EAGE (European Association of Geoscientists and Engineers)’s Post

100% of the technical/special topic articles in this month’s First Break use rainbows for data visualization. It would be helpful if the First Break “Guidelines for authors” contained recommendations to use scientific colourmaps for accurate, intuitive and accessible data visualization so we can easily and accurately interpret the published figures. See “The misuse of colour in science communication” Crameri et al, 2020 for more information: https://lnkd.in/eT7QfCqx EAGE (European Association of Geoscientists and Engineers) #FirstBreak #EAGE #usescientificcolourmaps #changethedefault #useviridis #usebatlow #usecmocean #appliedgeosciences #oceanography #hydrology #geology #geophysics #gpr #seismic #hydrogeology #environmentalgeosciences #sciencecommunication #gravity #magnetics #geotechnical

MOTIVO: Revolutionizing Seismic Data Interpretation   WAVERITY introduces MOTIVO, an automated interpretation stream poised to transform traditional interpretation workflows into robust AI-based solutions. This innovation meets the urgent demand for rapid and accurate seismic data analysis, providing geoscientists with an organized, data-centric approach that results in substantial cost savings and improved exploration and production outcomes. Today we are presenting the first solution of the MOTIVO stream – AI-based 3D seismic clusters. At its core is the idea of using a full frequency spectrum to identify and cluster similar seismic patterns, which can then be combined into geological facies by the interpreter. This approach allows for the construction of robust geologic architectural element distribution maps, distinctively delineates the target, and calculates the volumetrics of the potential hydrocarbon reservoir.   Join Us at IMAGE 2024 For those interested in a deeper understanding of this groundbreaking methodology, WAVERITY will present the paper "3D Seismic Facies Clustering through Spectral Decomposition Using Unsupervised ML" at IMAGE 2024 - International Meeting for Applied Geoscience & Energy. Join us for the poster presentation in the session "Machine Learning Applications 3" (Session ID: SP P18) on August 29, 2024, from 10:20 AM to 12:00 PM. Visit our website to learn more about this product: https://lnkd.in/dWT6KbX2 #WAVERITY #MOTIVO #SeismicData #AI #Geoscience #SeismicAnalysis #DataInterpretation #ExplorationAndProduction #EnergyTech #TechInnovation #CostSavings #HydrocarbonExploration #SeismicClusters #DataDriven #GeologicalMapping # #FutureOfEnergy #TechSolutions

🌟 Exciting News! RadiXplore 2.0 is now live and packed with powerful new features! 🚀 I'm thrilled to announce the release of RadiXplore 2.0, featuring cutting-edge tools designed to transform your document analysis experience. Here's a glimpse of what's new: 🔍 **RadixEye:** Utilize our #GeoAI vision model to search and extract images and tables from documents effortlessly. 🖼️ **RadixEyeGPT:**Describe an image to find it or use an image to search for similar images across millions of PDF reports. 💬 **RadixGPT:** Engage in a question-and-answer session with your documents using our Language Model (LLM). To demonstrate the potential of RadiXplore 2.0, we tasked RadixGPT with analyzing Ontario Geological Survey | Commission géologique de l’Ontarios' GeologyOntario Reports to identify rocks and minerals associated with Lithium in Ontario. The results were impressive! RadixGPT provided accurate answers, pinpointed the source documents, and even mapped out their geospatial locations. Ready to explore the future of document analysis? Sign up for free on our website and unlock the full potential of RadiXplore 2.0 today! #AI #DocumentAnalysis #RadiXplore #GeoAI #MachineLearning #DataAnalysis #Technology #Innovation #radixploremining #radixploreusecase

