Depth Point Research Articles

Universal bovine identification via depth data and deep metric learning

This paper proposes and evaluates, for the first time, a top-down (dorsal view), depth-only deep learning system for accurately identifying individual cattle and provides associated code, datasets, and training weights for immediate reproducibility. An increase in herd size skews the cow-to-human ratio at the farm and makes the manual monitoring of individuals more challenging. Therefore, real-time cattle identification is essential for the farms and a crucial step towards precision livestock farming. Underpinned by our previous work, this paper introduces a deep-metric learning method for cattle identification using depth data as a novel biometric measure acquired with an off-the-shelf 3D camera. In contrast to our previous work, which was limited to breeds with distinct coat patterns, this study introduces a breed-agnostic pipeline for universal cattle identification. The results show that depth, as a biometric, can potentially broaden the real-world applicability of our method to the rest 68% of UK cattle breeds that lack distinctive coat patterns. The method relies on Convolutional Neural Network (CNN) and Multi-Layered Perceptron (MLP) backbones that learn well-generalised embedding spaces from the body shape to differentiate individuals — requiring neither species-specific coat patterns nor close-up muzzle prints for operation. The network embeddings are clustered using a simple algorithm such as k-Nearest Neighbours (k-NN) for highly accurate identification, thus eliminating the need to retrain the network for enrolling new individuals. We evaluate two backbone architectures, Residual Neural Network (ResNet), as previously used to identify Holstein Friesians using RGB images, and PointNet, which is specialised to operate on 3D point clouds. We also present CowDepth2023, a new dataset containing 21,490 synchronised colour-depth image pairs of 99 cows, to evaluate the backbones. Both ResNet and PointNet architectures, which consume depth maps and point clouds, respectively, led to high accuracy that is on par with the coat pattern-based backbone. This new universal methodology also addresses the case of all-black and all-white breeds, where the previous coat pattern-based approach fell short. The ResNet colour backbone resulted in 99.97% k-NN identification accuracy, while the PointNet accuracy was 99.36%. Furthermore, we also show that the PointNet architecture is robust to noise and missing data by significantly reducing the number of 3D points and observing the drop in accuracy. Our research indicates that these techniques can identify animals using dorsal-view depth maps alone. Regardless of the substantial inter-class variety in the body shape, we show that the models spatially rely on similar body surfaces using Gradient-weighted Class Activation Mapping (Grad-CAM) and Point Cloud Saliency Mapping (PC-SM).

CAP-UDF: Learning Unsigned Distance Functions Progressively From Raw Point Clouds With Consistency-Aware Field Optimization.

Surface reconstruction for point clouds is an important task in 3D computer vision. Most of the latest methods resolve this problem by learning signed distance functions from point clouds, which are limited to reconstructing closed surfaces. Some other methods tried to represent open surfaces using unsigned distance functions (UDF) which are learned from ground truth distances. However, the learned UDF is hard to provide smooth distance fields due to the discontinuous character of point clouds. In this paper, we propose CAP-UDF, a novel method to learn consistency-aware UDF from raw point clouds. We achieve this by learning to move queries onto the surface with a field consistency constraint, where we also enable to progressively estimate a more accurate surface. Specifically, we train a neural network to gradually infer the relationship between queries and the approximated surface by searching for the moving target of queries in a dynamic way. Meanwhile, we introduce a polygonization algorithm to extract surfaces using the gradients of the learned UDF. We conduct comprehensive experiments in surface reconstruction for point clouds, real scans or depth maps, and further explore our performance in unsupervised point normal estimation, which demonstrate non-trivial improvements of CAP-UDF over the state-of-the-art methods.

