Image Logs Research Articles

Examples of Paleokarst in Mesozoic Carbonate Formations in the Carpathian Foreland Area

A paleokarst system developed in the Upper Jurassic–Lower Cretaceous carbonate complex has been recognized in the Carpathian Foreland area. Well logs testing and core data as well as an acoustic imager, a microresistivity scanner and six-arm dipmeter images were used to identify and analyze the character of the paleokarst features. A detailed interpretation of microresistivity and acoustic image logs allowed for the identification of different types of karst forms, such as caverns; multidirectional fractures, including fractures widened by dissolution; and the type of sediments filling them. The analysis of the seismic survey was conducted by linking the paleokarst characteristic features recognized in the seismic image to the karst intervals determined from borehole data. The set of seismic attributes calculated from the analyzed 3D seismic data, including the RMS amplitude, instantaneous frequency, consistent dip, variance, sweetness and relative acoustic impedance, helped to delineate the zones of the paleokarst distribution. Within the interpreted paleokarst surface developed in the carbonate formations in the study area, there are sinkholes, limestone pavements and valleys. Furthermore, in the northwestern part of the analyzed area, the development of paleokarst forms is related to the presence of a relatively deep branch of a paleovalley formed in the Paleogene, as well as to numerous discontinuities developed in carbonate formations. The development of this type of larger karst form was probably controlled primarily by tectonic faults. The research conducted by the authors of this paper showed the widespread presence of paleokarst features in Upper Jurassic–Lower Cretaceous carbonate formations in the study area. A good spatial identification of the paleokarst surface can be important in a regional context, since the highest part of the profile of carbonate formations is the most important reservoir for geothermal or hydrocarbon resources in this region.

Natural fractures at depth in shale reservoirs: new insights from the southern Sichuan Basin marine shales

Natural fractures are pervasive in southern Sichuan Basin marine shales, China, and provide a useful opportunity to try to understand subsurface fracture networks in shale reservoirs. Based on cores and electrical imaging logs from vertical and horizontal petroleum wells in the southern Sichuan Basin, four types of natural fractures were identified in terms of orientation, size, filling properties and spatial distribution. Uncemented bed-parallel shear fractures develop at or in the vicinity of mechanical interfaces and are inclined to present in shale layers with a dip angle greater than 12°. Cemented bed-parallel fractures are characterized by a crack-seal texture marked by multiple bands of fibrous cement, and their intensity decreases upwards and shows a positive relationship with total organic carbon (TOC) values. Uncemented bed-oblique fractures are a type of fracture that rarely develops, and accommodate limited open space. Cemented bed-oblique/perpendicular fractures are the most developed fracture type and are distributed on a regional scale and are subdivided into two types. The results imply that these shale fractures could be formed sequentially by local and regional tectonic deformation, and by abnormally high pressure. Most natural fractures cannot contribute to reservoir storage or efficiently enhance its permeability, yet can act as planes of weakness and be potentially reactivated during hydraulic fracture treatments.

Technology Focus: Formation Evaluation (February 2024)

