Lithological Classification Research Articles

Formation lithology classification using scalable gradient boosted decision trees

The classification of underground formation lithology is an important task in petroleum exploration and engineering since it forms the basis of geological research studies and reservoir parameter calculations. Hence, there have recently been increased efforts to automate lithology classification by incorporating various data science tools and principles. In this regard, efforts were made recently to evaluate machine learning methods to classify formation lithology by using data from the Daniudui gas field (DGF) and Hangjinqi gas field (HGF), both located in China. Although the boosted ensemble learners utilized in the studies performed well, there is still scope for improvement with respect to the prediction metrics. Additionally, the issue of scalability of some of these algorithms is also of concern. Hence, building upon the success of these algorithms in the previous studies, we tap into the state of the art of scalable ensemble decision tree algorithms, in our study. Specifically, we applied recently developed gradient boosted decision tree (GBDT) systems, namely, XGBoost, LightGBM and CatBoost, after combining well log data obtained from DGF and HGF. We compare their performance with random forests (RFs), AdaBoost and gradient boosting machines (GBMs) which serve as a baseline. We evaluated the algorithms using metrics such as the micro average, macro average and weighted average of precision (Pr), recall (Re) and F1-score (F1) on the test set after hyperparameter tuning. In our analysis, among the applied algorithms, we found that LightGBM possessed the highest metrics. Our work identifies LightGBM and CatBoost as good first-choice algorithms for the supervised classification of lithology when utilizing well log data.

The Upper Critical Zone of the Rustenburg Layered Suite in the Swartklip Sector, north-western Bushveld Complex, on the farm Wilgerspruit 2JQ: I. Stratigraphy and PGE mineralization patterns

AbstractThe Upper Critical Zone of the Rustenburg Layered Suite (RLS) in the Swartklip Sector, north-western Bushveld Complex, is considerably attenuated relative to other parts of the Complex. The interval between the UG2 chromitite and the Merensky Reef amounts to as little as 25 m in places. Within this interval, the aggregate thickness of orthopyroxenite-dominated ultramafic layers hosting the UG1 and UG2 chromitites and the Merensky and Bastard reefs does not differ significantly from the area around Rustenburg, to the south. The total thickness of ultramafic lithologies is, in fact, increased by the presence of the 3 to 5 m thick olivine-rich Pseudo Reef Unit, which is developed between the UG2 and Merensky Reef units in the Swartklip Sector, but does not occur in any significant form elsewhere in the Bushveld intrusion. The substantial thinning of the succession is due almost entirely to the fact that plagioclase-rich rocks (norite and anorthosite) between the ultramafic layers are radically thinned in the Swartklip Sector relative to virtually all other parts of the Bushveld Complex. The ultramafic layers, although dominated by orthopyroxenite, are characterized by higher proportions of olivine than in other parts of the Bushveld Complex. In our logging of the substantial number of exploration drill cores that form the basis of this study, we have found it expedient to define stratigraphic units that are either exclusively plagioclase-rich (norite and anorthosite) or plagioclase-poor (consisting of varying proportions of orthopyroxenite, harzburgite and chromitite). This effectively binary system of lithological classification has no overt genetic connotations.Our nomenclature has, in fact, enabled us to rigorously document the nature of contacts between ultramafic and plagioclase-rich units, and thus to identify unconformities between the ultramafic units (orthopyroxenite and harzburgite) and intervening noritic and anorthositic units, which have in the past been ascribed to localized thermo-chemical erosion of pre-existing plagioclase-rich cumulates. Apart from the well-documented evidence of erosional unconformities at the basal contacts of ultramafic units, we also provide evidence for unconformities at the tops of these units.

