Figure 2 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Examples of oil slicks marked in yellow polygons (a-i) and lookalike phenomena (j-l) extracted from Sentinel-1 data.
Source publication
Ocean surface monitoring, emphasizing oil slick detection, has become essential due to its importance for oil exploration and ecosystem risk prevention. Automation is now mandatory since the manual annotation process of oil by photo-interpreters is time-consuming and cannot process the data collected continuously by the available spaceborne sensors...
Contexts in source publication
Context 1
... generally, any patch that is darker than the surrounding area could be an oil slick. Illustrations of spills, seeps, and lookalikes are shown in Figure 2. To summarize, the main factors involved in oil slick detection from SAR images are related to contextual information such as wind conditions, sensor characteristics, and the presence of lookalikes. ...
Context 2
... account for small slick patches separated from the main slick and displaced. The difference is shown in Figure 2a-j. Figure 7 shows an example of image annotation. ...
Context 3
... better understand the prediction of the considered improved version of FC-DenseNet models, particularly their false detections, the slick probability maps outing from the model are analyzed. Figure 20 shows an example with a heat map representing the slick probability levels. One can visually observe that the slick inner surfaces generally have a high probability (red) and that the low probability values are always on the boundaries of the slicks (yellow). ...
Context 4
... their contribution to the decision is extremely low and always countered by the considered pixel area. As a result, the prediction associated with the slick class is always close to 0%. • For a pixel classified as slick, as for the explained pixel in Figures 22b, 23c, and 24b: the prediction is based on a limited area centered around the considered pixel. The maximal SHAP value in these cases is the highest observed (about 0.5). ...
Context 5
... the network prediction for the pixel is impacted by one input feature containing mainly black pixels, which is enough to classify the pixel as a slick, with a probability above 80%. • For a pixel located on a slick edge or within a narrow slick, as for the explained pixel in Figure 23b,d: networks tend to detect and base their decision on the slick edge or the narrow slick length. This shows that networks can detect the slick edges around the selected pixel in all observed images, which significantly influences their decision. ...
Context 6
... observation is that the ground truth annotation made by the photo-interpreters is not perfect and suffer from imprecise boundaries, as shown in Figure 2, revealing a diversity of annotations that position our detection task in a noisy reference context [70], as mentioned in Section 2.2. Hence the validation process requires a second pass of expertise on the large images to check the false alarms and eventually correct the annotations to integrate some false alarms as real slicks. ...
Context 7
... are thus, in a real context, allowing a continuous progression of the models and improvement of the annotations. This also becomes apparent in the prediction probability maps obtained by the models, as shown in Figure 20, where the lowest probabilities are always found on the slick boundaries. The imprecision of the annotation boundaries is apparent in the slick IoUs limited to 0.31 in Table 5 and also in the prediction probability maps obtained by the models, where the lowest probabilities are always found on the slick boundaries, as shown in Figure 20. ...
Context 8
... also becomes apparent in the prediction probability maps obtained by the models, as shown in Figure 20, where the lowest probabilities are always found on the slick boundaries. The imprecision of the annotation boundaries is apparent in the slick IoUs limited to 0.31 in Table 5 and also in the prediction probability maps obtained by the models, where the lowest probabilities are always found on the slick boundaries, as shown in Figure 20. Despite this, the semantic and instance segmentation models are quite capable of detecting slicks but suffer from false alarms. ...
Context 9
... further, based on the comparison of the explanations presented in Section 3.3, it is possible to understand the effect of wind speed data on the model predictions. Figure 24a shows a lookalike that corresponds to waves. More precisely, one focuses on the explanation of the pixel of interest. ...
Citations
... In fact, we choose a "late fusion" strategy, fusing softmax layer scores from both networks. This method has proven effective in the state-of-the-art methods [16], by involving a parallel decision fusion strategy. Among fusion options, like averaging and maximum probability calculation, we go for averaging the class probabilities obtained from both classifiers. ...
