Chapter 10 - Oil spill detection using optical remote sensing images and machine learning approaches (case study: Persian Gulf)

Variable-complexity machine learning models for large-scale oil spill detection: The case of Persian Gulf

Considering the form diversity of tree species composition in the Bagong Mountain National Forest Park of China, we mapped tree species utilizing Machine Learning Algorithms (support vector machines (SVM) and random forest (RF) classifiers) based on the OHS-2 hyperspectral satellite image by different datasets which combined spectral information and hyperspectral-derived vegetation indices (VIs) for improving tree species classification and explored the best performance of them. To verify the improvement, the results of physically-based spectral classifiers (spectral angle mapper (SAM) and maximum likelihood (ML) classifiers) were applied to compare with the results of machine learning algorithms. The results indicated an overall accuracy of 94.01%, 96.08%, 82.9% and 79.3% for SVM, RF, SAM and ML classifiers of the best performance using different datasets. Highest accuracies resulted from two machine learning algorithms classifiers; SVM and RF compared to SAM and ML classifiers. Although SVM outperformed RF when using all hyperspectral bands and VIs, the overall accuracy of the RF classifier is higher when compared to the SVM classifier using VIs combined selected features. Meanwhile, the RF classifier performed better than SVM after removing the redundancy of spectral data in training samples. Moreover, the machine learning algorithms successfully classified a small number of tree species (Cedrus deodara and Pterocarya stenoptera C. DC.) in the study area, but the physical spectroscopy-based method failed to classify these species. Such integration strategy improved the effectiveness of enhancing the accuracy of tree species classification and mapping their distribution on broad spatial and temporal scales using machine learning algorithms and hyperspectral imagery.

Optical image provides a unique spectral and textural opportunity to acquire data free of charge in cloud-free weather conditions. However, textural information from recent Sentinel-2A data has not been sufficiently recognized. Several researches have shown that texture features are essential for extracting crop information. This study presents a new scheme that integrates spectral bands and textural features for mapping maize fields and other land cover types (i.e., build-up, trees, others, and water) and evaluates the effectiveness of random forest (RF) and support vector machine (SVM) classifiers with different kernel functions based on Sentinel 2A (S-2A) data. The textural features were derived from multiple bands of S-2A image using a grey-level co-occurrence matrix (GLCM) algorithm. To normalize the analysis of these features, Random Forest feature selection was employed. The outcomes show that S-2A images are appropriate for mapping maize fields and other land covers when the image is accessible during the growing season. The SVM classifier with a radial basis function (RBF) kernel is superior to a polynomial (P), linear (L), and sigmoid (S) kernel functions, and the RF classifiers. For the SVM_RBF, the highest overall accuracy was obtained from the integration of spectral bands and textural features at scenario 2 SVM_RBF (87.40%), followed by the combination of spectral and textural features at scenario 2 SVM_P kernel (86.56%). However, scenario 5 SVM_P kernel (85.54%), scenario 4 SVM_P kernel function (85.30%), the spectral features alone, and the SVM_P kernel function, have the highest O/A (83.90%) than that of the SVM_RBF (83.66%). The inclusion of textural features contributed significantly in mapping the smallholder’s maize field in the study area. This demonstrated the value of S-2A in agricultural research.

Timely identification of sugarcane cultivation is essential for precision agriculture, policy formulation, and sustainable land management. Our study presents a multimodal framework that integrates spectral, morphological, and textural features from Sentinel-2 imagery with hybrid deep learning techniques for sugarcane mapping. A comprehensive feature space combining NDVI, geometrical descriptors, and GLCM-based texture metrics was used to train Random Forest (RF) and Support Vector Machine (SVM) classifiers. A hybrid convolutional neural network (CNN+MLP) architecture further merged visual embeddings from image patches with tabular descriptors. The RF model achieved a macro F1-score of 0.91, a Kappa coefficient of 0.86, and a macro-AUC of 0.94 across agricultural regions in Brazil. DBSCAN clustering was applied to filter noisy labels and improve the quality of the training dataset. A domain-specific confusion cost matrix illustrated the economic impact of misclassification, showing that the RF and CNN+MLP models improved accuracy and financial efficiency by reducing crop identification errors. Although the analysis was limited to São Paulo and Mato Grosso do Sul, the framework showed consistent and robust performance within these regions. Overall, it highlights the benefits of combining morphospectral, textural, and deep features for crop classification, providing a scalable foundation for time-series modeling and real-time land-use analytics. • Sugarcane regions were accurately mapped using Sentinel-2 satellite imagery across different agricultural zones in Brazil. • A hybrid CNN+MLP model combined visual embeddings and tabular features to improve classification. • Morphological, spectral (NDVI), and GLCM-based texture features were fused to improve crop discrimination. • Random Forest and Support Vector Machine classifiers were benchmarked and compared with the deep learning approach. • Unsupervised DBSCAN clustering was applied to detect and remove noisy samples, improving training set quality.