Examination of the potential for geothermal energy in parts of the Benue trough, Nigeria, through the use of high-resolution aeromagnetic data

Aeromagnetic data from nine sheets covering parts of Benue Trough Nigeria were analyzed to identify potential geothermal locations. The aeromagnetic data sheets were analyzed with Oasis Montaj 8.4, Matlab 7.5, Arcmap 10.7.1, and Surfer 13 combined. The centroid depth method was used to spectrally evaluate the depth of high-resolution aeromagnetic data. The findings showed that the research area's geothermal heat flow values range from 88.52259 mW/m2 to 166.2844 mW/m2, while the Curie Point Depth values range from 8.55 km to 16.38 km with a mean depth of 12.2068 km. The Geothermal Gradient values vary from 35.40904 ℃/km to 66.51376 ℃/km with a mean value of 48.85639 ℃/km. The region's mean thermal transfer is 122.1410 mW/m2. It appears that the Curie Point Depth is highest for the lowest thermal transfer values and lowest for the highest thermal transfer values. The study area's heat flow measurements indicate that the crust is oceanic. The comprehensive geothermal data this study provides may aid the study area's exploitation of geothermal energy. Given the extremely high heat flow recorded, there is a likelihood that the studied area has geothermal energy sources. As a result, it was observed that practically the entire study region had substantial heat flow (>80 mW/m2), suggesting the possibility of a geothermal energy source.

CONSTRUCTION OF EARTH CRUST SECTIONS DOWN TO THE MOHO BOUNDARY BASED ON LOW AMPLITUDE REFLECTED WAVES OF RIVER SEISMIC SURVEY USING CPD-2D METHOD (Vitim River, junction zone of the Angara-Lena monocline of the Siberian platform and the Bodaibo-Patom folded system)

In 2019, employees of the Research and Production Enterprise Luch jointly with specialists of the Seismological Branch of the Geophysical Survey of the Russian Academy of Sciences (GS RAS) carried out river seismic exploration works using the common depth point method (CDP-2D) along a 170 km long profile in the lower reaches of the Vitim River. The research was carried out according to the original methodology developed in the Siberian branches of the GS RAS, using the “Baikal” equipment recording seismic signals continuously (in contrast to traditional cable bottom or ground-based systems with time-limited recording). In this case, patented air guns “Malysh” were used in the water for the excitation of elastic waves, which have been recorded by ground registration method on the riverbank. To this day, the archives store primary materials from each receiving point (about 7000, with a distance of 25 m between them) in the form of a continuous hours-long digital recording of seismic signals and noise. These materials make it possible to generate seismograms of significant duration – up to 23 s, equal to the time interval between wave generations, in contrast to 6–10 second seismograms traditionally used to construct sections of the upper part of the Earth’s crust. In addition, the 24-bit Baikal recorder makes it possible to record signals with an amplitude two orders of magnitude lower than the amplitude of seismic noise. On the seismic sections constructed by the processor at arrival times of up to 13–14 sec, due to an increase in the stack fold of signals up to 1000–2000, we can select low-amplitude waves, reflected from boundaries in the middle and lower parts of the Earth’s crust to the Moho boundary. High fold is achieved by increasing the size of the stacking site (bin) by several times. By processing seismograms, generated from archival materials, for the first time, we have a complete vertical section of the Earth’s crust along a 170 km long profile, using low-amplitude signals. The profile passes through the junction zone of the Angara–Lena monocline and the Bodaibo-Patom folded system. The proposed approach can be used to obtain preliminary (and relatively cheap) information about the deep structure of the Earth’s crust along profiles performed using the CDP-2D river seismic exploration method.

Estimating the vertical profile of water quality variables in reservoirs: Application of remotely sensed data and machine learning techniques

Water quality assessment and management of reservoirs depend on accurate, large-scale, and continuous monitoring of the vertical profile of Water Quality Variables (WQVs). Remote sensing data have been widely used to retrieve high spatiotemporal water quality data; however, their application has practically been limited to evaluating surface WQVs. In this paper, a novel and efficient approach is introduced for assessing the profile of WQVs in reservoirs that depend on stratification, by taking into account the shape of profile as prior knowledge. First, an appropriate function is fitted to the WQVs vertical profile, and second, the parameters of that function as representative of the WQVs vertical profile are estimated using machine learning techniques. The model's inputs are day, maximum depth of point, and remote sensing data. Finally, PAWN sensitivity analysis is applied to show the extent to which each input influences different parts of the vertical profile. This method is applied in the Wadi Dayqah Reservoir, the largest dam in Oman, to evaluate water temperature, dissolved oxygen, pH, and chlorophyll-a profile. The results show that the predicted profiles are properly representative of in situ measurements, with a mean absolute error of 0.28 °C, 0.25 mg/L, 0.052, and 0.33 μg/L on test data sets of water temperature, dissolved oxygen, pH, and chlorophyll-a, respectively. Finally, PAWN sensitivity analysis illustrates that satellite data not only influence the parameters representing surface WQVs but also contribute to the estimation of other curve parameters.