The scientific research process begins as one tries to find explanations for a phenomenon. We make observations, define the problem statement, and review the existing domains of research that could be used. Another approach is to explore theoretical problems, those that are purely conceptual at present but provide a solution when a related observation is made in the future. Though these approaches sound isolated, both are part of characterizing uncertainty, and uncertainty comes in all scales and dimensions. This challenges us to learn at all scales possible, from the fume hoods in the laboratory to magnificently exposed outcrops and through deep narrow boreholes that drill through subsurface reservoirs. The combined efforts often convert learnings to actionable intelligence. At a smaller scale, porosity and permeability are probably the two most-studied rock properties among those that have meaningful implications for hydrocarbon reservoirs. Paper SPE 216856 considers machine-learning (ML) methods for classifying reservoir texture at a microscale. Borehole-image logs long have been used to obtain a picture of subsurface reservoirs. Unfortunately, a majority of the observations are qualitative. Quantifying these features faces the challenge of continuity, upscaling, and regional correlation. As we explore the latitude of ML-based applications, the use of these techniques for quantifying image logs becomes very relevant. The authors of that paper contribute to quantifying textural features at a “fume-hood scale” and develop a work flow with the potential for estimating properties such as porosity and permeability from a different domain of reservoir characterization. I often wonder how much the domain on formation evaluation encompasses. While geoscience-driven reservoir characterization is a big part of it, how reservoirs change over time also is a complementary observation. Paper URTeC 3864861 discusses various aspects of geomechanical changes that a hydraulically fractured reservoir goes through during its life cycle. The authors here study the relationship between measured strain from the fiber-optic sensors and wellhead pressure. Research like this could be extended to predicting production profiles and estimating recovery factors, which are important considerations in designing a stimulation program for sustaining production, maximizing recovery, and improving financial matrices for the capital program. I believe information could be categorized as learning, knowledge, and intelligence. Any scientific process starts with set of careful observations bound by an envelope of hypotheses. This is learning. Learning, which could be verified by predictable and repeatable outcomes from carefully designed experiments of complementing domains, becomes knowledge. Actionable knowledge, which then could be used to alter an outcome or a process, becomes intelligence. Paper URTeC 3871303 discusses a development strategy in a restrictive development unit with an existing parent well. Here, considerations are heavily weighted toward optimizing both interwell spacing and capital efficiency. The search for answers to a problem like this must seek guidance from a variable-scale experiment. The study here establishes the big picture with the structural elements of the basin that could restrict both the continuity of the reservoir and the nature of the producible fluid. With this framework, the model is then set to iterate from several different perspectives. Potential interwell communications are explored by measuring fracture-driven interactions (FDIs) and quantifying stimulated reservoir volume. What is impressive here is the different domains from which the authors seek answers. Direct observations from acoustic-fiber measurements for FDI and the geochemistry of produced fluids for identifying unique signatures from vertically separated formations are individual domains that seek the same answers in various scales. The study recommends the optimal spacing between wells and a stimulation design that minimizes well interference, reduces competition for resources between wells, and avoids overcapitalizing the program. This is how knowledge transforms into intelligence. I hope the readers appreciate the scales of characterization in these three papers. As a student of geology, I have always been fascinated by the concept of scale and its relation to the domains of science that we deal with. Unlike general relativity and quantum mechanics, most geologic phenomena are observed in all scales. It is just the uncertainty that needs to be quantified.

Parameter evaluation method of tight carbonate reservoir using electrical imaging pores diameter spectrum

Currently, high-yield gas reservoirs have been discovered by exploring Lower Paleozoic tight carbonate oil and gas reservoirs in the Changqing area. These reservoirs, however, are impacted by subsequent karstification and diagenesis processes, leading to the formation of complex pore spaces with varying pore diameters ranging from microns to millimeters. Conventional well curves, with a resolution at the decimeter level, can only provide an overview of the reservoir porosity, making it challenging to evaluate the pore structure. On the other hand, electrical imaging logging data has a sampling interval of 0.1 inches and a resolution as fine as 5 mm, providing valuable information for pore structure evaluation in tight reservoirs. Since formation resistivity is influenced by factors like porosity, mud filtrate resistivity, and cementation, the cementation index serves as a reflection of the pore structure. Therefore, by calibrating the cementation index using data obtained from experiments at different scales, the pore structure parameters of tight carbonate reservoirs can be evaluated. This study conducted a comparison between numerous core experiments at various scales and both conventional and imaging log data. A method was proposed for extracting pore size distribution characteristics and calculating the pore size spectrum based on electrical imaging logging, ultimately leading to the establishment of a quantitative evaluation model for the pore structure of tight carbonate reservoirs. The results demonstrated improved accuracy and consistency in interpretation, as confirmed by well test data.

DDViT: Advancing lithology identification on FMI image logs through a dual modal transformer model with less information drop

Lithology is an essential topic in oil and gas reservoir studies. Lithological observation lays the foundation for assessing oil and gas prospects and guides future exploration and development. Currently, the prevailing approach in lithological observation heavily depends on the manual analysis of core samples. However, such an approach is highly subjective and time-consuming. With the development of deep learning, some automated deep learning-based methods have been proposed for lithology interpretation from logging curves. However, the Fullbore Formation MicroImager (FMI) image logging, while widely used in the oil field exploration and development process, is occasionally deployed and utilized for lithology identification. In this work, we proposed a Dual-modal Drop-less-information Vision Transformer (DDViT), an FMI image lithology identification model based on transformer architecture. DDViT uses a dual-modal architecture to identify lithology using two different image modalities, namely dynamic FMI images (FMI_DYN) and static FMI images (FMI_STAT). These modalities reflect the local and overall characteristics, respectively. Furthermore, DDViT uses a less-information dropping module to drop the blank band information inherent in the FMI images to make our model more rational and stable. DDViT achieved a 90.81% lithology identification accuracy on Fengxi Well A of the western Qaidam Basin, providing a new approach to lithology identification and demonstrating the great potential of deep learning in geological images.