How to Handle Non-Uniformly Spaced Data Using NUFFT and Sinc Interpolation in fields like MRI and geophysics. Integrating NUFFT and Sinc Interpolation, our new blog explores advanced techniques for handling non-uniformly spaced data in fields like MRI and geophysics. By simulating non-uniform sampling and regridding the data onto a uniform grid, you can accurately process and reconstruct the signal. The combination of NUFFT and ResampleSINC provides a powerful solution, allowing direct handling of non-uniform data without interpolation while ensuring minimal distortion in signal reconstruction. If you want to learn more about these cutting-edge methods and how they can transform your signal processing workflows, keep reading! #SignalProcessing #NUFFT #NonUniformSampling #SincInterpolation #MRIProcessing #DataReconstruction #QuakeLogic #SeismicData #EngineeringResearch #TechInnovation #DigitalSignalProcessing

Confirmed speaker! Frederico Melo will talk about “Unlocking the Equatorial Margin with AI-based reconstructed seismic”. Seismic acquisition is one of the most important and effective tools for deriving insights during early exploration activities. It gathers regional understanding of the subsurface, which is intended to map and identify potential plays in new venture areas. Frontier exploration studies typically employ 2D seismic acquisition experiments designed to sparsely cover a large area of exploration interest before committing to larger upfront financial investments, such as 3D seismic acquisition and processing. Therefore, efforts to reduce the gap between sparse 2D and 3D seismic experiments are valuable, as plausible regional 3D seismic images of the subsurface can be effectively used to extract regional insights in areas where only 2D seismic images are available. Deep learning (DL) is a promising tool for this application due to its unique ability to handle nonlinear optimization without using pre-designed kernels and its capability of incorporating arbitrary a priori subsurface information. In this work, we developed a novel multi-directional learning algorithm to convert multiple sparse 2D seismic images into a full 3D seismic volume. The structural information is extracted from the 2D images and subsequently fed into the evolving 3D image volume through a proxy model and a DL network.The Brazilian Society of Geophysics - SBGf presents the second edition of the Workshop: The South America Equatorial Margin — Exploration for a Sustainable Energy Transition, which will take place on July 2nd and 3rd, 2024, at Petrobras/CENPES, Rio de Janeiro. More information and registrations on the website. https://lnkd.in/dfHCiV8e

How to Handle Non-Uniformly Spaced Data Using NUFFT and Sinc Interpolation in fields like MRI and geophysics. Integrating NUFFT and Sinc Interpolation, our new blog explores advanced techniques for handling non-uniformly spaced data in fields like MRI and geophysics. By simulating non-uniform sampling and regridding the data onto a uniform grid, you can accurately process and reconstruct the signal. The combination of NUFFT and ResampleSINC provides a powerful solution, allowing direct handling of non-uniform data without interpolation while ensuring minimal distortion in signal reconstruction. If you want to learn more about these cutting-edge methods and how they can transform your signal processing workflows, keep reading! https://lnkd.in/eWkja54a #SignalProcessing #NUFFT #NonUniformSampling #SincInterpolation #MRIProcessing #DataReconstruction #QuakeLogic #SeismicData #EngineeringResearch #TechInnovation #DigitalSignalProcessing

