data are continuously increasing. To fully utilize the data, integrate them with other information sources, and create applications relevant to real-world problems, extensive training data are required. We present TAIGA, an open dataset including continuous and categorical forestry data, accompanied by airborne hyperspectral imagery with a pixel size of 0.7 m. The dataset contains over 70 million labeled pixels belonging to more than 600 forest stands. To establish a baseline on TAIGA dataset for multitask learning, we trained and validated a convolutional neural network to simultaneously retrieve 13 forest variables. Due to the size of the imagery, the training and testing sets were independent, with strictly no overlap for patches up to <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$45\times 45$ </tex-math></inline-formula> pixels. Our retrieval results show that including both spectral and textural information improves the accuracy of mapping key boreal forest structural characteristics, compared with an earlier study including only spectral information from the same image. TAIGA responds to the increased availability of hyperspectral and very high resolution imagery, and includes the forestry variables relevant for forestry and environmental applications. We propose the dataset as a new benchmark for spatial–spectral methods that overcomes the limitations of widely used small-scale hyperspectral datasets.
Highlights
Satellite-based remote sensing is a widely accepted tool in forest inventory and for monitoring the extensive forests covering approximately one-third of the world’s land surface
We used the basic statistics commonly used in forest remote sensing studies, such as the root-mean-square error (RMSE), relative RMSE, relative mean bias, and coefficient of determination (R2)
We suggest that combining VHR and HyperSpectral Imaging (HSI) data from different sensors with advanced machine learning approaches may considerably improve the robustness of forestry data, especially for regions where national forest inventory data are not available
Summary
Satellite-based remote sensing is a widely accepted tool in forest inventory and for monitoring the extensive forests covering approximately one-third of the world’s land surface. This essential service is largely based on high- to mediumresolution multispectral optical instruments with pixel sizes of 10 m or more. While the availability of imagery with better spatial resolution, one meter or less (called Very High Resolution, VHR) is increasing rapidly, its utility in forestry applications has not been demonstrated yet [1]. VHR images are useful in visual analyses, but lack large-scale automated processing chains required for routine forest applications. A natural approach to automate VHR image processing is deep learning by making use its breakthroughs in computer. Box 15400, FI-00076 Aalto, Finland, and is with Purdue University, USA
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
