Deep Learning Solutions for Tandem-X-Based Forest Classification

Mazza, A; Sica, Francescopaolo

doi:10.1109/igarss.2019.8900441

Cited by 3 publications

(3 citation statements)

References 16 publications

(27 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the sake of brevity, a more extensive evaluation with respect to the input configuration is presented for the TDX-Res case only, without loss of generality. A few models correspond to the solutions discussed in Mazza and Sica [39].…”

Section: Numerical Assessmentmentioning

confidence: 99%

“…For each architecture, we tested different input combinations, ranging from the single-band SAR image to a 4-band stack that encloses three additional features: the incidence angle, the interferometric coherence, and the volume decorrelation contribution, which carry relevant information on the nature of the illuminated target. Some preliminary results can be found in Mazza and Sica [39].The paper is organized as follows. Section 2 provides a brief summary of the baseline reference method and introduces basic concepts about CNNs.…”

mentioning

confidence: 99%

See 1 more Smart Citation

TanDEM-X Forest Mapping Using Convolutional Neural Networks

et al. 2019

Self Cite

View full text Add to dashboard Cite

In this work, we face the problem of forest mapping from TanDEM-X data by means of Convolutional Neural Networks (CNNs). Our study aims to highlight the relevance of domain-related features for the extraction of the information of interest thanks to their joint nonlinear processing through CNN. In particular, we focus on the main InSAR features as the backscatter, coherence, and volume decorrelation, as well as the acquisition geometry through the local incidence angle. By using different state-of-the-art CNN architectures, our experiments consistently demonstrate the great potential of deep learning in data fusion for information extraction in the context of synthetic aperture radar signal processing and specifically for the task of forest mapping from TanDEM-X images. We compare three state-of-the-art CNN architectures, such as ResNet, DenseNet, and U-Net, obtaining a large performance gain over the baseline approach for all of them, with the U-Net solution being the most effective one.2 of 18 between the master and slave images, enabling high-resolution interferometric measurements with an unprecedented quality. The constellation comprises two twin satellites flying in a bistatic close-orbit configuration, which allows for a flexible selection of the acquisition geometries and, in particular, of the interferometric baselines [8]. The main goal of the mission was the generation of a global consistent high-resolution digital elevation model (DEM) with unprecedented accuracy, which has been successfully completed in 2016 [9]. Besides the nominal DEM product, for each bistatic interferometric TanDEM-X acquisition, additional quantities can be computed as by-pass products. Indeed, the bistatic acquisition is not affected by temporal decorrelation, allowing for an accurate isolation of volume scattering phenomena from the interferometric coherence. This feature was exploited in References [10,11], where the authors presented a framework for the development of a global TanDEM-X forest/non-forest map [12] as described more in details in Section 2.1.Deep learning approaches and, in particular, Convolutional Neural Networks (CNNs) have been massively used in computer vision and image processing in the last few years, since the publication of the breakthrough work of Krizhevsky et al. on image classification in 2012 [13]. Thanks to the CNNs capability to learn very complex nonlinear relationships from huge labeled datasets with the help of commercial GPUs, unprecedented results have been obtained for many typical tasks such as super-resolution [14,15], segmentation [16], denoising [17], object detection [18,19], classification [20][21][22], and many others.Recently, deep learning has started to significantly impact remote sensing applications as well, as testified by the recent survey of Zhu et al. [23]. Established techniques in remote sensing concern, e.g., pansharpening [24,25], vehicle detection [26] with optical images, crop classification [27,28], anomaly detection with hyperspectral data [29], despeckling [30,31],...

show abstract

Section: Numerical Assessmentmentioning

confidence: 99%

mentioning

confidence: 99%

TanDEM-X Forest Mapping Using Convolutional Neural Networks

et al. 2019

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the last years, deep learning base methods are showing impressive results in many application of natural image processing such as classification, segmentation and detection (He et al, 2017). Actually, good results are achieved also in several remote sensing application like land classification and segmentation (Mazza, Sica, 2019), super-resolution (Vitale, 2019) and detection (Gargiulo et al, 2019). Clearly, the deep learning base method for despeckling have been proposed (Wang et al, 2017), (Chierchia et al, 2017), (Vitale et al, 2019b).…”

Section: Introductionmentioning

confidence: 99%

Edge Preserving CNN Sar Despeckling Algorithm

Vitale

Ferraioli

Pascazio

2020

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

Abstract. SAR despeckling is a key tool for Earth Observation. Interpretation of SAR images are impaired by speckle, a multiplicative noise related to interference of backscattering from the illuminated scene towards the sensor. Reducing the noise is a crucial task for the understanding of the scene. Based on the results of our previous solution KL-DNN, in this work we define a new cost function for training a convolutional neural network for despeckling. The aim is to control the edge preservation and to better filter man-made structures and urban areas that are very challenging for KL-DNN. The results show a very good improvement on the not homogeneous areas keeping the good results in the homogeneous ones. Result on both simulated and real data are shown in the paper.

show abstract