The objective of this study was to develop a machine learning (ML) application to determine the optimal dosage of sodium hypochlorite (NaOCl) to curtail corrosion and odor by H2 S in the headworks of a water resource recovery facility (WRRF) without overly consuming volatile fatty acids (VFAs) that are essential for the enhanced biological phosphorus removal. Given the highly diverse datasets available, three subproblems were formulated, and three cascaded ML modules were developed accordingly. The final ML models, chosen based on performance, were able to predict various targeted variables. More specifically, in Module 1, a recurrent neural network (RNN) was designed to predict wastewater characteristics. In Module 2, a random forest (RF) classifier and a support vector machine (SVM) classifier were built with the information from Module 1 along with other datasets to predict the concentrations of VFAs and H2 S, respectively. Finally, in Module 3, with the information obtained from Module 2, another RF classifier was developed to predict NaOCl dosage to reduce H2 S but keeping VFAs within the target range. These efforts are relevant and informative for WRRFs that are considering developing Intelligent Water Systems to predict the wastewater characteristics to make operational improvements. PRACTITIONER POINTS: A recurrent neural network (RNN) using long short-term memory (LSTM) successfully predicted influent wastewater parameters. A support vector machine classifier predicted hydrogen sulfide (H2 S) with 97.6% accuracy. The concentration of VFAs, an important parameter in EBPR, was predicted using a random forest classifier with 93.4% accuracy. The optimal NaOCl dosage for H2 S control can be predicted with a random forest classifier using H2 S, VFAs, and flow.

Emotion is a property by which human beings and machines can be differentiated as machines are emotionless while human beings are not. If the emotion of a speaker is recognized then others can interact accordingly. This paper presents a new approach for recognizing all the six basic emotions (Happy, anger, fear, sadness, boredom and neutral) from the speech signals more effectively. To recognize the emotion of a speaker, pitch value and two wavelet packet feature vectors derived from speech signals are used. Principal Component Analysis (PCA) has been applied to reduce the dimension of feature vectors. Random Forest (RF) and Support Vector Machine (SVM) classifiers are trained separately based on these reduced feature vectors. The experimental results show that the accuracy of emotion recognition with Random Forest classifier is 86.11% while with SVM classifier it is 84.41%. Experimentally, it is also found that clean speech of 1 sec duration is sufficient enough to recognize emotion of the speaker.

Land-cover classification using remote sensing imagery is an important part of environmental research because it provides baseline information for ecological vulnerability and risk assessment, disaster management, landscape conservation, local and regional planning, and so on. Rural-land-cover classification is challenging for both object-based image analysis methods and classifiers. The objective of this study is to improve the object-oriented classification accuracy of rural land cover by combining two models based on high spatial resolution imagery. We apply the C5.0 algorithm in R to combine support vector machines (SVMs) and random forest (RF) to create the model RS_C5.0. The effectiveness of the model combination is assessed by comparing the classification results with the state-of-the-art machine learning algorithm, namely extreme gradient boosting (XGBoost). The comparisons are done based on the classification results of both the study area and the case area. Results show that in the classification of the study area, RF performs slightly better than SVM, and XGBoost performs worse than RF but better than SVM. However, in the classification of the case area, SVM performs slightly better than RF and both SVM and RF perform better than XGBoost. Furthermore, RS_C5.0 obtains the highest overall accuracies and kappa coefficients in the classifications of both the study area and the case area. In terms of training time, XGBoost runs the slowest in the classifications of both the study area and the case area. SVM and RF as well as the combined model (RS_C5.0) run much faster than XGBoost classifier. To summarize, the combination of SVM and RF classifiers using C5.0 algorithm is found to be a fast and effective way to improve rural-land-cover classification.