Image change detection is an active research topic in the field of remote sensing, as it allows monitoring environmental changes that occur on temporal and spatial scales. However, most of the existing change detection methods suffer from a lack of adaptability to different image types and lack of large-scale validation. In this study, we propose an automatic change detection method, called "CD-ResUNet," based on multi-spectral NDVI imagery. It is an end-to-end deep learning method based on the fusion of two complementary deep learning networks: UNet and residual networks (ResNet). Extensive experiments have been conducted on low-resolution as well as high-resolution datasets using four represented geographical areas, which are Colombia, California, Brazil, and Duluth, each containing 145,161 patches, and the Change Detection Dataset containing 16,000 patches. For all the investigated regions, the proposed method outperforms many relevant state-of-the-art methods with an accuracy up to 99.5% and an F1-score of 99.40%.
... Features of oil slicks on SAR images may differ significantly due to the weathering process, including spreading, evaporation, emulsion, etc (Alpers, Holt, and Zeng 2017). Within this context, a deeper understanding of the physics behind sea surface scattering from mineral surfactants is a theoretical step of primary importance to design robust and effective SAR-based oil spill surveillance methods (Amri et al. 2022;Buono et al. 2018;Li et al. 2018). Within this context, it is worth investigating the backscattering from oil-covered sea surfaces to provide a better understanding of SAR measurements related to marine oil spills (Fingas and Brown 2014). ...
Model-based comparisons of near-coincident TerraSAR-X and COSMO-SkyMed VV-polarized SAR measurements over sea surface with and without oil slicks, Geo-spatial Information Science, ABSTRACT This paper contrasts predicted X-band sea surface backscattering from slick-free and oil-covered sea surfaces with actual measurements acquired by the X-band satellite TerraSAR-X and COSMO-SkyMed Synthetic Aperture Radar (SAR) missions. Two SAR scenes were acquired with a temporal difference of about 36 minutes, under similar met-ocean conditions, during the North Sea's Gannet Alpha oil spill accident. The normalized radar cross section of the slick-free sea surface is predicted using the Advanced Integral Equation Model (AIEM) while the backscatter from the oiled sea surface is predicted by the AIEM augmented with the Model of Local Balance (MLB) to include the damping effect of oil slicks. Experimental results show that X-band co-polarized numerical predictions agree reasonably well with both TSX and CSK actual measurements collected over slick-free sea surfaces. When dealing with oil-covered sea surfaces, the predicted backscattering reasonably agrees with TSX measurements, while it overestimates the CSK ones. This is likely due to the different spreading conditions of the oil imaged by the two satellite missions. ARTICLE HISTORY
... Ker-nelSHAP and DeepSHAP are some methods proposed for estimating SHAP values, outperforming other techniques such as LRP and LIME (Linardatos et al. 2021;Van et al. 2004). SHAP also provides insights into feature importance and their impact on network decisions (Lundberg et al. 2017;Knapič et al. 2021;Amri et al. 2022). ...
The automated detection of plant species carries immense significance across diverse domains such as agriculture, environmental monitoring, biodiversity preservation, and climate change research. In an era marked by environmental shifts and climate change's profound implications, precise plant species identification plays a pivotal role in decoding our evolving world. This study focuses on the plant classification within Saudi Arabia's distinctive botanical landscape. Initiating with the creation of the Saudi Arabia Flora Dataset (SAFD), an extensive collection of ten rare plant species and their various subspecies. This dataset, available upon request, originates from the diverse landscapes of Saudi Arabia, aiming to bolster biodiversity preservation and ecological research. This study conducts a comprehensive comparative analysis of relevant deep learning models, laying the foundations for our main contribution ; the introduction of Collaborative Ensemble Plant Detection Model (CEPDM). This novel model harnesses the strengths of established architectures, achieving an accuracy rate of 99%, 96% precision and 98% F1 score. Furthermore, to enhance transparency, we incorporate eXplainable Artificial Intelligence (XAI) techniques, providing insights into the decisive factors guiding our model's predictions. Automated plant detection in Saudi Arabia contributes to biodiversity preservation, ecological research, and climate resilience.