Photogrammetric survey and 3D model as experimental tool for mapping of polychrome marbles in artworks: the case of two Baroque altars in Bari (Italy)

ABSTRACTThis paper focuses on the use of the photogrammetric method to obtain 3D model of artworks for their characterisation and conservation. The high-resolution reproduction of a piece of art allows both the identification of constituent materials, decorative elements and detailed features and its digitalisation and enjoyment and then enhancement. In this study the approach was applied to the polychrome marble altars in the crypt of the St. Sabino’s cathedral and in the St. Giacomo’s church, both in Bari (Italy), significant evidences of virtuosity of Baroque sculptors and of use and reuse of precious Roman marbles combined with more recent and local stones. For this purpose, firstly, a three-dimensional model and high definition survey were performed by collection of numerous photographs and their processing using photogrammetric method and digital survey software; in addition, identification of Roman marbles and other decorative stones was carried out in terms of lithological classification, provenance locality, the historical period of employ. Results, obtained through an inexpensive, fast and complete survey, allowed to produce a high definition representation of the altars for a complete mapping of used material and decorative features and to guarantee their conservation, sharing and enhancement.

Rock lithological classification by hyperspectral, range 3D and color images

Mine operations in the future will require automatic rock characterization at many different stages, since it can be used to supervise and optimize various processes in the laboratory as well as at the mine locations for planning and exploitation. Different methods for classifying rocks based on image analysis have been proposed in the past. In this paper, we report the use of hyperspectral sensors in the classification of rock lithology. The lithology provides information about the chemical composition of the rock, and its physical properties. According to our literature review, this is the first time hyperspectral sensors have been employed in rock type classification. Additionally, it is the first time the use of three different technologies in rock type classification has been reported: hyperspectral sensors, laser range and a color camera. We use two hyperspectral sensors, one has sensibility within the visible and near-infrared range of 400–1000 nm, and the second has sensibility within the short-wavelength infrared range of 900–2500 nm. The range and high definition color images are used to perform accurate segmentation of the rock samples. Images are tessellated into sub-images in which various features from the three sensor types are extracted. In a first stage, the sub-images are classified by using a support-vector machine (SVM) classifier with the extracted features as inputs. In a second stage, the rock segmentation is used to perform a voting process among all the sub-images of each rock and improve the classification. The method was tested using a database with 13 lithologies from a copper mine in Chile. The results show that lithological classification performance obtained by using hyperspectral images greatly exceeds the performance of the color and range images. The achieved classification performance was 98.62% using sub-image classification and 99.95% using a voting process among sub-images. The number of features was also reduced by using the CMIM (Conditional Mutual Information Maximization) feature selection method, achieving a performance of over 99% with using only 3% of the total number of features.

Group Method of Data Handling (GMDH) Lithology Identification Based on Wavelet Analysis and Dimensionality Reduction as Well Log Data Pre-Processing Techniques

Although the group method of data handling (GMDH) is a self-organizing metaheuristic neural network capable of developing a classification function using influential input variables, the results can be improved by using some pre-processing steps. In this paper, we propose a joint principal component analysis (PCA) and GMDH (PCA-GMDH) classifier method. We investigated well log data pre-processing techniques composed of dimensionality reduction (DR) and wavelet analysis (WA), using the southern basin of the South Yellow Sea as a case study, with the aim of improving the lithology classification accuracy of the GMDH. Our results showed that the dimensionality reduction method, which is composed of PCA and linear discriminant analysis (LDA), minimized the complexity of the classifier by reducing the number of well log suites to the relevant components and factors. On the other hand, the WA decomposed the well log signals into time-frequency wavelets for the GMDH algorithm. Of all the pre-processing methods, only the PCA was able to significantly increase the classification accuracy rate of the GMDH. Finally, the proposed joint PCA-GMDH classifier not only increased the accuracy but also was able to distinguish between all the classes of lithofacies present in the southern basin of the South Yellow Sea.