Background: Methamphetamine is a chemical substance which affects the brain electrical activity of addicted individuals. The study aimed to assess the potential of using electroencephalography (EEG) signals and machine learning (ML) techniques to distinguish individuals with methamphetamine (MA) dependence from healthy individuals. Methods: The researchers utilized highly comparative time-series analysis (hctsa) for feature extraction. Three ML algorithms, logistic regression (LR), support vector machine (SVM), and random forest (RF), were employed to process the data. Various combinations of top 40 features were used to test the possibility of reaching 100% accuracy. Results: Although individual features did not achieve 100% accuracy, combinations of two features resulted in two distinct states with a prediction accuracy of 100% when using the SVM approach. Even more combinations of features with 100% accuracy were found when utilizing more features. Conclusion: Based on the findings, SVM, LR, and RF classifiers, combined with feature extraction through the hctsa method, demonstrated exceptional accuracy in identifying MA users among healthy individuals using a single EEG channel. The classification accuracy reached up to 100%.

Although advancements in remote sensing technology have facilitated quick capture and identification of the source and location of oil spills in water bodies, the presence of other biogenic elements (lookalikes) with similar visual attributes hinder rapid detection and prompt decision making for emergency response. To date, different methods have been applied to distinguish oil spills from lookalikes with limited success. In addition, accurately modeling the trajectory of oil spills remains a challenge. Thus, we aim to provide further insights on the multi-faceted problem by undertaking a holistic review of past and current approaches to marine oil spill disaster reduction as well as explore the potentials of emerging digital trends in minimizing oil spill hazards. The scope of previous reviews is extended by covering the inter-related dimensions of detection, discrimination, and trajectory prediction of oil spills for vulnerability assessment. Findings show that both optical and microwave airborne and satellite remote sensors are used for oil spill monitoring with microwave sensors being more widely used due to their ability to operate under any weather condition. However, the accuracy of both sensors is affected by the presence of biogenic elements, leading to false positive depiction of oil spills. Statistical image segmentation has been widely used to discriminate lookalikes from oil spills with varying levels of accuracy but the emergence of digitalization technologies in the fourth industrial revolution (IR 4.0) is enabling the use of Machine learning (ML) and deep learning (DL) models, which are more promising than the statistical methods. The Support Vector Machine (SVM) and Artificial Neural Network (ANN) are the most used machine learning algorithms for oil spill detection, although the restriction of ML models to feed forward image classification without support for the end-to-end trainable framework limits its accuracy. On the other hand, deep learning models’ strong feature extraction and autonomous learning capability enhance their detection accuracy. Also, mathematical models based on lagrangian method have improved oil spill trajectory prediction with higher real time accuracy than the conventional worst case, average and survey-based approaches. However, these newer models are unable to quantify oil droplets and uncertainty in vulnerability prediction. Considering that there is yet no single best remote sensing technique for unambiguous detection and discrimination of oil spills and lookalikes, it is imperative to advance research in the field in order to improve existing technology and develop specialized sensors for accurate oil spill detection and enhanced classification, leveraging emerging geospatial computer vision initiatives.

The classification of IP traffic is important for many reasons, including network management and security, quality of service (QoS) monitoring and provisioning, and high hardware utilisation. Recently, many machine learning-based IP traffic classifiers have been developed. Unfortunately, most of them need to be trained on large datasets and thus require a long training time and significant computational power. In this paper, I investigate this problem and, as a solution, present a hybrid system, which I call the ISITC, that combines the random forest (RF) and XGBoost (XGB) machine learning techniques with the support vector classifier (SVC) as the final estimator, the stacking classifier. This design leads to the development of a model that performs the classification of IP traffic and internet applications efficiently and with high accuracy. I evaluate the performance of the ISITC and various IP traffic classifiers, including neural network (NN), RF, decision tree (DT), and XGB classifiers and SVCs. The experimental results show that the ISITC provides the best IP traffic classification, with an accuracy of 96.7, and outperforms the other IP traffic classifiers: the NN classifier has an accuracy of 59, the RF classifier has an accuracy of 88.5, the DT classifier has an accuracy of 90.5, the XGB classifier has an accuracy of 89.8, and the SVC has an accuracy of 64.8.

