Search

Sea-surface petroleum pollution is observed as "oil slicks" (i.e., "oil spills" or "oil seeps") and can be confused with "look-alike slicks" (i.e., environmental phenomena, such as low-wind speed, upwelling conditions, chlorophyll, etc.) in synthetic aperture radar (SAR) measurements, the most proficient satellite sensor to detect mineral oil on the sea surface. Even though machine learning (ML) has become widely used to classify remotely-sensed petroleum signatures, few papers have been published comparing various ML methods to distinguish spills from look-alikes. Our research fills this gap by comparing and evaluating six traditional techniques: simple (naive Bayes (NB), K-nearest neighbor (KNN), decision trees (DT)) and advanced (random forest (RF), support vector machine (SVM), artificial neural network (ANN)) applied to different combinations of satellite-retrieved attributes. 36 ML algorithms were used to discriminate "ocean-slick signatures" (spills versus look-alikes) with ten-times repeated random subsampling cross validation (70-30 train-test partition). Our results found that the best algorithm (ANN: 90%) was >20% more effective than the least accurate one (DT: ~68%). Our empirical ML observations contribute to both scientific ocean remote-sensing research and to oil and gas industry activities, in that: (i) most techniques were superior when morphological information and Meteorological and Oceanographic (MetOc) parameters were included together, and less accurate when these variables were used separately; (ii) the algorithms with the better performance used more variables (without feature selection), while lower accuracy algorithms were those that used fewer variables (with feature selection); (iii) we created algorithms more effective than those of benchmark-past studies that used linear discriminant analysis (LDA: ~85%) on the same dataset; and (iv) accurate algorithms can assist in finding new offshore fossil fuel discoveries (i.e., misclassification reduction). [ABSTRACT FROM AUTHOR]

A novel empirical approach to categorize oil slicks’ sea surface expressions in synthetic aperture radar (SAR) measurements into oil seeps or oil spills is investigated, contributing both to academic remote sensing research and to practical applications for the petroleum industry. We use linear discriminant analysis (LDA) to try accuracy improvements from our previously published methods of discriminating seeps from spills that achieved ~70% of overall accuracy. Analyzing 244 RADARSAT-2 scenes containing 4562 slicks observed in Campeche Bay (Gulf of Mexico), our exploratory data analysis evaluates the impact of 61 combinations of SAR backscatter coefficients (σ°, β°, γ°), SAR calibrated products (received radar beam given in amplitude or decibel, with or without a despeckle filter), and data transformations (none, cube root, log10). The LDA ability to discriminate the oil-slick category is rather independent of backscatter coefficients and calibrated products, but influenced by data transformations. The combination of attributes plays a role in the discrimination; combining oil-slicks’ size and SAR information is more effective. We have simplified our analyses using fewer attributes to reach accuracies comparable to those of our earlier studies, and we suggest using other multivariate data analyses—cubist or random forest—to attempt to further improve oil-slick category discrimination.

For remote sensing of Great Lakes ice cover, a field experiment campaign was conducted in the 1997 winter season across the Straits of Mackinac and Lake Superior. The campaign was coordinated in two expeditions on two different United States Coast Guard icebreaker vessels, the Biscayne Bay in February and the Mackinaw in March. Aboard these icebreakers, the Jet Propulsion Laboratory C-band polarimetric scatterometer was used to measure backscatter signatures of various ice types and open water at incidence angles from 0° to 60°. The radar measurements include incidence angles and polarizations of spaceborne Synthetic Aperture Radars (SAR) on ERS, RADARSAT, and Envisat satellites. The radar data together with in situ measurements form a signature library that can be used to interpret SAR data for ice classification and mapping. Results are presented for backscatter signatures of Great Lakes ice types from thin lake ice to thick brash ice with different snow-cover and surface conditions. The signature library indicates that several ice types can be identified with multi-polarization SAR data; however, single-polarization data can result in misclassification of ice and open water at different ranges of incidence angle and wind conditions. For incidence angles larger than 30°, thick brash ice, the most difficult for icebreaking operations and the most hazardous for ship navigation, can be uniquely identified by co-polarized backscatter for all wind conditions below the gale force. [ABSTRACT FROM AUTHOR]