Convolutional neural networks for quality and species sorting of roundwood with image and numerical data

Roundwood sorting is still a manual process in many Swiss sawmills, requiring employees to visually inspect and categorize thousands of logs per day. The heavy workload can be both physically and mentally taxing and can lead to increased rates of human error. State-of-the-art automation systems like X-ray log scanners are expensive and difficult to integrate into existing process lines. This paper proposes a novel recommendation system that leverages recent advances in image classification to automate roundwood classification by quality and species. The system integrates a camera to capture cross-sectional images of logs and record numerical data, such as length, taper, and diameter. The analysis of the resulting dataset highlights the challenges of data imbalance and noise, which makes classification difficult and, in some cases, impossible. However, by using selected datasets with reduced noise, state-of-the-art Convolutional Neural Networks (CNNs) can extract quality and species features. Quality models learn from a manually selected and simplified dataset, featuring samples that experts can clearly classify based on the image’s information. Species models are trained on a label-noise-reduced dataset, reflecting real-world complexity. The accuracy on the selected dataset for three quality classes is 80%. The species determination is less challenging and reaches 91% accuracy on a synchronized dataset for the main species spruce and fir. Overall, this paper highlights the potential of Machine Learning in augmenting the roundwood sorting processes and presents a novel system that can improve the efficiency and accuracy of the process.

High-fidelity source mask optimization for suppressing line-end shortening.

Source mask optimization (SMO) is a widely used computational lithography technique for compensating lithographic distortion. However, line-end shortening is still a key factor that cannot be easily corrected and affects the image fidelity of lithography at advanced nodes. This paper proposes a source mask optimization method that suppresses line-end shortening and improves lithography fidelity. An adaptive hybrid weight method is employed to increase the weights of the line end during the optimization, which adaptively updates the weights in each iteration according to the edge placement error (EPE). A cost function containing a penalty term based on the normalized image log slope (NILS) is established to ensure the fidelity of the overall feature when paying more attention to the line-end region. The scope of this penalty term is regulated by widening and extending the split contour to further reduce the line-end shortening. Simulation results show that the proposed method can effectively suppress the line-end shortening and improve the lithography fidelity compared with the traditional SMO method.

Reconstructing the Zama (Mexico) discovery source to sink palaeogeography, Part II: Sediment routing from the Late Miocene shelf‐margin to deepwater basin

AbstractThe Late Miocene source terrane tectonic history in the southern Gulf of Mexico Basin, as informed by detrital zircon geothermochronology data, supports a detailed regional palaeogeographic reconstruction from palaeoshoreline to the deepwater Zama minibasin of the Sureste salt basin. Seismic mapping points to a trio of pathways that converge upon two entry points into the Zama minibasin, illuminating how sediment gravity flows transit a complex seascape defined by shallow salt bodies. Consideration of empirical scaling relationships within and between segments of this sediment dispersal system allows for testable predictions of Upper Miocene submarine fan‐runout lengths over basin exploration areas. Distances from the reconstructed shelf‐margin to the Zama wells vary around 100 km, an increase of 20% over a straight‐line distance as flows likely navigated around extant salt stocks, walls and sheets. This 100‐km fan length is about 40% of the reconstructed minimum palaeo‐river length, within predicted ranges for smaller source‐to‐sink systems in tectonically active areas (25 to 50%). The estimated fan‐runout distance can be extended even further basinwards, considering the contemporaneous passage of the mobile Chortis block along the Tonala shear zone, expanding the Palaeo‐Rio Grijalva drainage network during the Tortonian. These Late Miocene deepwater systems linked to the Palaeo‐Rio Grijalva differ substantially from onshore Mexico‐sourced turbidity flows feeding into the axis of the north‐trending Veracruz Trough. Textural data from wells here suggests these systems were less effective at larger grain transport and sorting. Local (intrabasinal) variations are also evident within the Zama minibasin, as well data (image logs and cores) indicate that axially oriented sediment gravity flows involved fewer high‐density turbidities, depositing lower net‐to‐gross sandstones and thicker shales than those flowing transverse to the basin axis from a southeastern basin entry point. These interpretations will guide both local exploitation of these economic resources and could also support future exploration for analogous salt‐influenced deepwater reservoir systems in the Sureste basin and globally.