The EU funded Iliad project builds on the assets resulting from two decades of investments in policies and infrastructures for the blue economy aiming to establish an interoperable, data-intensive, and cost-effective Digital Twin of the Ocean (DTO). Iliad DTO combines ocean observations, state-of-the-art forecast models, citizen science, AI and advanced computing infrastructures (cloud computing, HPC, Internet of Things, Big Data, social networking, and more) to create high-resolution, multi-variable and multi-dimensional near real-time virtual representations of the ocean. Continuously integrating data from numerous in-situ sensors and satellites, Iliad DTO can provide a platform for researchers, policy makers, the industry, and citizens, to monitor the impact of climate change and human activities on ocean health and productivity and make data driven management decisions for a sustainable blue economy. Several Digital Twin (DT) pilots will be undertaken in key thematic areas such as offshore wind energy, wave and tidal energy, biodiversity assessments, fisheries and aquaculture, marine pollution and more. The current work presents the Oil Spill Response Digital Twin pilot developed in the frame of Iliad. The DT focuses on Cretan Sea and aims to provide early detection of marine oil spills and operational forecasting of spill trajectories to support immediate response to pollution events, minimizing thus the impact on marine ecosystems, coastal communities and the economy and reducing the time for environmental recovery. A multi-model approach is followed for predicting the fate and transport of oil spills, employing MEDSLIK-II and OpenDrift particle tracking models, coupled to operational, high-resolution numerical weather (WRF), hydrodynamic (NEMO) and sea state (WAVEWATCH III) models for Cretan Sea. Real-time observations (current speed/direction, sea water temperature, wave height) from novel, low-cost on-line sensors are integrated in the DT and assimilated into the operational metocean forecasting chain for validation purposes and for improving models’ forecasting skills. A near-real time automatic oil spill detection system, from Sentinel-1 SAR images, is developed, which allows early detection of marine oil spills and triggers the oil spill forecasting system, producing accurate short-term forecasting of spills’ trajectories and fate. Automatic detection and classification of oil spill events employs a trained FCOS which performs initial object detection, fine-tuned with a dataset of +1000 SAR images, including 4 different classes (oil spill, look-alike, ship, land). Image pre-processing and oil spill mask delineation is performed using SNAPpy-based (Sentinel Application Platform Python toolbox) adaptive thresholding algorithms. The oil spill detection and forecasting system is tested by reconstructing the Ulysse-Virginia oil spill, which occurred off the coast of Corsica on October 7th, 2018. Sentinel-1 images are used to detect and delineate the coverage of the spill, to initiate Medslik-II and OpenDrift models for simulating the oil spill fate and transport. Oil spill predictions are produced using different metocean forcings, highlighting the importance of high-resolution metocean data in accurate forecasting of oil spill trajectories. Quantitative metrics are used to evaluate the ability of the oil spill models to reproduce the satellite oil spill observations.    Acknowledgement: This research has received funding from the European Union’s H2020 RIA programme under GA No 101037643.

Oil spill has tremendous effect on marine environment. Therefore, it is significant for protecting marine environment to get oil spill information timely. Presently, the methods which most countries adopt to detect oil spill are mainly direct detection or remote sensing methods. Among these methods, Synthetic Aperture Radar (SAR), one technique of remote sensing methods, has been becoming a hot research field nowadays. This paper concludes main characters of SAR images in oil spill detection. It introduces, compares and analyses the oil spill detection steps and realized techniques. There are four steps in oil spill detection process, including filtering, target detection, feature extraction and classification. Filtering can be completed by several filtering techniques, such as Lee, advanced Lee and Frost. General speaking, there are four methods can realize target detection, which are single threshold, adaptive threshold, wavelet transform and max entropy method. About the third step, feature extraction, 12 features can be used to represent oil spill. Classification is a key and difficult step in oil spill detection, in which algorithms based on Bayes and neural network can be adopted. Finally, the paper points out the development direction of oil spill detection in SAR images.

