Multi-task fully convolutional network for tree species mapping in dense forests using small training hyperspectral data

TL;DR

This paper introduces a multi-task convolutional neural network that improves tree species mapping accuracy in dense forests using limited hyperspectral UAV data, by combining semantic segmentation with boundary-aware distance regression.

Contribution

It presents a novel multi-task architecture with a partial loss and boundary constraints, achieving state-of-the-art accuracy with scarce training data.

Findings

Boosts semantic segmentation accuracy by up to 11%.

Achieves an average user's accuracy of 88.63%.

Achieves an average producer's accuracy of 88.59%.

Abstract

This work proposes a multi-task fully convolutional architecture for tree species mapping in dense forests from sparse and scarce polygon-level annotations using hyperspectral UAV-borne data. Our model implements a partial loss function that enables dense tree semantic labeling outcomes from non-dense training samples, and a distance regression complementary task that enforces tree crown boundary constraints and substantially improves the model performance. Our multi-task architecture uses a shared backbone network that learns common representations for both tasks and two task-specific decoders, one for the semantic segmentation output and one for the distance map regression. We report that introducing the complementary task boosts the semantic segmentation performance compared to the single-task counterpart in up to 11% reaching an average user's accuracy of 88.63% and an average producer's accuracy of 88.59%, achieving state-of-art performance for tree species classification in tropical forests.