Snow Avalanche Segmentation in SAR Images With Fully Convolutional Neural Networks

Bianchi, Filippo Maria; Grahn, Jakob; Eckerstorfer, Markus; Malnes, Eirik; Vickers, Hannah

doi:10.1109/jstars.2020.3036914

Cited by 34 publications

(23 citation statements)

References 20 publications

(23 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…That our F1 score of 0.81 is a lot higher than the 0.61 produced in [15] can have two reasons. On the one hand, our data set is less varied in snow conditions than that of [15] which used images acquired from multiple periods between 2014 and 2017. This facilitates the task of mapping avalanches in our case.…”

Section: E Avalanche Segmentation With a U-netmentioning

confidence: 79%

“…These numbers being quite similar is a further indicator that the method works quite well. It still has room for improvements and research such as [15] which focuses on the optimal application of deep learning independently of the preprocessing needs to be continued.…”

Section: E Avalanche Segmentation With a U-netmentioning

confidence: 99%

“…While these two methods map avalanches on the pixel level, other methods localize avalanches within a bounding box: convolutional neural networks and variational autoencoders have been used to detect avalanches in predefined zones at the end of known avalanche corridors [13], [14]. Bianchi et al applied a U-Net on S1 images of Norway and achieved an F1 score of 66% at mapping avalanches [15]. Common auxiliary information used for automatic mapping are terrain slope and land cover.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Backscatter Characteristics of Snow Avalanches for Mapping With Local Resolution Weighting

Tompkin

Leinß

2021

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

View full text Add to dashboard Cite

Snow avalanches cause a sudden change of snow properties making them detectable with synthetic aperture radar (SAR). However, steep alpine terrain combined with the slant view geometry of SAR sensors complicates detection: the avalanche brightness depends on the incidence angle and the observed area is limited by radar layover and shadow. Likewise, the spatial resolution varies strongly with the local incidence angle relative to the terrain. To increase the avalanche brightness and to improve the imaging coverage and resolution we apply local resolution weighting (LRW) on Sentinel-1 (S1) backscatter images from ascending and descending orbits. LRW merges acquisitions by averaging them, weighted with the local groundrange resolution. To analyze the relative avalanche brightness with respect to the local incidence angle and polarization, and to quantify the benefit of LRW, we created a dataset of 914 manually drawn avalanche outlines based on S1 imagery of an extreme avalanche event on January 4, 2018 in the Swiss Alps. We show that avalanches appear brightest at slopes facing away from the radar at local incidence angles of 55 ± 20 • ; such slopes are weighted considerably stronger through LRW. With a processing pipeline for avalanche segmentation using a fixed threshold on the backscatter difference we obtain a higher F1 score (0.75) with LRW compared to an unweighted orbit average (F1 = 0.68) or single orbit acquisitions (F1 = 0.5). For automatic segmentation, we used the manually drawn dataset for training and testing of a deep neural U-Net and achieved a F1 score of 0.81 on LRW backscatter differences.

show abstract

Section: E Avalanche Segmentation With a U-netmentioning

confidence: 79%

Section: E Avalanche Segmentation With a U-netmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Backscatter Characteristics of Snow Avalanches for Mapping With Local Resolution Weighting

Tompkin

Leinß

2021

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

View full text Add to dashboard Cite

show abstract

“…Lastly, neural networks have in recent years gained much popularity in detection applications, such as land cover classification [12], ground military target detection [13], sea ice concentration mapping [14], and more recently avalanche detection [15], having demonstrated their superiority over traditional image processing and machine learning methods in these applications. Several studies have already investigated the use of neural networks for surface water detection from Landsat imagery [16][17][18][19][20] and Pham-Duc et al [21] documented promising results for flood mapping by using Sentinel-1 SAR data as inputs and Landsat-8 imagery as targets to train a neural network which was shown to give accurate water detection of >90% at 30 m resolution.…”

Section: Introductionmentioning

confidence: 99%

Long-Term Water Surface Area Monitoring and Derived Water Level Using Synthetic Aperture Radar (SAR) at Altevatn, a Medium-Sized Arctic Lake