Quantitative Interpretation using OpendTect Predicting rock properties from seismic data has been a cornerstone expertise of dGB Earth Sciences since 1995. Starting as a QI specialist company, we pioneered unique workflows based on stochastic forward modeling of pseudo-wells and non-linear (neural network-based) predictions. Our services have significantly expanded since then, but our passion for QI projects remains unwavering. The attached video shows some of the highlights that are available in OpendTect. Our QI Services Include: ⭐️ Rock-physics analysis ⭐️ AvO attribute analysis ⭐️ Forward modeling of synthetic data ⭐️ Seismic inversion products ⭐️ Rock-property predictions ⭐️ Integrated interpretation services Integrated interpretation services OpendTect's Inversion & Rock Physics Bundle offers a comprehensive suite of tools for innovative QI work, including the unique HitCube inversion approach supported by SynthRock in which a stochastic simulator generates thousands of 1D models (pseudo-wells), and the synthetics of these pseudo-wells are cross-matched against real seismic data to create 3D probability volumes of rock properties. Understanding geology is crucial for drilling successful wells. We predict the 3D distribution of lithologies, rock properties, and fluid fill from 3D seismic data, integrating control points at well locations. We enhance reservoir prediction by leveraging geological models and neural networks, avoiding oversimplified linear models. 🔧 Reservoir Characterization Services: ⭐️ Well Modeling Products: Log-log predictions, Lithology clustering, Pattern analysis, Sensitivity modeling, Stochastic pseudo-well modeling, AVO analysis, rock physics. ⭐️ Inversion Products: HitCube, Seismic Spectral Blueing, Clustered/classified seismic, Impedance inversion, HorizonCube-based geo-models, HorizonCube-based low frequency models, Rock properties prediction. ⭐️ Additional Services: Full 3D calibration to well data, 4D time-lapse studies (monitoring saturation changes). If you would like to know more, please contact us: info@dgbes.com #OpendTect #Geoscience #QuantitativeInterpretation #Seismic #RockPhysics #ReservoirCharacterization #InnovationInGeology #NeuralNetworks #SeismicInversion #GeologicalModeling #AI #ML

𝐂𝐨𝐦𝐩𝐚𝐫𝐢𝐬𝐨𝐧 𝐁𝐞𝐭𝐰𝐞𝐞𝐧 𝐋𝐞𝐚𝐬𝐭-𝐒𝐪𝐮𝐚𝐫𝐞𝐬 𝐑𝐞𝐯𝐞𝐫𝐬𝐞 𝐓𝐢𝐦𝐞 𝐌𝐢𝐠𝐫𝐚𝐭𝐢𝐨𝐧 𝐚𝐧𝐝 𝐅𝐮𝐥𝐥-𝐖𝐚𝐯𝐞𝐟𝐨𝐫𝐦 𝐈𝐧𝐯𝐞𝐫𝐬𝐢𝐨𝐧 #SeismicDataProcessing 𝑷𝒂𝒓𝒕-1 (Least Squares in Terms of Kirchhoff and Reverse Time Migration) Inversion is a very common technique in seismic processing, consisting of creating a model that can predict acquired data. This approach has three main components, the model m, the data d, and the operator L that connects them (d = Lm). Choosing the kind of model defines the operator. In seismic we use many types of models. For example, in Full Waveform inversion (FWI) we use some function of velocity, like slowness square. In Least squares migration (LSMIG) we use a model that resembles reflectivity (Kirchhoff modeling), or velocity perturbations (scattering or Born modeling) (Fig.1) by #TGS. We minimize the residual square or average difference between data and predictions. This minimization always proceeds in some form of optimization, which depends on whether the operator represents a linear transformation (the operator is independent of the model we seek), or non-linear (the operator changes as we improve our estimation of the model). Generally, linear operators are used in forward geophysical modelling calculations. A common task is to find the inverse of these calculations, i.e., given the observed data, find a model that can predict these data. The LSM mentioned above is applicable for linear problems. However, recorded seismic data depend nonlinearly on earth parameters. FWI is a nonlinear inversion method which uses all information in the seismogram to get the earth model. 𝐋𝐞𝐚𝐬𝐭 𝐒𝐪𝐮𝐚𝐫𝐞𝐬 𝐦𝐢𝐠𝐫𝐚𝐭𝐢𝐨𝐧 𝐢𝐧 𝐓𝐞𝐫𝐦𝐬 𝐨𝐟 𝐌𝐢𝐠𝐫𝐚𝐭𝐢𝐨𝐧 To understand the differences of LSMIG and migration we have to look at the Hessian. Therefore, Mls is closer to the true model because it removes the blurring effect of the Hessian. The Hessian is often diagonal dominant, so the real improvement achieved by LSMIG depends on how large the off-diagonal terms of the Hessian are. 𝐈𝐦𝐩𝐥𝐞𝐦𝐞𝐧𝐭𝐚𝐭𝐢𝐨𝐧𝐬 𝐅𝐨𝐫 𝐊𝐢𝐫𝐜𝐡𝐡𝐨𝐟𝐟 𝐚𝐧𝐝 𝐑𝐞𝐯𝐞𝐫𝐬𝐞 𝐓𝐢𝐦𝐞 𝐌𝐢𝐠𝐫𝐚𝐭𝐢𝐨𝐧 We will use the same framework to design both LSK and LSRTM as a linearized inversion of forward-adjoint operator pairs. Since the inversion formulation is identical, all differences are in the modeling operators, which leads to different degrees of mismatch between physics and data. 𝐀 𝐂𝐨𝐦𝐩𝐚𝐫𝐢𝐬𝐨𝐧 𝐁𝐞𝐭𝐰𝐞𝐞𝐧 𝐋𝐒 𝐊𝐢𝐫𝐜𝐡𝐡𝐨𝐟𝐟 𝐚𝐧𝐝 𝐋𝐒𝐑𝐓𝐌 We will look at results from LSMIG for both RTM and Kirchhoff. In both cases, we will use a similar implementation with adjoint-pairs that pass the adjoint test with similar precision. (Fig. 2.1) shows an RTM for the Marmousi model (A complex 2D structural model used to compare depth-migration and velocity determination models.) and (Fig. 2.2) shows the LSRTM result after 9 iterations. Smoothing the velocity model (Fig. 3.1 and 3.2).