The present-day in-situ stress field and its effect on shale gas development in Zigong area of southern Sichuan Basin

The Zigong area in southern Sichuan Basin is one of the key shale gas production areas in China, and the Longmaxi and Wufeng Formations within it are currently the key exploration and development layers. The present-day in-situ stress field has a significant impact on well trajectory deployment, horizontal well construction, hydraulic fracturing and other aspects. However, it has not been finely quantified in shale gas reservoirs of Zigong area, making it difficult to effectively guide development practice. Therefore, this study constructed a geomechanical model of the target layer in the Zigong area, quantitatively characterized the distribution of present-day in-situ stress, and explored its shale gas development effects. The results show that: (1) Based on the analysis of drilling induced fractures obtained from rock acoustic emission experiments and imaging logging interpretation, the dominant orientation of the maximum horizontal principal stress in the target layer is mainly NW-SE-trending; (2) The horizontal principal stress difference in the Layer S1l1-1−1 and Wufeng Formation is 5–30 MPa (the majority is 6–15 MPa) and 5–35 MPa (the majority is 6–18 MPa), respectively. Low stress difference values are mainly concentrated in the southeastern and northern parts. The overall prediction accuracy is high, with an error rate of less than 8% for the Layer S1l1-1−1 and less than 10% for the Wufeng Formation; (3) Based on rock acoustic emission experiments and finite element model simulation results, the study area is mainly under strike-slip stress faulting mechanism; (4) Due to the influence of the strike-slip faulting stress mechanism, when deploying horizontal wells in the area, priority should be given to selecting areas with low stress mechanism factor (Aφ) and low stress differences, to achieve better fracturing and transformation effects and reduce the risk of casing deformation and wellbore instability.

Low-Cost Quantized Compressed Sensing and Transmission Method for Ultrasonic Imaging Logging

Low-Cost Quantized Compressed Sensing and Transmission Method for Ultrasonic Imaging Logging

库车坳陷白垩系巴什基奇克组成像测井沉积学研究

库车坳陷白垩系巴什基奇克组成像测井沉积学研究

Numerical Simulation and Characteristic Analysis of Remote Detection Acoustic Reflection Imaging Logging of Near-Borehole Fracture and Vuggy Reservoirs

Numerical Simulation and Characteristic Analysis of Remote Detection Acoustic Reflection Imaging Logging of Near-Borehole Fracture and Vuggy Reservoirs

Interpretation of Mud Losses in Carbonates Based on Cuttings Description, Well-Logging, Seismic and Coherency Data

Hartha Formation is an overburdened horizon in the X-oilfield which generates a lot of Non-Productive Time (NPT) associated with drilling mud losses. This study has been conducted to investigate the loss events in this formation as well as to provide geological interpretations based on datasets from nine wells in this field of interest. The interpretation was based on different analyses including wireline logs, cuttings descriptions, image logs, and analog data. Seismic and coherency data were also used to formulate the geological interpretations and calibrate that with the loss events of the Hartha Fm. The results revealed that the upper part of the Hartha Fm. was identified as an interval capable of creating potential mud losses, resulting in high NPT. This is due to its diagenetic features such as succrosic dolomites and vuggy zones that could act as thief zones. Seismic potential was used for the prediction of the geological related non-productive drilling time in the Hartha interval. The seismic data quality in this interval was good, with geological observations already made. Detailed interpretation and analysis of the Hartha interval were performed and integrated with the existing seismic interpretation, rock properties, and NPT database to calibrate wells with the loss events to the seismic observations.

A new method for quantitative evaluation of shale laminae using electrical image logging

Shale oil reservoirs are generally characterized by complex mineral compositions, rapid lithofacies changes, and thin laminae. Explorations have confirmed that the type and density of shale laminae significantly influence reservoir quality, highlighting the importance of accurately identifying these laminae through well logging for effective shale reservoir evaluation. Presently, relevant technologies primarily focus on the qualitative identification of shale laminae using vertical slab images from image logs. However, influenced by the complex borehole conditions and image logging quality, this approach is less effective in identifying millimeter-scale laminae. This study proposes a new method for achieving high-resolution slab images and quantitatively evaluating the laminae using electrical image logs. The new method effectively improves the processing accuracy of slab images by delicately flattening and aligning the button electrode curves derived from electrical image logs point by point. Meanwhile, it allows for the accurate quantitative evaluation of the lamina number through precise identification of peaks and troughs in microelectrode curves. As demonstrated by the applications in shale oil reservoirs in the Gulong area in Daqing and the Ganchagou area in Qinghai, the proposed method can significantly improve accuracy compared to traditional slab images. Furthermore, the lamination index calculated using this method is highly consistent with the lamina number observed in cores. This study provides a new technical method for the quantitative lamina evaluation and rock structure analysis of shale reservoirs.