Gas content, geochemical characteristics and implications of coalbed methane from the Deep Area of Qi’Nan Coalmine in Huaibei Coalfield

The objective of this work is to investigate the implications of geological influence factors on gas content and geochemical characteristics of deep-buried (> 800 m) coalbed methane (CBM) reservoirs. Results show that bituminous coal accounts for the majority, which exhibits similar maturity but differ in maceral and chemical constituents. CBM reservoirs show low porosity, low permeability and moderate temperature, with thickness of 0.85–4.15 m. In addition, the total gas content is 4.58–12.33 m3/t (average of 8.83 m3/t). CH4 is the main component with concentration of 92.83−99.22% (average of 96.68%), and the δ13CCH4 and δ13DCH4 is − 53.78‰–−44.62‰ (average of − 48.82‰) and − 223.93‰––−4.49‰ (average of − 215.37‰), respectively. All CBM samples are the mixtures of thermogenic gases and secondary biogenic gases with CO2 reduction. In addition, gas content characteristic at the critical point of burial depth is the result by positive and negative geological effects. CH4 concentration shows a wide range with the increases of buried depth, while the numerical values of which for the selected samples display complex variation characteristics. Furthermore, the values of δ13CCH4 and δ13DCH4 become heavier with the increases of buried depth. Besides, the above two geochemical parameters are related to Ro, max and reservoir temperature.

Automatic Conditioning of Marine Seismic Angle Stack Data With a Convolutional Neural Network

Misalignment of reflections is a common problem in migrated prestack seismic data, which occurs due to moveout correction with incomplete knowledge of the velocity model. This issue persists in partial angle stack data often used for amplitude versus offset (AVO) analyses, and is traditionally treated with residual moveout and trim static corrections. However, these methods require time-consuming velocity selection and parameter tuning, and commonly lead to spurious alignment of reflections with noise. Therefore, we train a convolutional neural network (CNN) on synthetic examples to automatically perform conditioning and alignment of angle stack data. First, two-dimensional synthetic common depth point (CDP) gathers are created using a convolutional modeling method, and then made into input-target pairs where the inputs are poorly moveout-corrected (using inaccurate normal moveout velocities) and the targets are accurately moveout-corrected. These input-target examples are split into angle stacks for the near, mid, and far offsets, and then used to train a CNN to transform the misaligned angle stacks into their more-aligned counterparts. In testing, the trained network increases the alignment of unseen synthetic data, improving the correlation with the target far offset trace from 0.76 to 0.85 on average. The network is applied on two field examples from offshore Norway with Class 3 and Class 2P AVO responses, respectively. In both cases, the angle stack data predicted by the network shows improved alignment, greater resolution in the far offset stacked section, and increased correspondence to a well-tie synthetic seismogram, all while retaining the AVO responses. The CNN predictions are compared to angle stack data following traditional conditioning (residual moveout and trim static corrections), with the results being of similar or greater quality. This method enables high quality conditioning of angle stack data at a fraction of the time required for conventional methods, and is easily adaptable to different datasets.

The improved Moho depth imaging in the Arabia-Eurasia collision zone: A machine learning approach integrating seismic observations and satellite gravity data

