Wide-Area Land Cover Mapping With Sentinel-1 Imagery Using Deep Learning Semantic Segmentation Models

Šćepanović, Sanja; Antropov, Oleg; Laurila, Pekka; Rauste, Yrjö; Ignatenko, Vladimir; Praks, Jaan

doi:10.1109/jstars.2021.3116094

Cited by 35 publications

(33 citation statements)

References 82 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since recently, DL models have been increasingly used in Earth Observation applications [13], [14], indicating potential for analyzing complex structured vegetation such as, e.g., boreal forests. Despite promising results for classification tasks, DL has not yet been used extensively for regression modeling tasks in environmental studies [15].…”

Section: Introductionmentioning

confidence: 99%

Improved Semisupervised UNet Deep Learning Model for Forest Height Mapping With Satellite SAR and Optical Data

et al. 2022

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

In this study, we introduce an improved semisupervised deep learning approach, and demonstrate its suitability for modeling the relationship between forest structural parameters and satellite remote sensing imagery and producing forest maps. The improved approach is based on a popular UNet model, modified and fine-tuned to improve the forest parameter prediction performance. Within the improved model, squeezeand-excitation blocks are embedded to re-calibrate the multisource features via retrieved channel-wise self-attention and a novel Cross-Pseudo Regression strategy is implemented to train the model in a semi-supervised way. The improvement imposes consistency learning on two perturbed network branches, generating regression pseudo-reference and expanding the dataset size. For demonstration, we used satellite synthetic aperture radar (SAR) Sentinel-1 and multispectral optical Sentinel-2 images as remote sensing data, complemented with reference data represented by forest tree height as one of the key forest structural variables. The study area is located in a boreal forestland in Central Finland. Proposed approach showed larger accuracy compared to traditional machine learning methods such as random forests and boosting trees, and baseline UNet model. Best accuracy figures for forest tree height were achieved with combined SAR and optical imagery and were as small as 24.1 % RMSE (root mean square error) on pixel-level and 15.4 % RMSE on forest stand level.

show abstract

Section: Introductionmentioning

confidence: 99%

Improved Semisupervised UNet Deep Learning Model for Forest Height Mapping With Satellite SAR and Optical Data

et al. 2022

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

Self Cite

View full text Add to dashboard Cite

show abstract

“…Semantic segmentation of remote sensing images aims to produce pixel-wise categorical labels to facilitate interpretation of the remote sensing data [1], [2], [3]. The semantically parsed annotation enables an intuitive perception of targets, therefore, has been widely adopted in downstream tasks, such as land-cover mapping [4], [5], water resources management [6], [7] and disaster assessment [8], [9], among others.…”

Section: Introductionmentioning

confidence: 99%

A Synergistical Attention Model for Semantic Segmentation of Remote Sensing Images

Liu

et al. 2023

IEEE Trans. Geosci. Remote Sensing

View full text Add to dashboard Cite

In remotely sensed images, high intra-class variance and inter-class similarity are ubiquitous due to complex scenes and objects with multivariate features, making semantic segmentation a challenging task. Deep convolutional neural networks can solve this problem by modelling the context of features and improving their discriminability. However, current learning paradigms model the feature affinity in spatial dimension and channel dimension separately and then fuse them in a sequential or parallel manner, leading to suboptimal performance. In this study, we first analyze this problem practically and summarize it as attention bias that reduces the capability of network in distinguishing weak and discretely distributed objects from widerange objects with internal connectivity, when modeled only in spatial or channel domain. To jointly model both spatial and channel affinity, we design a synergistic attention module (SAM), which allows for channel-wise affinity extraction while preserving spatial details. In addition, we propose a synergistic attention perception neural network (SAPNet) for the semantic segmentation of remote sensing images. The hierarchicalembedded synergistic attention perception module aggregates SAM-refined features and decoded features. As a result, SAPNet enriches inference clues with desired spatial and channel details. Experiments on three benchmark datasets show that SAPNet is competitive in accuracy and adaptability compared with stateof-the-art methods. The experiments also validate the hypothesis of attention bias and the efficiency of SAM.

show abstract

“…Delalay et al, [56] used SegNet for crop type classification by using a fusion of optical and SAR data. [57] used this network to classify the land cover features on SAR images, and many researchers have used the SegNet network on satellite images. In 2017, Cehn et al, [58] proposed the DeepLab architecture.…”

Section: Introductionmentioning

confidence: 99%

“…Image level features are also sent to the ASPP module in this DeepLabV3+ module, incorporating a decoder module to refine segmentation results [59]. Classifying SAR image formats using DeeplabV3+ [57,60,61].…”

Section: Introductionmentioning

confidence: 99%

Performance Evaluation of DL-based Models for LULC segmentation using SAR time series data of regions in Gujarat

Jain

Varghese²,

Joshi

et al. 2023

Preprint

View full text Add to dashboard Cite

Both spatial and non-spatial data are critical to a country's social and economic development. Land use and land cover (LULC) maps are data that can be used to retrieve land information and detect environmental changes. Several recent studies have used deep learning approaches to generate LULC maps from various remote-sensing image formats. They have correctly recognized several land cover elements such as built-up areas, forests, water, etc. On the other hand, spaceborne radar images must be used to test deep learning methods because radar sensors operate in all weather and lighting circumstances and have diverse image characteristics depending on the wavelength. Although many researchers have applied deep learning techniques to classify the land cover features, most of them need to do more by taking less or more several land cover features and taking proper evaluation metrics to justify the classification techniques. The current research shows segmentation on Sentinel-1 C-band synthetic aperture radar (SAR) images of Gujarat's Ahmedabad, Anand, and Kheda regions using four deep-learning models. U-Net, SegNet, FCN, and DeepLabV3+ have primarily been employed in prior research to classify SAR and other satellite image formats and showed excellent results in identifying different land cover features. Training samples of five primary classes for ground-truth data were obtained using Google Earth and Sentinel-2 images as reference maps. For segmenting the Sentinel-1 image, above mentioned four deep-learning semantic segmentation models were chosen and fine-tuned. All models did exceptionally well, with overall accuracy ranging from 76% to 90%. U-Net fared well, followed by DeepLabV3+ in classifying all five land cover features, with mean intersection over union (IoU) of 0.67 and 0.55, respectively. The results reveal that using a self-generated reference map, the Sentinel-1 dual-pol images can categorize the primary land cover features.

show abstract

Wide-Area Land Cover Mapping With Sentinel-1 Imagery Using Deep Learning Semantic Segmentation Models

Cited by 35 publications

References 82 publications

Improved Semisupervised UNet Deep Learning Model for Forest Height Mapping With Satellite SAR and Optical Data

Improved Semisupervised UNet Deep Learning Model for Forest Height Mapping With Satellite SAR and Optical Data

A Synergistical Attention Model for Semantic Segmentation of Remote Sensing Images

Performance Evaluation of DL-based Models for LULC segmentation using SAR time series data of regions in Gujarat

Contact Info

Product

Resources

About