Geophysical exploration methods on bauxite reservoirs in L gasfield, Ordos Basin

Introduction: The L gasfield on the eastern edge of the Ordos Basin mainly focuses on the exploration and development of coalbed methane and tight sandstone gas. As the overlying strata of Majiagou Formation of Lower Ordovician in Lower Paleozoic, bauxite has always been considered a regional cap rock and has not been widely followed with interest. The drilled wells has revealed that bauxite can be used as a unconventional reservoir with good reservoir properties. In the future, it will become a new natural gas exploration target in the L region. This article is dedicated to the comprehensive study of geophysical methods for the bauxite exploration in the L region.Methods: Firstly, we use seismic data to restore paleo-geomorphology of the Lower Paleozoic. Secondly, we analyze the well log data of different paleo-geomorphic unit. Then, we establish a mineral model based on the logging data to calculate the lithological composition and porosity, and use imaging logging to evaluate the pore structure and fluid type.Result and Discussion: Finally, we obtained favorable reservoir development zones for bauxite, guiding exploration and evaluation of bauxite.

Contribution of mine borehole data toward high-resolution stress mapping: An example from northern Bowen Basin, Australia

Most of in-situ stress data in the Australian continent comes from wellbore stress analysis in deep hydrocarbon reservoirs, and earthquake focal mechanism solutions near the Australian plate boundaries, where geophysical tools facilitate understanding of the present-day stress patterns. This resulted in a paucity of stress information in many other regions such as the northern Bowen Basin, which is an active mining province, but with low seismicity rates and limited deep petroleum exploration. The mining industry runs several hundred kilometres of image logs annually to characterise geotechnical attributes. These logs provide an image from the borehole wall, which facilitates analysis of stress-related borehole deformations for in-situ stress characterisation. This paper examines the orientation of horizontal in-situ stress using different types of image logs in mine boreholes across the northern Bowen Basin. Analyses of 128 km of image logs in 680 vertical boreholes resulted in the interpretation of 9046 pairs of stress-related indicators including 735 drilling induced fractures and 8311 borehole breakouts. Our comprehensive database comprises 890 quality-ranked data records for the orientation of maximum horizontal stress (SHmax), which makes the Bowen Basin as a basin with the highest data density in the world in terms of quality-ranked stress information according to the World Stress Map. Statistical analysis of SHmax orientation reveals that the mean SHmax orientation in northern Bowen Basin is N018° ± 16°. The results show that this orientation is consistent over long distances, which is in contrast with several eastern Australian basins. This uniform stress pattern agrees well with plate-scale geomechanical model predictions, which further highlights the impact of plate boundary forces in the contemporary stress pattern of this region. Detailed image log investigation did not show any systematic rotation of stress; however, some small-scale stress perturbations were observed in the vicinity of sharp stiffness contrasts and geological structures.

Fracture identification and characteristics of carbonate underground gas storage: an example from the eastern area of Sulige gas field, ordos Basin, China

The carbonate rock formations have obvious dual media characteristics, fracture development and good physical conditions, which are the main seepage channels and storage spaces for gas after the reconstruction of underground gas storage. The carbonate strata of the Ordovician system are important natural gas reservoirs in the eastern area of Sulige Gas Field in the Ordos Basin, and the identification and characterization of their fractures are of great significance for the modeling of fractures in the later stage and the improvement of the operation scheme of the gas storage. At present, there is little research on fractures, which restricts exploration and development. Therefore, taking the 39–61 gas storage reservoir in the eastern area of Sulige Gas Field in the Ordos Basin as the research object, this paper identifies and studies the characteristics of the fractures by core, microscopic, conventional logging curves, and imaging logging identification. The results show that the fracture length ranges from 5 to 15 cm and the width ranges from 0.1 to 3 mm. The fracture angles are mostly between 75° and 90° and the main direction is NW–SE. In conventional logging curves, porosity logging has a good response to fractures, while resistivity logging has a general response to fractures; In layers with more developed fractures, natural gamma values are mostly higher than 40API, rock volume density is less than 2.8 g/cm3, neutron porosity is greater than 12.5%, and acoustic time difference is greater than 160 μ s/m. This study is of great significance for improving the identification of carbonate fractures, enriching the relevant theories, and providing guidance for the construction of carbonate gas storage.

