SPICE-Based SAR Tomography over Forest Areas Using a Small Number of P-Band Airborne F-SAR Images Characterized by Non-Uniformly Distributed Baselines

Abstract

Synthetic aperture radar tomography (TomoSAR) has been proven to be a useful way to reconstruct vertical structure over forest areas with P-band images, on account of its three-dimensional imaging ability. In the case of a small number of non-uniformly distributed acquisitions, compressive sensing (CS) is generally adopted in TomoSAR. However, the performance of CS depends on the selected hyperparameter, which is closely related to the noise of a pixel. In this paper, to overcome this limitation, we propose a sparse iterative covariance-based estimation (SPICE) approach based on the wavelet and orthogonal sparse basis (W&O-SPICE) for application over forest areas. SPICE is a sparse spectral estimation method that achieves a high vertical resolution, and takes account of the noise adaptively for each resolution cell. Thus, it does not require the user to select a hyperparameter. Furthermore, the used sparse basis not only ensures the sparsity of the forest canopy scattering contribution, but it can also keep the original sparse information of the ground contribution. The proposed method was tested in simulated experiments and the results demonstrated that W&O-SPICE can successfully reconstruct the vertical structure of a forest. Moreover, three P-band fully polarimetric airborne SAR images with non-uniformly distributed baselines were applied to reconstruct the vertical structure of a tropical forest in Mabounie, Gabon. The underlying topography and forest height were estimated, and the root-mean-square errors (RMSEs) were 6.40 m and 4.50 m with respect to the LiDAR digital terrain model (DTM) and canopy height model (CHM), respectively. In addition, W&O-SPICE showed a better performance than W&O-CS, beamforming, Capon, and the iterative adaptive approach (IAA).

UAV Laser Scans Allow Detection of Morphological Changes in Tree Canopy

“…This greatly limited the possibility of not detecting the actual tree because individual tree canopies were not often connecting, as is mentioned by [27]. The limitations of the IWS segmentation from ULS data in close canopy forest were tested and compared with several other methods by [15]. The study of [26] mentioned reaching high detection potential of individual standing trees.…”

“…In both highly-cited studies [10,11], the TLS method was used for data acquisition, and both author collectives concluded that the method was time-demanding. In contrast to TLS, the ULS data acquisition provides lower point cloud density matched with high time efficiency [12][13][14][15]. A yearlong investigation of the branch inclination angles using TLS was not performed on forest trees.…”

“…The photogrammetry RS methods are based on a developed scale-invariant feature transform algorithm [28]. One study [15] described and evaluated several methods for tree detection, including IWS, individual tree segmentation (ITS), and point cloud segmentation (PCS), and ULS has also been used for individual tree detection [25]. The advantage of low cost and feasible multi-temporal data acquisition for the derivation of DTM and individual tree change detection based on CHM will also be used for estimation of carbon stock, describing the carbon dynamics [29], and for describing the forest vertical structure [30] from the International Space Station (ISS) module, NASA's Global Ecosystem Dynamics Investigation (GEDI) data.…”

“…In contrast to our study, in which we used three scans throughout the year, Barak et al [51] only used two time-distanced scans per tree; thus, we could determine if the branches returned to their "initial" position. The TLS method is, for this purpose, time-consuming, in contrast to the ULS method, which is adequate for the derivation of representative individual tree point clouds for many common forest inventory values, and its time efficiency compared to TLS methods makes it more practical for surveys across larger areas [12,13,15].…”

Individual Tree-Crown Detection and Species Classification in Very High-Resolution Remote Sensing Imagery Using a Deep Learning Ensemble Model

“…The influence of canopy cover on tree detection accuracy has not been evaluated. In a recent study [53], different tree-crown detection algorithms were assessed under various canopy cover conditions and a reduction in accuracy directly related to an increase in canopy cover rate was reported. However, the authors report a high dependency on spatial resolution of the tree detection methods and did not employ deep learning techniques.…”