Block stream characteristics in Southern Carpathians (Romania)

Block streams, as common features in the Southern Carpathians, are essential for understanding the evolution of alpine environments. Despite their widespread occurrence, they have frequently been omitted from previous inventories of periglacial landforms and from the general understanding of landscape development. This study evaluated the environmental controls on spatial distribution and dimensions of 2056 block streams for a better understanding of their evolution and climatic significance. Different lithology classes and topographic characteristics were assessed using standard least-square linear regression and one-way analysis of variance (ANOVA). The majority of the block streams belong to the linear morphological type, but six other types of morphologies were found (e.g. sinuous, curved, and braided). Bivariate and multivariate statistical analysis revealed a high density of landforms on southern slopes (59%), granite terrains (47%), and above 1750 m. Large features occur mainly on southern slopes and in areas with granites. Landform dimensions and spatial distribution are influenced to a certain degree by lithology, slope aspect, and elevation, making it difficult to identify a dominant factor. Block streams developed extensively at the end of the Pleistocene under permafrost conditions. Today, these features suggest inactivity, but during the cool episodes of the Holocene (e.g. Little Ice Age - LIA), the formation of block streams above 2100 m was re-initiated. Future absolute dating and monitoring of landform dynamics will support these findings.

Lithological classification of cement quarry using discriminant algorithms

As such in any industrial raw material site characterization study, making a lithological evaluation for cement raw materials includes a description of physical characteristics as well as grain size and chemical composition. For providing the cement components in accordance with the specifications required, making the classification of the cement raw material pit is needed. To make this identification in a spatial system at a quarry stage, the supervised pattern recognition analysis has been performed. By using four discriminant analysis algorithms, lithological classifications at three levels, which are with limestone, marly-limestone (calcareous marl) and marl, have been made based on the main chemical components such as calcium oxide (CaO), alumina (Al2O3), silica (SiO2), and iron (Fe2O3). The results show that discriminant algorithms can be used as strong classifiers in cement quarry identification. It has also recorded that the conditional and mixed classifiers perform better than the conventional discriminant algorithms.

Using UAV for automatic lithological classification of open pit mining front

Abstract Mine planning is dependent on the natural lithologic features and on the definition of their limits. The geological model is constantly updated during the life of the mine, based on all the information collected so far, plus the knowledge developed from the exploration stage up to the mine closure. As the mine progresses, the amount of available data increases, as well as the experience of the geological modeller and mine planner who deliver the short, medium, and long-term plans. This classical approach can benefit from the automation of the geological mapping on the mining faces and outcrops, improving the speed of repetitious work and avoiding exposure to intrinsic dangers like mining equipment, falling rocks, high wall proximity, among others. The use of photogrammetry to keep up with surface mining activities boarded in UAVs is a reality and the automated lithological classification using machine learning techniques is a low-cost evolution that might present accuracies above 90% of the contact zones and lithologies based on the automated dense point cloud classification when compared to the manual (or reality) classified model.

THE CONTRIBUTION OF SENTINEL-2 SATELLITE IMAGES FOR GEOLOGICAL MAPPING IN THE SOUTH OF TAFILALET BASIN (EASTERN ANTI-ATLAS, MOROCCO)

Abstract. Geological mapping in desert, mountainous or densely vegetated areas are sometimes faced with many constraints. Recently several remote sensing methods are used on ASTER or LANDSAT imagery for making that task easier. The aim of this paper is to evaluate the applicability of some of these methods on Sentinel-2A images. The study, therefore, focuses on a lithological classification using these multispectral images in the south of the Tafilalet basin. To achieve this goal, two L1C level images were used. Decorelation stretch combined with the optimal index factor (OIF) and Minimum Noise Fraction (MNF) were the main improvements used for RGB combination images. The classifiers Spectral Angle Mapper (SAM) and the Maximum Likelihood classifier (MLC) have been evaluated for a most accurate classification to be used for our lithofacies mapping. The latest drawn geological maps and RGB images of false colour combinations were used to select regions of interest (ROI) as the endmembers to use for these classifiers. Obtained results showed a clear discrimination of the different lithological units of the study area. Classifications evaluation showed that the Maximum likelihood classifier is more accurate with an overall accuracy of 76% and a Kappa coefficient is 0.74. Finally, this study has shown the importance of the use of sentinel-2 multispectral images in geological mapping and has shown that the high spectral resolution of the VNIR and SWIR bands creates a synergy with the high spatial resolution for optimal lithological mapping.