ANALYSIS OF CEMENTING WELLS X-1 and Y-1 USING MULTIFUNCTION ULTRASONIC IMAGING LOGGING TOOL

By analyzing the measurement principle of the ultrasonic pulse method, a multifunctional ultrasonic image recording device and the application of the measurement technique were designed. It can perform surface imaging well in high-resolution imaging mode in open holes. Can also perform casing inspection and cement bond evaluation in full wave mode at casing holes. The proposed tool was tested in practice. This shows that this tool can evaluate geological feature information such as fractures and layered structures in open holes. Imaging casing thickness, inner and outer diameters and cement impedance can also be measured at casing holes. Cement analysis in wells Logging Tool (MUIL), cement slurry design and operations during cementing. The logging results above can be analyzed quantitatively including amplitude readings which can then determine compressive strength prices and bond index prices. Meanwhile, qualitative analysis was carried out comprehensively from VDL recordings and using the Enhanced Logging Imaging System. For analysis from a design perspective, it looks at the cement slurry planning, while from an operational perspective it looks at the suitability between planning and implementation in the field.

Assessment of Fracture Density Distribution from Image Logs for Sensitivity Analysis in the Asmari Fractured Reservoir

Characterizing fracture properties in naturally fractured reservoirs poses a significant challenge. While welltesting remains valuable, it often fails to provide precise descriptions of these properties. Bridging this gap requires the integration of geological expertise to enhance fracture assessment. This study addresses the limitations of well-test analysis and explores the application of Conventional Image Logs in structural, fracture, and geomechanical analysis. However, effectively combining these applications with well-test analysis on a field scale reveals a substantial knowledge gap. A critical challenge in this context is the absence of a defined procedure for calculating the variable "σ," a crucial parameter for simulating fractured carbonate reservoirs using image log fracture density. Integrating geological knowledge is essential to reduce uncertainties associated with well-test analysis and provide more accurate characterizations of fracture properties. Image log data processing emerges as a valuable avenue for gaining insights into the static attributes of naturally fractured reservoirs. This study focuses on Characterizing fractures using data from ten image logs and Developing a more accurate simulation model through the interpretation of images, with a particular emphasis on OBM imaging. The main goals of this fracture study revolve around establishing correlations between fracture densities well by well within the simulation and enhancing the accuracy of the simulation model by incorporating fracture data from image logs. Borehole imaging tools such as FMI/FMS and OBMI-UBI play a pivotal role in identifying significant structural features, including faults, fractures, and bedding. Fine-tuning fracture parameters during the history matching process, while potentially time-consuming, significantly impacts other historical match parameters. Consequently, the reliability of reservoir simulation results, predictions, and recovery enhancement strategies hinges on the precision of fracture properties and their distribution within the model. Recent advances in interpretation techniques have expanded the horizons of image interpretation, enabling the creation of more accurate simulation models for fractured reservoirs using fracture data obtained from image logs. The overarching goal of this project is to comprehensively evaluate a fractured reservoir field by integrating data from ten individual wells.

A deep-learning framework for borehole formation properties prediction using heterogeneous well-logging data: A case study of a carbonate reservoir in the Gaoshiti-Moxi area, Sichuan Basin, China

The properties of borehole formations, such as porosity, permeability, and water saturation, play a crucial role in characterizing and evaluating subsurface reservoirs. Although core sample experiments offer precise measurements, they are time consuming and cost intensive. An alternative method is to use the logging data to construct an empirical model that predicts formation properties, which is widely studied due to its speed and affordability. Nevertheless, because the response of a logging point reflects its surrounding formation, conventional logging methods relying on point-to-point (P2P) mapping perform poorly in complex reservoirs. Furthermore, the resolution of conventional logging is lower compared with imaging logging. To address these limitations, this study presents a novel approach to predict formation properties based on a deep-learning framework using heterogeneous well-logging data. Our neural network framework takes short sequences of conventional logging data and windowed imaging logging data as inputs. The neural network applies 1D convolution to extract features from the conventional logging sequences and 2D convolution to extract features from the resistivity imaging data. Then, these two feature vectors are fused and fed into a multilayer fully connected neural network to predict formation properties. A case study of a carbonate reservoir demonstrates that our method delivers more accurate predictions of formation porosity, permeability, and water saturation than the P2P, sequence-to-point, and image-to-point prediction methods. Moreover, it is expected that our paradigm will serve as a source of inspiration for forthcoming research endeavors aimed at enhancing the accuracy of predicting borehole formation properties in complex reservoirs.