Oil spillage over a sea or ocean surface is a threat to marine and coastal ecosystems. Spaceborne synthetic aperture radar (SAR) data have been used efficiently for the detection of oil spills due to their 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, lookalikes, 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 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, lookalikes, 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 $F_1$ 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.

Nitrogen dioxide (NO2) is one among several constituents of air pollution. To restrict its surface-level concentration to within the limits prescribed by regulatory authorities, dedicated monitoring of its spatiotemporal spread is needed. Satellite-based remote sensing of the tropospheric composition can be used to estimate NO2 concentration. However, this is not a direct measurement of the surface-level NO2 concentration, though several studies have shown that the tropospheric vertical column density (VCD) estimated by the satellite sensor is correlated with surface-level concentration. This review article covers various aspects related to the estimation of surface-level NO2 using remotely sensed data. It provides detailed literature, tracing the evolution of the various methods developed for the estimation, from scaling methods to the initial linear regression (LR) models onward to the more recent deep learning (DL) architectures. The performance of these models is critically reviewed.

Oil spills cause a significant threat to marine and coastal ecosystems. It is one of the major causes of water pollution. This research focuses on the use of deep learning for oil spills detection and classification. UNet is a convolutional neural network, originally proposed for biomedical image segmentation and modified for the discrimination of oil spills and look-alikes. The model is trained on a publicly available benchmark oil spill detection dataset of Sentinel-1 synthetic aperture radar (SAR) images. The images have been semantically segmented into multiple regions of interest such as sea surface, oil spills, look-alikes, ships and land. The proposed UNet-based model achieves intersection over union (IoU) value of 95.69% for sea surface, 60.85% for oil spills, 54.90% for look-alikes, 70.27% for ships and 96.79% for land class. The mean intersection over union (mIoU) value for all the classes is 75.70% which consitutes a nearly 10% increase compared to state of the art for this dataset.

Oil spillages on a sea’s or an ocean’s surface are a threat to marine and coastal ecosystems. They are mainly caused by ship accidents, illegal discharge of oil from ships during cleaning and oil seepage from natural reservoirs. Synthetic-Aperture Radar (SAR) has proved to be a useful tool for analyzing oil spills, because it operates in all-day, all-weather conditions. An oil spill can typically be seen as a dark stretch in SAR images and can often be detected through visual inspection. The major challenge is to differentiate oil spills from look-alikes, i.e., low-wind areas, algae blooms and grease ice, etc., that have a dark signature similar to that of an oil spill. It has been noted over time that oil spill events in Pakistan’s territorial waters often remain undetected until the oil reaches the coastal regions or it is located by concerned authorities during patrolling. A formal remote sensing-based operational framework for oil spills detection in Pakistan’s Exclusive Economic Zone (EEZ) in the Arabian Sea is urgently needed. In this paper, we report the use of an encoder–decoder-based convolutional neural network trained on an annotated dataset comprising selected oil spill events verified by the European Maritime Safety Agency (EMSA). The dataset encompasses multiple classes, viz., sea surface, oil spill, look-alikes, ships and land. We processed Sentinel-1 acquisitions over the EEZ from January 2017 to December 2023, and we thereby prepared a repository of SAR images for the aforementioned duration. This repository contained images that had been vetted by SAR experts, to trace and confirm oil spills. We tested the repository using the trained model, and, to our surprise, we detected 92 previously unreported oil spill events within those seven years. In 2020, our model detected 26 oil spills in the EEZ, which corresponds to the highest number of spills detected in a single year; whereas in 2023, our model detected 10 oil spill events. In terms of the total surface area covered by the spills, the worst year was 2021, with a cumulative 395 sq. km covered in oil or an oil-like substance. On the whole, these are alarming figures.

Nitrogen dioxide (NO 2 ) is one of the six gaseous air pollutants that need regular monitoring in big cities around the world. It contributes to particle pollution and can trigger chemical reactions that lead to increased concentration of ozone in the troposphere. Lahore, a metropolitan city of Pakistan is among the most polluted cities in the world. Area-wide monitoring of NO 2 is necessary in this region to devise a long-term emission control policy. However, it lacks a dense network of ground-based air quality monitoring stations (AQMS), which is need of the hour. The installation of AQMS requires huge financial resources. In this paper, we investigate a machine learning-based approach to estimate surface level concentration of NO 2 using remote sensing and modeled meteorological data. We use multiple linear regression (M1) and a polynomial fitted regression (M2) techniques to model ambient NO 2 , using remotely sensed vertical column density (VCD) of NO 2 , acquired by tropospheric monitoring instrument (TROPOMI), onboard Sentinel 5P satellite, and modeled meteorological parameters such as surface pressure, dew point temperature, and wind speed. Results show that M2 outperformed M1 with an $\mathbf{R}^{2}$ value of 0.49 and root mean square error (RMSE) value of $\mathbf{19}.\mathbf{27}\ \mu \mathbf{g}/\mathbf{m}^{3}$ . There is a moderate positive correlation between in-situ measurements and remotely sensed VCD of NO 2 , which makes it an interesting problem that needs to be explored further to achieve desirable results.

Glacial lake outburst floods (GLOFs) are a major threat to the local communities and important infrastructures in the high mountain regions. This paper focuses on the development of a benchmark dataset for glacial lakes classification in Sentinel 2 multi-spectral data and subsequent detection of glacial lakes prior to a glacial lake outburst flood (GLOF). Towards this end, we collected Sentinel 2 true color scenes of High-Mountain Asia (HMA) region using glacial lakes inventory of this region. It covers an area of 2080.12 km ² with nearly 30,121 glacial lakes. After data collection, we retained 1200 cloud free true color images and manually generated their ground truth masks. The dataset covers lakes with different shapes, sizes and radiometric signatures. For detection of glacial lakes, we used an encoder-decoder based convolutional neural network (CNN). The model is trained on the labelled dataset of glacial lakes for semantic segmentation of true color images into two relevant classes: lake and no lake. The performance of the proposed model is evaluated using intersection over union (IoU) score. It classifies glacial lakes correctly with an IoU score of 79.90%, which is quite good as far as complexity of the problem is concerned.