In this exciting bit of #research led by Haorui Peng, we present a new #deeplearning based approach to #seismic #inversion for #subsurface impedance in the image-depth domain. This is similar to what you may be used to in state-of-the-art impedance inversion, but instead accounts for multidimensional image blurring to #acquisition footprint, #processing and #illumination effects. The novelty here is the use of the invertible Recurrent Inference Machine (i-RIM) #machinelearning architecture - a type of #loopunrolled #network - applied to the multidimensional deblurring of seismic images. By employing i-RIMs, we construct a learning and #inference framework that is informed by a known forward operator that captures the #physics of seismic imaging, and the acquisition and processing characteristics of the data. Our approach outperforms widely-used U-Nets, particularly in yielding robust, successful inferences in out-of-distribution data. This is only the beginning for this line of work - so stay tuned. There are many exciting applications from #advancedprocessing, #fullwaveform imaging, through their use in #monitoring for various #energytransition scenarios. For now, enjoy this great #openaccess article by Haorui.

I'm very excited about the results we are achieving through Nimbus Insights, applying advanced data science and remote sensing techniques to geospatial data. In this analysis, we combined the results of a principal component analysis applied to surface reflectance data from the Harmonized Sentinel-2 MSI collection with elevation data from Alos Palsar (Hi-Res Terrain Corrected). Through meticulous analysis of the eigenvector matrix, factor loadings, and the images produced by the newly identified factors, we have developed color compositions that elucidate various characteristics of the study area. This analysis enables the extraction of structural, mineralogical, textural, and geomorphological information, proving to be very useful for structural mapping, geological studies, and mineral exploration, especially in areas with sparse vegetation. The analyzed area is located in Iran, in the Hormozgan province, located in the south of Iran, known for its coasts and islands in the Persian Gulf, as well as its historical and cultural significance. The chosen area encompasses the ancient Amir mine and the chromite occurrences of Kuh-E-Sorkh. I would like to extend my gratitude to my colleagues at Nimbus Insights, Murillo Costa and Ramon Arouca Jr., for their collaboration in this significant endeavor of applying technological advancements to geosciences. Interactive 3D models of our findings are hosted in the cloud and can be accessed through the links provided below. For demonstration, we have selected three color compositions and one RGB image. RGB 3D Model Link: https://p3d.in/BxHOr Color Composite 3D Model 1: https://p3d.in/xuJNT Color Composite 3D Model 2: https://p3d.in/3SR5n Color Composite 3D Model 3: https://p3d.in/Br29S