Automated lithological classification using UAV and machine learning on an open cast mine

ABSTRACTMine planning is directly dependent on the lithological features and the definition of contacts between materials. Geological modelling is a continual duty that is performed using observation data, which includes open faces information. New data must be continuously acquired and more details are added to the model. This task can benefit from the automation of lithological detection. Unmanned aerial vehicles (UAVs) are widely used in open pit mining projects, with low risk to the operators, to the aircraft or third parties. Topographic modelling using UAV imagery is now common in the mining industry. The next step, presented here, is to automate the surface feature detection using machine learning (ML) algorithms to classify a complete detailed geological model. An inexpensive aircraft was used on a Brazilian phosphate mine with point spacing as small as 10 cm.

Multi-physics data integration for interpretation and lithology classification

In recent years, the progress of data acquisition and analysis technology is remarkable in the field of non-seismic. Comprehensive evaluation of seismic data and non-seismic data is becoming important to increase success probability in exploration field. However, there are many issues to be considered for effective utilization, such as sensitivity and resolution of each data. In this paper, we introduce current status and problems of attempts to estimate structure and lithofacies by integrated analysis using multi-species geophysical data.

Towards Optimization of Boosting Models for Formation Lithology Identification

Lithology identification is an indispensable part in geological research and petroleum engineering study. In recent years, several mathematical approaches have been used to improve the accuracy of lithology classification. Based on our earlier work that assessed machine learning models on formation lithology classification, we optimize the boosting approaches to improve the classification ability of our boosting models with the data collected from the Daniudi gas field and Hangjinqi gas field. Three boosting models, namely, AdaBoost, Gradient Tree Boosting, and eXtreme Gradient Boosting, are evaluated with 5‐fold cross validation. Regularization is applied to the Gradient Tree Boosting and eXtreme Gradient Boosting to avoid overfitting. After adapting the hyperparameter tuning approach on each boosting model to optimize the parameter set, we use stacking to combine the three optimized models to improve the classification accuracy. Results suggest that the optimized stacked boosting model has better performance concerning the evaluation matrix such as precision, recall, and f1 score compared with the single optimized boosting model. Confusion matrix also shows that the stacked model has better performance in distinguishing sandstone classes.

Integrated Geophysical and Geological Mapping of the Lithological and Geological Structures at a Proposed Generator Site

The discovery of oil in commercial quantity in Ghana has led to increase in infrastructural development to provide the needed services to the oil industry. The siting of these industries requires the characterization of the subsurface to determine its suitability to host these facilities. It is in this line that seismic refraction and geo-electrical surveys were conducted on a site near Takoradi, Ghana that has been planned to host a generator and its appurtenances for the oil industry. The objective of the survey was to provide subsurface information critical for the design of earthing system for the proposed generator plant and its appurtenances. The ABEM SAS4000 Terrameter with the Lund imaging system and a 24-channel ABEM Terraloc MK6 were respectively used for the electrical and seismic refraction data collection. The seismic refraction and resistivity results identify three lithological units namely, dry gravel, sand and silty clay layer, water saturated sand and silty clay layer and granodiorites basement rocks. The first layer has a thickness of 10 m and the second layer with thickness in the range of 5 to 10 m. The lithological classifications as obtained from these results correlate with the drill logs. Electrical grounding will be suitable at a depth of 10 m beneath resistivity profile two at a distance between 32 to 52 m.

The Utility of Machine Learning in Identification of Key Geophysical and Geochemical Datasets: A Case Study in Lithological Mapping in the Central African Copper Belt