... Researchers have continually investigated several approaches to properly handle the change detection. Currently, powerful deep learning algorithms are transforming the field of change detection, promising more accurate and efficient solutions across a wide range of applications[ [21], [15]]. Advancements in remote sensing technologies, coupled with the rapid growth of deep learning methodologies, have opened new avenues for advanced studies in change detection. ...
Image change detection in remote sensing is crucial for monitoring environmental changes at different temporal and spatial scales. The primary goal is to identify changed pixels in multi-temporal images accurately. However, challenges such as limited large-scale validation and response latency persist. In this study, we propose an accurate automated deep learning change detection method called "ResNet-Unet" based on multi-spectral NDVI imagery. ResNet-Unet combines UNet and residual networks, while employing deep learning based features for precise change detection. We evaluate the proposed method on low-resolution data from three geographical regions: Colombia, California, and Duluth. In each region, there are 145,161 patches, giving us extensive coverage for our experiments. Our method is validated in three different areas, where we achieve an accuracy of 99.50% and an F1 score of 99.41%.
... Most current approaches for detecting oil pollution on the water surface involve either oil spill detection on the sea surface using various remote sensing monitoring methods to detect the type and area of the oil spill [10][11][12][13], or using remote sensing images to detect the oil slick [14,15]. Zhao Dong et al. [16] proposed an oil slick detection method based on multispectral remote sensing technology for sea surface oil slick detection, but the method is less effective in detecting oil slicks heavily polluted by sunlight and needs to be verified by more multispectral images. ...
The oil in hydropower station catchment wells is a source of water pollution which can cause the downstream river to become polluted. Timely detection of oil can effectively prevent the expansion of oil leakage and has important significance for protecting water sources. However, the poor environment and insufficient light on the water surface of catchment wells make oil pollution detection difficult, and the real-time performance is poor. To address these problems, this paper proposes a catchment well oil detection method based on the global relation-aware attention mechanism. By embedding the global relation-aware attention mechanism in the backbone network of Yolov5s, the main features of oil are highlighted and the minor information is suppressed at the spatial and channel levels, improving the detection accuracy. Additionally, to address the problem of partial loss of detail information in the dataset caused by the harsh environment of the catchment wells, such as dim light and limited area, single-scale retinex histogram equalization is used to improve the grayscale and contrast of the oil images, enhancing the details of the dataset images and suppressing the noise. The experimental results show that the accuracy of the proposed method achieves 94.1% and 89% in detecting engine oil and turbine oil pollution, respectively. Compared with the Yolov5s, Faster R-CNN, SSD, and FSSD detection algorithms, our method effectively reduces the problems of missing and false detection, and has certain reference significance for the detection of oil pollution on the water surface of catchment wells.
... A novel automatic detection of offshore oil slicks was developed using multi-modal deep learning [127]. An expert photo-interpreter annotated Sentinel-1 SAR data over four years and three areas, specifically, the Atlantic Ocean coast in Southern Africa (Nigeria and Namibia), and the Mediterranean Sea in Western Asia (Lebanon). ...
In 1978, the SEASAT satellite was launched, carrying the first civilian synthetic aperture radar (SAR). The mission was the monitoring of ocean: application to land was also studied. Despite its short operational time of 105 days, SEASAT-SAR provided a wealth of information on land and sea, and initiated many spaceborne SAR programs using not only the image intensity data, but also new technologies of interferometric SAR (InSAR) and polarimetric SAR (PolSAR). In recent years, artificial intelligence (AI), such as deep learning, has also attracted much attention. In the present article, a review is given on the imaging processes and analyses of oceanic data using SAR, InSAR, PolSAR data and AI. The selected oceanic phenomena described here include ocean waves, internal waves, oil slicks, currents, bathymetry, ship detection and classification, wind, aquaculture, and sea ice.
... At present, research on the interpretability of SAR deep learning algorithms is in its infancy [60][61][62][63]. In 2021, Shendryk et al. explained the contribution of different inputs to machine learning models for predicting sugarcane yield using SHAP [60]. ...