The Arabia-Eurasia convergences created one of the earth's topographic highs on the Central Tethys collisional belt. Despite the area's geological significance, a comprehensive and high-resolution map of Moho depth has been lacking due to the sparse and uneven distribution of seismically constrained Moho depth data. This study addresses this deficiency by compiling an extensive dataset of nearly 2500 seismically measured Moho depth points from 68 seismic local scale studies, resulting in the development of an updated seismically constrained Moho depth model (S-Moho) at a 0.5° × 0.5° spatial resolution. Despite some coverage gaps in the remote areas, the S-Moho model offers a more detailed view than previously available. To further improve the coverage of the S-Moho depth model, an incremental data-driven approach was employed. Initially, a gravity-based regression Moho depth model (SB-Moho) was developed by correlating S-Moho depth points with corresponding Bouguer anomalies. However, its accuracy was constrained by unaccounted isostatic and non-isostatic components. To address this limitation, a sliding window approach was applied to derive a windowed SB-Moho model (WSB-Moho). Additionally, a machine learning-based Moho model (ML-Moho) was developed using seismic Moho depth points along with 11 predictive variables. Both WSB-Moho and ML-Moho models demonstrated consistent and smooth Moho depth variations. The Zagros region reveals a prominent NW-SE oriented Moho depression (45-60 km thick), attributed to the underthrusting of the Arabian Plate beneath the Iranian Plateau. The models suggest that crustal thickening extends beyond tectonic boundaries, likely influenced by the dip of suture zones. In contrast, the crustal thickening in eastern Anatolia, northwest of the Zagros, is less pronounced, indicating different geodynamic processes. Strike-slip faulting and magmatic activity in this area contribute to a broader distribution of deformation compared to the more localized crustal thickening in the Zagros. In southeastern Zagros, strike-slip faults in central Iran accommodate much of the northward convergence of the Arabian Plate, thereby limiting the extent of crustal thickening.

Automatic detection of multi-antibiotic residues simultaneously in aquatic products by visual protein microarray chips

A 96-well microplate-based protein microarray chip which can simultaneously detect three different families of antibiotics using a multiplexed approach has been developed for the first time. All steps in the microarray chip can be completed through an automated biochip analyzer, enabling high throughput analysis of 96 samples within 60 min according to the set program. The microarray chip can be visually evaluated by the color depth of array points, and can also be quantified through a scanner. Under the optimal conditions, the limits of detection (LODs) of nitrofuran metabolites (NFMs), chloramphenicol (CAP), and fluoroquinolones (FQs) were 0.11–0.15 μg/kg, 0.07 μg/kg and 6.42 μg/kg, respectively. This microarray chip can be applied to different aquatic products, including crucian carp, shrimp, and scallops, with the recovery ranged from 89.4 % to 111.2 %. These results demonstrate new strategies for the high-throughput detection of multi-antibiotic residues and present potential of protein chip technologies for daily supervision.

Enhancing quadrotor robustness control using image-based visual servoing (IBVS) with fuzzy logic

Visual servoing is a commonly employed approach in robotics and unmanned aerial vehicles (UAVs) that facilitates accurate object positioning and movement control by utilizing visual feedback. As quadrotors gain popularity, more control methods have been developed. This article presents a method for enhancing quadrotor robustness control using image-based visual servoing (IBVS) with fuzzy logic. Unlike traditional visual servoing, which relies on a fixed gain and often encounters challenges with velocity convergence and maintaining the object in the field of view, this method is designed to enhance the visual servo control of quadrotors by dynamically adjusting the gain of the IBVS system through a fuzzy logic controller. This controller adaptively adjusts the servo gain in response to feature errors and the depth of the object's feature points. MATLAB simulations clearly demonstrate the superior performance of this fuzzy logic integrated method compared to classical approaches, showcasing enhanced control capabilities in challenging environments.

Identification of prospective oil-bearing areas in the peripheral parts of productive horizons U-12 and U-13 of the Zhetybai field