Random Forests, a supervised machine learning algorithm, provides a robust, data driven means of predicting lithology from geophysical, geochemical and remote sensing data. As an essential part of input selection, datasets are ranked in order of importance to the classification outcome. Those ranked most important provide, on average, the most decisive split between lithological classes. These rankings provide explorers with an additional line of reasoning to complement conventional, geophysical and geochemical interpretation workflows. The approach shows potential to aid in identifying important criteria for distinguishing geological map units during early stage exploration. This can assist in directing subsequent expenditure towards the acquisition and further development of datasets which will be the most productive for mapping. In this case study, we use Random Forests to classify the lithology of a project in the Central African Copper-Belt, Zambia. The project area boasts extensive magnetic, radiometric, electromagnetic and multi-element geochemical coverage but only sparse geological observations. Under various training data paradigms, Random Forests produced a series of varying but closely related lithological maps. In this study, training data were restricted to outcrop, simulating the data available at the early stages of the project. Variable ranking highlighted those datasets which were of greatest importance to the result. Both geophysical and geochemical datasets were well represented in the highest ranking variables, reinforcing the importance of access to both data types. Further analysis showed that in many cases, the importance of high ranking datasets had a plausible geological explanation, often consistent with conventional interpretation. In other cases the method provides new insights, identifying datasets which may not have been considered from the outset of a new project.

Integrated Approach for Lithological Classification Using ASTER Imagery in a Shallowly Covered Region—The Eastern Yanshan Mountain of China

This paper studied the applicability of Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) data for lithological classification in the shallowly covered Eastern Yanshan Mountain of Eastern China. The interpreted objects include quartz sandstone, carbonate rock, gneiss, and andesite. An integrated approach was employed to process the remote sensing data: First, the matched filtering method, accompanied by the ASTER library and imagery spectra as reference, was used to enhance the targeted information, namely the remotely visible (typical) rock outcrops. As the andesite is frequently weathered to iron-rich minerals in the field, so the high-resolution SPOT6 3/1 image, instead of the ASTER image, was used to locate the outcrops. Second, the fractal digital number-frequency algorithm was developed to preliminarily extract the lithology anomaly patches; and the obtained geological anomalies were generalized into three types: anomalies related to outcrops, anomalies caused by different forms of the same lithology, and random noise. Third, digital elevation model derived slope masks, in combination with the spatial intersection operation, was used to eliminate the pseudo-outcrop anomalies and noise. Finally, the accuracy assessment was conducted by referencing the local rock-outcrop database, and the misclassification rates for quartz sandstone, carbonate rock, gneiss, and andesite are 8.9%, 12.5%, 23%, and 48.3%, respectively. This study has contributed a useful case study for remote-sensing lithology mapping in shallowly covered areas, and the proposed method should have a great potential to be applied to many similar cases.

Lithofacies distribution characteristics and its controlling factors of shale in Wufeng Formation-Member 1 of Longmaxi Formation in the Jiaoshiba area

It is essential to investigate shale lithofacies distribution and controlling factor of the shale for geological evaluation of shale gas exploration and development. Through comprehensive analysis of cores, thin sections, cathode luminescence, whole-rock X-ray diffraction, element capture spectroscopy, major/trace element and other data, three major types and eight sub-type shale lithofacies in the shale of Wufeng Formation-Member 1 of Longmaxi Formation in Jiaoshiba area are identified by the three-end-member method and shale lithological classification nomenclature, and the spatiotemporal distribution law and main development controlling factors of shale lithofacies are well studied. In the Jiaoshiba area, vertically, the marine shale develops siliceous shale, mixed shale and argillaceous shale from bottom to top. Besides, lateral distribution of the shale is different from north to south; the shale lithofacies in the north area changes rapidly, the mixed shale in the north area is much thicker than that in the south area, while the siliceous shale in the south area is relatively thicker. Difference in the shale lithofacies is controlled by special sedimentary geologic events; development of the siliceous shale is controlled by the Ordovician-Silurian global volcanic event to some extent, while the mixed shale is significantly influenced by effect of bottom current, and the argillaceous shale is mainly affected by supply of terrestrial clastic material.

Lithological facies classification using deep convolutional neural network

The aim of this paper is the development of an effective model based on deep learning for geological facies classification in wells. Facies classification is carried out by studying the lithological properties of rocks, which are characteristic of modern sediments, accumulating in certain physical and geographical conditions. In this study, a new 1D-CNN model, which is trained on various optimization algorithms, is proposed. The photoelectric effect, gamma ray, resistivity logging, neutron-density porosity difference, average neutron density porosity and geologic constraining variables are considered as input data of the model. Acceptable accuracy and the use of conventional well log data are the main advantages of the proposed intellectual model. The proposed model is compared with a recurrent neural network model, a long short-term memory model, a support vector machine model, and a k-nearest neighbor model and shows more accurate results in comparison with them. The model shows successful results in the study of well log data and can, therefore, be recommended as a suitable and effective approach for well log data processing required for lithological discrimination.