Background and purposeRenal cell carcinoma (RCC) is a heterogeneous group of cancers. The collagen fiber content in the tumor microenvironment of renal cancer has an important role in tumor progression and prognosis. A radiomics model was developed from dual-energy CT iodine maps to assess collagen fiber content in the tumor microenvironment of ccRCC.MethodsA total of 87 patients with ccRCC admitted to our hospital were included in this retrospective study. Among them, 59 cases contained large amounts of collagen fibers and 28 cases contained a small amount of collagen fibers. We established a radiomics model using preoperative dual-energy CT scan Iodine map (IV) imaging to distinguish patients with multiple collagen fibers from those with few collagen fibers in the tumor microenvironment of ccRCC. We extracted features from dual-energy CT Iodine map images to evaluate the effects of six classifiers, namely k-nearest neighbor (KNN), support vector machine (SVM), extreme gradient boosting (XGBoost), random forest (RF), logistic regression (LR), and decision tree (DT). The effects of the models built based on the dynamic and venous phases are also compared. Model performance was evaluated using quintuple cross-validation and area under the receiver operating characteristic curve (AUC). In addition, a clinical model was developed to assess the clinical factors affecting collagen fiber content.ResultsCompared to KNN, SVM, and LR classifiers, RF, DT, and XGBoost classifiers trained with higher AUC values, with training sets of 0.997, 1.0, and 1.0, respectively. In the validation set, the highest AUC was found in the SVM classifier with a size of 0.722. In the comparative test of the active and intravenous phase models, the SVM classifier had the best effect with its validation set AUC of 0.698 and 0.741. In addition, there was a statistically significant effect of patient age and maximum tumor diameter on the collagen fiber content in the tumor microenvironment of kidney cancer.ConclusionRadionics features based on preoperative dual-energy CT IV can be used to predict the amount of collagen fibers in the tumor microenvironment of renal cancer. This study better informs clinical prognosis and patient management. Iodograms may add additional value to dual-energy CTs.

Ballistics (the linkage of bullets and cartridge cases to weapons) is a common type of evidence encountered in criminal cases around the world. The interest lies in determining whether two bullets were fired using the same firearm. This paper proposes an automated method to classify bullets from surface topography and Land Engraved Area (LEA) images of the fired pellets using machine and deep learning methods. The curvature of the surface topography was removed using loess fit and features were extracted using Empirical Mode Decomposition (EMD) followed by various entropy measures. The informative features were identified using minimum Redundancy Maximum Relevance (mRMR), finally the classification was performed using Support Vector Machines (SVM), Decision Tree (DT) and Random Forest (RF) classifiers. The results revealed a good predictive performance. In addition, the deep learning model DenseNet121 was used to classify the LEA images. DenseNet121 provided a higher predictive performance than SVM, DT and RF classifiers. Moreover, the Grad-CAM technique was used to visualise the discriminative regions in the LEA images. These results suggest that the proposed deep learning method can be used to expedite the linkage of projectiles to firearms and assist in ballistic examinations.In this work, the bullets that were compared were air pellets fired from both air rifles and a high velocity air pistol. Air guns were used to collect the data because they were more accessible than other firearms and could be used as a proxy, delivering comparable LEAs. The methods developed here can be used as a proof-of-concept and are easily expandable to bullet and cartridge case identification from any weapon.

Background: Mammography is commonly utilised for the detection of breast cancer, but it has some drawbacks like inaccurate findings. There is a good possibility of improvements in diagnostic accuracy by ML algorithms. Preventive healthcare requires early cervical cancer screening. Machine learning techniques are used in a variety of ways to analyse data on breast cancer. This study evaluated RF and SVM models for early identification of breast cancer. The results can potentially be improved by the use of comparative analysis. It is necessary to consider the significant contribution of screening programs of cancers. The results obtained in this paper can be utilized as a reference to enhance healthcare policy. Finally, a faster and more accurate detection of cancer can save lives. Method: This research presents computational machine learning (ML) experiments to improve the detection accuracy of breast cancer. It proposes two computational algorithms: RF and SVM to measure the precision of breast cancer prediction. The WDBC dataset is taken as input for analysis. It consists of 569 cases and 30 patient features. Typical data preparation methods include the division of training and testing datasets and normalising the data. The Python language is used to create a Random Forest (RF) classifier and a Support Vector Machine (SVM) algorithm. The model’s hyperparameters are done using grid search and cross-validation methods. The models are evaluated by performance indicators. Results: RF classifier attained an accuracy rate of 96%, which confirms its position as an essential tool for the early identification of diseases. The SVM classifier was able to obtain an overall accuracy of 95%, which also indicates significant potential for early detection techniques. Finally, the findings reveal RF and SVM are equally useful in the early diagnosis of breast cancer. Each approach has specific advantages and disadvantages.