Background: To date, a comprehensive range of scientific, geological exploration, and appraisal work has been conducted at the Zhetybai field. The overall structural map of the field has been relatively well established. However, several critical issues remain insufficiently studied or entirely unexplored. Based on the results of seismic exploration (3D common depth point method) and additional appraisal work, peripheral (edge) zones of the field structure have been identified, including its cross-faults, as well as transverse flexures and longitudinal faults on the southern wing of the structure, which require further investigation. Aim: To further study the field structure, identify new potential hydrocarbon areas in the peripheral parts of the field, and assess the oil potential of the U-12 and U-13 horizons in the Zhetybai field. Materials and methods: This study involved a qualitative assessment of the informativeness of seismic attributes, analysis of reflection geometry (configuration), dynamic amplitude parameters, continuity of frequency, and other indicators. The analysis of reflection characteristics, in conjunction with all available data, primarily geophysical survey data, allowed for hypothesizing sedimentary conditions and obtaining acceptable lithological assessments. Results: The evaluation wells J-1, J-2, and J-3, drilled in 2023-2024 in the peripheral parts beyond the approved boundary of hydrocarbon-bearing horizons U-12 and U-13, yielded positive results. The productivity of the new reservoir was also confirmed by newly drilled production wells 5333, 5652, and 5367. Conclusion: To further study the structure of the Zhetybai field and identify new potential hydrocarbon-bearing areas in the peripheral parts, as well as structural uplifts and sandstone bodies of horizons U-12 and U-13, evaluation wells J-1, J-2, and J-3 were drilled. The detailed identification of sandstone bodies allows for the prediction of prospective areas not covered by drilling. The positive results from drilling evaluation wells have refined the geological structure of horizons U-12 and U-13 and confirmed the productivity of the peripheral parts of the reservoirs.

3D Head Pose Estimation via Normal Maps: A Generalized Solution for Depth Image, Point Cloud, and Mesh

Head pose estimation plays a crucial role in various applications, including human–machine interaction, autonomous driving systems, and 3D reconstruction. Current methods address the problem primarily from a 2D perspective, which limits the efficient utilization of 3D information. Herein, a novel approach, called pose orientation‐aware network (POANet), which leverages normal maps for orientation information embedding, providing abundant and robust head pose information, is introduced. POANet incorporates the axial signal perception module and the rotation matrix perception module, these lightweight modules make the approach achieve state‐of‐the‐art (SOTA) performance with few computational costs. This method can directly analyze various topological 3D data without extensive preprocessing. For depth images, POANet outperforms existing methods on the Biwi Kinect head pose dataset, reducing the mean absolute error (MAE) by ≈30% compared to the SOTA methods. POANet is the first method to perform rigid head registration in a landmark‐free manner. It also incorporates few‐shot learning capabilities and achieves an MAE of about on the Headspace dataset. These features make POANet a superior alternative to traditional generalized Procrustes analysis for mesh data processing, offering enhanced convenience for human phenotype studies.

TAG-fusion: Two-stage attention guided multi-modal fusion network for semantic segmentation

In the current research, leveraging auxiliary modalities, such as depth information or point cloud information, to improve RGB semantic segmentation has shown significant potential. However, existing methods mainly use convolutional modules for aggregating features from auxiliary modalities, thereby lacking sufficient exploitation of long-range dependencies. Moreover, fusion strategies are typically limited to singular approaches. In this paper, we propose a transformer-based multimodal fusion framework to better utilize auxiliary modalities for enhancing semantic segmentation results. Specifically, we employ a dual-stream architecture for extracting features from RGB and auxiliary modalities, respectively. We incorporate both early fusion and deep feature fusion techniques. At each layer, we introduce mixed attention mechanisms to leverage features from other modalities, guiding and enhancing the current modality's features before propagating them to the subsequent stage of feature extraction. After the extraction of features from different modalities, we employ an enhanced cross-attention mechanism for feature interaction, followed by channel fusion to obtain the final semantic features. Subsequently, we provide separate supervision to the network on the RGB stream, auxiliary stream, and fusion stream to facilitate the learning of representations for different modalities. The experimental results demonstrate that our framework exhibits superior performance across diverse modalities. Specifically, our approach achieves state-of-the-art results on the NYU Depth V2, SUN-RGBD, DELIVER and MFNet datasets.