... In 2022, Al-Najjar et al. employed SHAP to measure the impact, interaction, and correlation of landslide condition factors in RF and SVM [61]. In addition, SHAP was applied to visualize and interpret the results of deep learning SAR oil slick detection [63]. In these studies, algorithms such as SHAP were used only for post-hoc interpretation analysis of model results and failed to fully incorporate the characteristics of SAR images. ...
Reliable and timely rice distribution information is of great value for real-time, quantitative, and localized control of rice production information. Synthetic aperture radar (SAR) has all-weather and all-day observation capability to monitor rice distribution in tropical and subtropical areas. To improve the physical interpretability and spatial interpretability of the deep learning model for SAR rice field extraction, a new SHapley Additive exPlanation (SHAP) value-guided explanation model (SGEM) for polarimetric SAR (PolSAR) data was proposed. First, a rice sample set was produced based on field survey and optical data, and the physical characteristics were extracted using decomposition of polarimetric scattering. Then a SHAP-based Physical Feature Interpretable Module (SPFIM) combing the long short-term memory (LSTM) model and SHAP values was designed to analyze the importance of physical characteristics, a credible physical interpretation associated with rice phenology was provided, and the weight of physical interpretation was combined with the weight of original PolSAR data. Moreover, a SHAP-guided spatial interpretation network (SSEN) was constructed to internalize the spatial interpretation values into the network layer to optimize the spatial refinement of the extraction results. Shanwei City, Guangdong Province, China, was chosen as the study area. The experimental results showed that the physical explanation provided by the proposed method had a high correlation with the rice phenology, and spatial self-interpretation for finer extraction results. The overall accuracy of the rice mapping results was 95.73%, and the kappa coefficient reached 0.9143. The proposed method has a high interpretability and practical value compared with other methods.
... Besides the above-mentioned applications in ocean observation, this Special Issue presents various advanced ocean remote sensing technologies and their applications, including the use of artificial intelligence (AI) technology to explore ocean information [61][62][63][64] and reconstruct missing values [65,66]. The applications of ocean remote sensing detailed in this Special Issue include methods for the observation of changes in the ocean environment [67][68][69][70][71][72][73][74][75][76][77] and fishing ground [78], as well as the dynamics of the ocean, such as internal tides [79], internal waves [80,81], eddies and wakes [82,83], upwelling [84,85], ocean current [86][87][88][89] and even bibliometric analysis applied to oil detection and mapping [90]. ...
... However, it is difficult to automatically distinguish between man-made (spill) and natural (seep) oil slicks from synthetic aperture radar (SAR) images using limited datasets. Amri et al. [62] introduced the application of deep learning for automated offshore oil slick detection in SAR images. The data used were derived from a large database of real and recent oil slick monitoring for both types of oil slicks. ...
The launch of Seasat, TIROS-N and Nimbus-7 satellites equipped with ocean observation sensors in 1978 opened the way for remote sensing applications in ocean observation [...]
Fluid particle detection technology is of great importance in the oil and gas industry for improving oil-refining techniques and in evaluating the quality of refining equipment. The article discusses the process of creating a computer vision algorithm that allows the user to detect water globules in oil samples and analyze their sizes. The process of developing an algorithm based on the convolutional neural network (CNN) YOLOv4 is presented. For this study, our own empirical base was proposed, which comprised microphotographs of samples of raw materials and water–oil emulsions taken at various points and in different operating modes of an oil refinery. The number of images for training the neural network algorithm was increased by applying the authors’ augmentation algorithm. The developed program makes it possible to detect particles in a fluid medium with the level of accuracy required by a researcher, which can be controlled at the stage of training the CNN. Based on the results of processing the output data from the algorithm, a dispersion analysis of localized water globules was carried out, supplemented with a frequency diagram describing the ratio of the size and number of particles found. The evaluation of the quality of the results of the work of the intelligent algorithm in comparison with the manual method on the verification microphotographs and the comparison of two empirical distributions allow us to conclude that the model based on the CNN can be verified and accepted for use in the search for particles in a fluid medium. The accuracy of the model was AP@50 = 89% and AP@75 = 78%.