A global erodibility index to represent sediment production potential of different rock types

Erosion drives landscape evolution and is fundamental for understanding the interplay between climate and tectonics as well as land-ocean matter fluxes. Large scale erosion models are an important part of the suite of Earth system models. However, their representation of the sediment production potential of different rock types is very limited. A new global erodibility index is proposed herein. It depicts the relative erosion potential of different lithological classes in the Global Lithological Map (GLiM). The erodibility index was developed by correlating observational hillslope data against climatic and tectonic setting. Hillslope data were compared for different lithological classes in regions of similar uplift across a range of landscapes. An erodibility index was defined for each lithological class, based on the relative hillslope compared to that of acid plutonic rocks in the respective area, which are here used as base lithology. The erodibility indices of different lithological classes fall within three groups: low erodibility - 1.0 for acid plutonic rocks, metamorphic rocks, and carbonate sedimentary rocks, 1.1 for acid volcanic rocks, 1.2 for mixed sedimentary rocks; medium erodibility - 1.5 for basic plutonic rocks and siliciclastic rocks of all grain sizes, 1.4 for basic volcanic rocks; high erodibility – 3.2 for unconsolidated sediments. This erodibility index is the first estimate of sediment production potential of different rock types aimed at regional to global scale application. The erodibility index presented here can help improve our understanding of global erosion and landscape evolution. A new lithology erodibility layer, the global erodibility index dataset (GEroID), is produced and used as input in a global sediment flux model (WBMsed) to demonstrate a potential utilization of the new dataset. Using an updated lithological map paired with the new erodibility index yields improved regional model results compared to the original model input.

Reservoir Lithology Determination by Hidden Markov Random Fields Based on a Gaussian Mixture Model

In this paper, geological prior information is incorporated in the classification of reservoir lithologies after the adoption of Markov random fields (MRFs). The prediction of hidden lithologies is based on measured observations, such as seismic inversion results, which are associated with the latent categorical variables, based on the assumption of Gaussian distributions. Compared with other statistical methods, such as the Gaussian mixture model or $k$ -Means, which do not take spatial relationships into account, the hidden MRFs approach can connect the same or similar lithologies horizontally while ensuring a geologically reasonable vertical ordering. It is, therefore, able to exclude randomly appearing lithologies caused by errors in the inversion. The prior information consists of a Gibbs distribution function and transition probability matrices. The Gibbs distribution connects the same or similar lithologies internally, which does not need a geological definition from the outside. The transition matrices provide preferential transitions between different lithologies, and an estimation of them implicitly depends on the depositional environments and juxtaposition rules between different lithologies. Analog cross sections from the subsurface or outcrop studies can contribute to the construction of these matrices by a simple counting procedure.

Digital geologic map and geochronologic, geochemical and geothermal database of the south-eastern part of the Sierra Madre Occidental, Mexico


 
 
 We present a first interactive digital geological map of the southeastern part of the Sierra Madre Occidental at its borders with the Mesa Central. The area of the geological cartography covers approximately 120,000 km2 that includes part of the states of Jalisco, Zacatecas, Aguascalientes and Durango. The geology has been compiled in ArcGIS through an interpretation of all the information available in the literature, integrated with our own geological mapping. Published maps were georeferenced and, as far as possible, the traces of the geologic limits and structures with a clear morphological expression were refined using digital elevation models and satellite images available in Google Earth. The map includes 15 geological units and the main tectonic structures. The informal geological units used in the map are chronostratigraphic and lithological and are designed to highlight the main magmatic episodes that shaped the region. The age assignment for each unit has been compared with a geochronological database that includes 304 ages. The lithological classification of each polygon has been compared for consistency with a geochemical database compiled from the literature that includes 313 samples. We also added a layer of information with the main hot springs of the region and their temperature.