INNOVATIVE EXPERIMENTAL TESTING PROGRAM OF DIRECT SHEAR TEST IN SOIL MECHANICS

The research work aims at analyzing for the first time the data set obtained on cohesive soil samples following the publication of the Romanian Invention Patent RO 134239. The standard test method for the direct shear test provides the shear strength parameter – internal friction angle in consolidated drained condition - of either undisturbed or remolded soil samples forcing the shear plane at the midsection of the sample in the horizontal direction. The samples are provided in parallelepipedal shape (6 cm x 6 cm x 2 cm) and the displacement rate in horizontal direction is 0.1 mm/min. The new equipment patented in Romania changes the direction of shearing, from horizontal to vertical, and the soil samples are of cubic shape (6 cm x 6 cm x 6 cm). The experimental program involves testing both the parallelepipedal and cubic samples using the same motorized mechanism, with simultaneous readings from their respective micro-comparators. The UU test is performed without allowing consolidation and shearing at 1.0 mm/min. For the CD test, samples are consolidated under vertical loads for 24 hours before shearing at 0.1 mm/min. The shear stresses for cubic samples were higher than those for parallelepipedal samples, with residual stresses reflecting this trend. For cubic samples, both the peak and residual shear stresses trend lines indicated higher cohesion (c) and lower internal friction angle (𝜙) for UU tests and CD tests in contrast to parallelepipedal samples in both testing conditions. The innovative testing program allows for variability in shear strength parameters along the soil failure surface in both natural and compacted soil structures. This differentiation divides the soil condition into drained and undrained states at the initiation, emergence points, and the point of maximum depth along the failure surface. This approach is significant for accurately assessing soil shear resistance and potential failure mechanisms. The study's findings suggest a nuanced approach to parameter selection for slope stability analysis, ensuring accurate representation of both cohesion and internal friction in stability models.

Development and Application of a Liquid Sampling Device with Timing, Quantitative and Depth Determination

This thesis describes the development process of a liquid sampling device that can be timed, quantitative, and deep. The structural composition, working principle, and innovative features of the liquid sampling device are introduced, and the application prospects are analyzed. The device has the function of controllable and adjustable sampling point depth and sampling volume, and adopts quantitative collection and collection methods without contact with accessories. The device has a silent function in the sample body, which can realize pre-booking of collection time points and multi-point synchronous collection. Wireless remote control automatic mode is adopted for sending and stopping collection instructions, which can be sent one-to-one or one-to-many at the same time. It provides automation links for water quality monitoring projects and meets the requirements of modern scientific and technological development.

A Novel Method for Extracting DBH and Crown Base Height in Forests Using Small Motion Clips

The diameter at breast height (DBH) and crown base height (CBH) are important indicators in forest surveys. To enhance the accuracy and convenience of DBH and CBH extraction for standing trees, a method based on understory small motion clips (a series of images captured with slight viewpoint changes) has been proposed. Histogram equalization and quadtree uniformization algorithms are employed to extract image features, improving the consistency of feature extraction. Additionally, the accuracy of depth map construction and point cloud reconstruction is improved by minimizing the variance cost function. Six 20 m × 20 m square sample plots were selected to verify the effectiveness of the method. Depth maps and point clouds of the sample plots were reconstructed from small motion clips, and the DBH and CBH of standing trees were extracted using a pinhole imaging model. The results indicated that the root mean square error (RMSE) for DBH extraction ranged from 0.60 cm to 1.18 cm, with relative errors ranging from 1.81% to 5.42%. Similarly, the RMSE for CBH extraction ranged from 0.08 m to 0.21 m, with relative errors ranging from 1.97% to 5.58%. These results meet the accuracy standards required for forest surveys. The proposed method enhances the efficiency of extracting tree structural parameters in close-range photogrammetry (CRP) for forestry. A rapid and accurate method for DBH and CBH extraction is provided by this method, laying the foundation for subsequent forest resource management and monitoring.

Numerical prediction of ground-borne vibrations due to continuous vibratory driving of circular closed-ended piles

Various types of pile driving work may induce high-intensity ground-borne vibrations. Predicting vibration intensity before adopting mitigation measures is vital for minimizing the impact of vibration on nearby structures and occupants. Vibratory pile driving is a commonly applied foundation construction method. However, numerical simulation models for ground vibrations during a complete process of vibratory driving have rarely been studied. This study introduces an axisymmetric finite element model that utilizes the arbitrary Lagrangian-Eulerian technique to simulate the continuous vibratory driving of a circular closed-ended pile penetrating from the ground surface to a target depth. The model validity was confirmed by assessing the calculated ground vibrations against the findings documented in earlier research. The results showed that the critical penetration depth of piles, at which the maximum peak particle velocity (PPV) occurs, varied with the radial distance and depth of points of interest, contradicting a common preconception. Moreover, the maximum PPV did not always occur on the ground surface across all radial distances. Parametric analysis revealed that an increase in the soil cohesion strength, pile diameter, or soil-pile friction, or a decrease in the driving frequency or soil damping ratio would increase ground vibrations due to vibratory pile driving.

