Identification of Forest Ground and Canopy Peaks From 3-D SAR Tomographic Profile Using Deep Learning
- 16 September 2025
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 63 (01962892) , 1-13
Abstract
Tomographic SAR (TomoSAR) at low frequency, i.e., P/L band, has become a promising tool for forest structure study. Forest canopy height and underlying topography are two of the most important parameters one can estimate using TomoSAR technique. One simple way to estimate these two parameters is to detecting the peaks of the tomographic profile, which, however, can lead to large biases due to complicated forest structure, sidelobes or insufficient TomoSAR resolution. Polarimetric TomoSAR (Pol-TomoSAR) provides a solution to this by exploring the polarimetric diversity to separate the ground and canopy components and then conduct independent TomoSAR analysis. However, Pol-TomoSAR technique suffers from low ground-to-volume ratio (GVR), which often leads to unsuccessful ground and canopy separation. To mitigate this propblem, in this paper, we provide a deep-learning solution to ground and canopy height estimation from 3D tomographic profile through the identification of the patterns of ground and canopy peaks. A 3D U-net model is introduced in our solution to grasp as much three-dimensional characteristics of the tomographic profile as possible. Moreover, our model can be well trained using only synthetic TomoSAR dataset, making it easy to implement when we don’t have enough real data with LiDAR references. The proposed method is validated on P-band real TomoSAR dataset from multiple test sites in AfriSAR campaign, showing that it can achieve more accurate ground and canopy height estimation than the state-of-the-art Pol-TomoSAR techniques. The maximum RMSE improvement reaches as high as 66.4% and 63.2% for ground and canopy top height, respectively.Keywords
Funding Information
- National Natural Science Foundation of China (U2241202)
- Academic Excellence Foundation of Beihang University (BUAA) for Ph.D. Students, in part by China Association for Science and Technology’s Young Talents Project for Ph.D. Students
- China Scholarship Council
- CNES/TOSCA
- DGA-AID (DOCTOSAR-3DFOR 2022-65-0082)
This publication has 24 references indexed in Scilit:
- MLC-net: A sparse reconstruction network for TomoSAR imaging based on multi-label classification neural networkISPRS Journal of Photogrammetry and Remote Sensing, 2024
- A Deep Learning Solution for Height Inversion on Forested Areas Using Single and Dual Polarimetric TomoSARIEEE Geoscience and Remote Sensing Letters, 2023
- A Deep-Learning Approach for SAR Tomographic Imaging of Forested AreasIEEE Geoscience and Remote Sensing Letters, 2023
- Canopy height and structure measurements of Gabon from medium-footprint waveform lidar2017
- SAR imaging of tropical African forests with P-band multibaseline acquisitions: Results from the AfriSAR campaign2016
- 3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation2016
- SAR tomography for the retrieval of forest biomass and height: Cross-validation at two tropical forest sites in French GuianaRemote Sensing of Environment, 2016
- U-Net: Convolutional Networks for Biomedical Image Segmentation2015
- Deep learningNature, 2015
- First demonstration of airborne SAR tomography using multibaseline L-band dataIEEE Transactions on Geoscience and Remote Sensing, 2000