2019

Self Cite

View full text Add to dashboard Cite

Monitoring water storage in lakes and reservoirs is critical to water resource management, especially in a changing climate. Satellite microwave remote sensing offers a weather and light-independent solution for mapping water cover over large scales. We have used 13 years of synthetic aperture radar (SAR) data from three different sensors (Sentinel-1, RADARSAT-2, and Envisat advanced synthetic aperture radar (ASAR)) to develop a method for mapping surface water cover and thereby estimating the lake water extent (LWE). The method uses the unsupervised K-means clustering algorithm together with specific post-processing techniques to create binary maps of the water area. We have specifically tested and validated the method at Altevatn, a medium-sized arctic lake in Northern Norway, by using in-situ measurements of the water level. The multi-sensor SAR LWE time series were used in conjunction with the water level measurements to derive the lake hypsometry while at the same time quantifying the accuracy of our method. For Altevatn lake we estimated LWE with a root mean squared error (RMSE) of 0.89 km2 or 1.4% of the mean LWE, while the inferred lake water level (LWL) was associated with an RMSE of 0.40 m, or 2.5% of the maximum annual variation. We foresee that there is potential to further develop the algorithm by generalizing its use to other lakes worldwide and automating the process such that near real-time monitoring of LWE may be possible.

show abstract

“…In contrast, semantic segmentation classifies the multiple classes of interest in a single image at 159 pixel-level, making it suitable for complex problems like oil spill detection and classification [5], 160 [6]. is also used in many remote sensing applications [8][9][10]. It consists of an encoder (contracting 164 path) and decoder (expansive path) part as shown in Fig.…”

mentioning

confidence: 99%

Comparison of CNNs and ViTs Based Hybrid Models Using Gradient Profile Loss for Classification of Oil Spills in SAR Images

Basit¹,

Siddique²,

Sarfraz³

2021

Preprint

View full text Add to dashboard Cite

Oil spillage over a sea or ocean’s surface is a threat to marine and coastal ecosystems. Spaceborne synthetic aperture radar (SAR) data has been used efficiently for the detection of oil spills due to its operational capability in all-day all-weather conditions. The problem is often modeled as a semantic segmentation task. The images need to be segmented into multiple regions of interest such as sea surface, oil spill, look-alikes, ships and land. Training of a classifier for this task is particularly challenging since there is an inherent class imbalance. In this work, we train a convolutional neural network (CNN) with multiple feature extractors for pixel-wise classification; and introduce to use a new loss function, namely ‘gradient profile’ (GP) loss, which is in fact the constituent of the more generic Spatial Profile loss proposed for image translation problems. For the purpose of training, testing and performance evaluation, we use a publicly available dataset with selected oil spill events verified by the European Maritime Safety Agency (EMSA). The results obtained show that the proposed CNN trained with a combination of GP, Jaccard and focal loss functions can detect oil spills with an intersection over union (IoU) value of 63.95%. The IoU value for sea surface, look-alikes, ships and land class is 96.00%, 60.87%, 74.61% and 96.80%, respectively. The mean intersection over union (mIoU) value for all the classes is 78.45%, which accounts for a 13% improvement over the state of the art for this dataset. Moreover, we provide extensive ablation on different Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs) based hybrid models to demonstrate the effectiveness of adding GP loss as an additional loss function for training. Results show that GP loss significantly improves the mIoU and F1 scores for CNNs as well as ViTs based hybrid models. GP loss turns out to be a promising loss function in the context of deep learning with SAR images.

show abstract

Snow Avalanche Segmentation in SAR Images With Fully Convolutional Neural Networks

Cited by 34 publications

References 20 publications

Backscatter Characteristics of Snow Avalanches for Mapping With Local Resolution Weighting

Backscatter Characteristics of Snow Avalanches for Mapping With Local Resolution Weighting

Long-Term Water Surface Area Monitoring and Derived Water Level Using Synthetic Aperture Radar (SAR) at Altevatn, a Medium-Sized Arctic Lake

Comparison of CNNs and ViTs Based Hybrid Models Using Gradient Profile Loss for Classification of Oil Spills in SAR Images

Contact Info

Product

Resources

About