During the 1997 winter season, shipborne polarimetric backscatter measurements of Great Lakes (freshwater) ice types using the Jet Propulsion Laboratory C-band scatterometer, together with surface-based ice physical characterization measurements and environmental parameters, were acquired concurrently with Earth Resource Satellite 2 (ERS-2) and RADARSAT Synthetic Aperture Radar (SAR) data. This polarimetric data set, composed of over 20 variations of different ice types measured at incident angles from 0° to 60° for all polarizations, was processed to radar cross-section to establish a laboratory of signatures (look-up table)for different ice types. The library is used in the computer classification of calibrated satellite SAR data. Computer analysis of ERS-2 and RADARSAT ScanSAR images of Great Lakes ice cover using a supervised classification technique indicates that different ice types in the ice cover can be identified and mapped, and that wind speed and direction can have an influence on the classification of water as ice based on single frequency, single polarization data. Once satellite SAR data are classified into ice types, the ice map provides important and necessary input for environmental protection and management, ice control and ice breaking operations, and ice forecasting and modeling efforts. [ABSTRACT FROM AUTHOR]

This study investigates the backscattering properties of C-band RADARSAT (SAR) signal over sloping mountainous areas, and evaluates the utility of RADARSAT in natural disaster areas (e.g., forest fires), geological mapping, monitoring of natural hazards, natural and associated Earths environmental changes. Conventional change detection techniques in remote sensing are adequate for most applications. However, the changes in backscattering coefficients (σ0) before and after, for example, forest fires are not apparent in certain cases, mainly because of the effects of steeply sloping target scattering surface areas. The slopes of a mountainous test site can be divided into two cases with respect to the incident SAR signal look direction: the front and back slopes, where local incidence angles can vary significantly. Depending on the slope of the scattering surface, the backscattering coefficient increases to the maximum of 3.3. dB in the front-slope, but decreases 4.5 dB in back-slope after forest fires. Forest fire burnt areas provide a good example of a natural sloping scattering surface, which exhibits subtle changes in backscattering coefficients similar to the lithological changes in nature. The April 2000 forest fire in Kangwon province, Korea, burned for several weeks and was one of the worst ones in recent years. Among several firedamaged sites, an area near Samchuk city was investigated in detail with a set of multi-temporal RADARSAT Synthetic Aperture Radar (SAR) data to estimate the fire damages as compared to the unburnt background. [ABSTRACT FROM AUTHOR]

A novel empirical approach to categorize oil slicks' sea surface expressions in synthetic aperture radar (SAR) measurements into oil seeps or oil spills is investigated, contributing both to academic remote sensing research and to practical applications for the petroleum industry. We use linear discriminant analysis (LDA) to try accuracy improvements from our previously published methods of discriminating seeps from spills that achieved ~70% of overall accuracy. Analyzing 244 RADARSAT-2 scenes containing 4562 slicks observed in Campeche Bay (Gulf of Mexico), our exploratory data analysis evaluates the impact of 61 combinations of SAR backscatter coefficients (σ°, β°, γ°), SAR calibrated products (received radar beam given in amplitude or decibel, with or without a despeckle filter), and data transformations (none, cube root, log10). The LDA ability to discriminate the oil-slick category is rather independent of backscatter coefficients and calibrated products, but influenced by data transformations. The combination of attributes plays a role in the discrimination; combining oil-slicks' size and SAR information is more effective. We have simplified our analyses using fewer attributes to reach accuracies comparable to those of our earlier studies, and we suggest using other multivariate data analyses—cubist or random forest—to attempt to further improve oil-slick category discrimination. [ABSTRACT FROM AUTHOR]