Study on the impact of camera ISP algorithms on stripe structured light 3D reconstruction and the improvement of 3D reconstruction accuracy by photon input

This paper investigates the impact of camera image signal processing (ISP) algorithms on stripe structured light 3D reconstruction and explores the improvement of 3D reconstruction accuracy by photon input. Existing 3D reconstruction technologies have broad applications in fields such as intelligent manufacturing, healthcare, and consumer electronics, but their high precision requirements often cannot be met by current ISP algorithms. By using handheld devices to capture images and combining key techniques such as the multi-step phase shifting method, multi-frequency phase unwrapping method, and triangulation method, this paper conducts an in-depth study on the application of photon input in 3D reconstruction. The study demonstrates that using non-visual information (RAW images) as input can significantly improve reconstruction accuracy, producing more accurate results compared to images processed by ISP. The paper also quantifies the impact of ISP processing on 3D reconstruction results by comparing the depth information and point cloud data of two sets of images. Experimental results show that disabling certain ISP algorithms, such as bilateral noise filtering (BNF), edge enhancement (EEH), and non-local means denoising (NLM), can further improve reconstruction accuracy and reduce errors. In conclusion, this paper proposes a photon image-based 3D reconstruction method that, combined with artificial intelligence technology and differentiable point cloud rendering techniques, holds promise for achieving higher precision and faster 3D reconstruction. This technology is of great significance in practical applications, particularly in the field of industrial close-range 3D reconstruction.

Resurgent dome and super-hot enhanced geothermal system: The Sahinkalesi Massif within the Hasandag Stratovolcanic Province, Central Anatolia, Turkey

The Sahinkalesi, a volcanic dome located NNE of Hasandağ, Türkiye exhibits anomalous heat flow value, geothermal gradient and the Curie point depth is located at very shallow depth in this region. Our investigation indicates presence of super-critical thermal regime (378°C) at about 4 km depth and the MT analysis indicate shallow magma chamber at about 5 km depth. The crust is relatively thin below this region with the low-velocity region located at depth of about 36 km. Thermo-Hydro-mechanical model investigation has been carried out using finite element discretization technique. For faulted zone reservoir models, 30 years of geothermal energy exploitation does not cause thermal breakthrough for mass flow rates up to 500 kg/s, however, the mean stress developed in the reservoir becomes much larger and may be unsustainable for the reservoir stability. To ensure the success of a fractured reservoir model, the use of multiple wellbores is recommended. In the case of a closed-loop geothermal system, the primary concern is the control of thermoelastic stress. This can be achieved either by increasing the wellbore depth while reducing the injection mass flow rate, or by extending the wellbore's horizontal component. The outlet temperature in both the cases maintained at 275°C. This is the first time a superhot EGS site has been identified in Türkiye.

Improving Accuracy and Efficiency of Monocular Depth Estimation in Power Grid Environments Using Point Cloud Optimization and Knowledge Distillation

In the context of distribution networks, constructing 3D point cloud maps is crucial, particularly for UAV navigation and path planning tasks. Methods that utilize reflections from surfaces, such as laser and structured light, to obtain depth point clouds for surface modeling and environmental depth estimation are already quite mature in some professional scenarios. However, acquiring dense and accurate depth information typically requires very high costs. In contrast, monocular image-based depth estimation methods do not require relatively expensive equipment and specialized personnel, making them available for a wider range of applications. To achieve high precision and efficiency in UAV distribution networks, inspired by knowledge distillation, we employ a teacher–student architecture to enable efficient inference. This approach maintains high-quality depth estimation while optimizing the point cloud to obtain more precise results. In this paper, we propose KD-MonoRec, which integrates knowledge distillation into the semi-supervised MonoRec framework for UAV distribution networks. Our method demonstrates excellent performance on the KITTI dataset and performs well in collected distribution network environments.