Search

Photogrammetry systems are widely used in industrial manufacturing applications as an assistance measurement tool. Not only does it provide high-precision feedback for assembly process inspection and product quality assessment, but also it can improve the flexibility and robustness of manufacturing systems and production lines. However, with growing global competition and demands, companies are forced to enhance production efficiency, shorten production lifecycle and increase product variety by incorporating reconfigurable factory design that can meet challenging timeline and requirements. Although dynamic facility layout is widely investigated, the position selection for the photogrammetry system in dynamic manufacturing environment is usually overlooked. In this paper, dynamic layout of the V-STARS photogrammetry system is investigated and optimised in a digital-twin environment using deep reinforcement learning. The learning objectives are derived from the field of view (FoV) evaluation from point clouds 3D reconstruction, and collision detection from the digital twin simulated in Visual Components. The application feasibility of the proposed dynamic layout optimisation of the V-STARS photogrammetry system is verified with a real world industrial application. [ABSTRACT FROM AUTHOR]

3D modeling of wear particles has proven to be a useful tool for monitoring mechanical failure conditions. In this work, a new method for 3D reconstruction of wear particles in uncontaminated oil (healthy oil) and contaminated oil (used oil) was proposed. The image acquisition device can capture multi-view images of moving wear particles in both healthy and used oil by using the reflected light. The images were pretreated first, and the image color inversion was conducted using the Pillow library. The pretreated wear particle images were used for 3D reconstruction using long short-term memory 3D recurrent reconstruction neural network. The current results were verified against existing results, and good agreement can be found. It can be concluded that we can reconstruct the similar 3D wear particle results with fewer images by comparison with other methods. Specifically, only 4–6 image samples were used for the 3D reconstruction of wear particles, and at least 8 image samples were needed for other existing reports. [ABSTRACT FROM AUTHOR]

To improve the accuracy of photogrammetric joint roughness coefficient (JRC) estimation, this study proposes two optimization models based on ground sample distance (GSD), point density, and the root mean square error (RMSE) of checkpoints. First, an algorithm that automatically generates spatial positions for equipment based on the convergence strategy was developed, using principles of Structure from Motion and Multi-View Stereo (SfM-MVS) and the shooting parameter selection algorithm (SPSA). Second, a portable positioning plate containing ground control points and checkpoints was designed based on optical principles, and a moving camera capture strategy guided by SPSA was proposed. Combining SPSA, portable positioning plate, and moving camera capture strategy, a photogrammetric experiment for small-scale rock samples in the field was conducted, collecting 48 datasets with different shooting parameters. Subsequently, a dataset incorporating GSD, point density, RMSE, and three JRC estimation metrics was established, revealing their correlations and sensitivities. Using seven machine learning algorithms, optimization models for photogrammetric JRC accuracy were developed, with Linear Multidimensional Regression and Gaussian Process Regression models improving JRC accuracy by an average of 85.73%. Finally, the applicability and limitations of the newly proposed method were further discussed. [ABSTRACT FROM AUTHOR]

With the rapid development of artificial intelligence and computer vision, numerous technologies have been introduced to automate manufacturing in the industry. Typical metal workpieces in the industry often have highly reflective surfaces, come in various sizes, and are positioned irregularly. The motor rotor presented in this paper is one such representative workpiece. Traditional grasping methods for workpiece loading and unloading are pre-programmed and often struggle to cope with complex and disordered situations. In this paper, we introduce a structured light (SL) sensor as the visual guide for the triaxial robot. Furthermore, we propose a high-precision hand-eye calibration method for the non-orthogonal coordinate system of the triaxial robot. Additionally, a motor rotor center localization method based on U-Net image segmentation is proposed. By combining the high-precision hand-eye calibration and localization, we can accurately and automatically locate and grasp the rotor. We have conducted sufficient experiments to verify the effectiveness and accuracy of our system. [ABSTRACT FROM AUTHOR]

The co‐generation of electricity and chemicals via direct internal reforming solid oxide fuel cells (DIR‐SOFCs) offers a promising route to carbon‐neutral energy solutions. However, challenges such as inadequate performance and fast degradation, particularly when using hydrocarbon fuels like CH4, hinder the deployment of DIR‐SOFC technology. This study addresses three critical issues: the effect of fuel composition on electrochemical properties, the mechanisms and microstructural impacts of carbon deposition, and the practical feasibility of DIR‐SOFCs at an industrial scale. First, comprehensive polarization and impedance analyses are conducted to assess the impact of varying fuel compositions—specifically p(CH4) and p(H2O)—on DIR‐SOFC performance. Second, advanced morphological characterization, machine learning‐assisted 3D reconstructions, and numerical simulations are utilized to reveal carbon deposition behavior and its effects on anode microstructures. Quantitative analysis of carbon's impact on pores, Ni, and YSZ phases provides novel insights into carbon‐induced microstructural changes. Finally, the industrial‐scale co‐generation of electricity and chemicals is validated, emphasizing both energy efficiency and operational stability. This study enhances the understanding of electrochemical and microstructural mechanisms, offering crucial insights for optimizing DIR‐SOFC design and operation, and laying the groundwork for their broader adoption in a carbon‐neutral future. [ABSTRACT FROM AUTHOR]

This article mainly studies the element content, phase composition, and fragmentation mechanism of SiO2–CaO–Al2O3–MgO composite inclusions that cause fatigue fracture of spring steel. The study found that the SiO2–CaO–Al2O3–MgO composite inclusions in the wire rod are not a homogeneous phase. According to the electron microscopy energy spectrum analysis results, although the SiO2–CaO–Al2O3–MgO composite inclusion is in the low‐melting‐point area of the phase diagram according to the component content ratio. However, field‐emission transmission electron microscopy analysis reveals that this type of composite inclusions includes MgAl2O4, CaAl2Si2O8, CaSiO3, MgSiO3, and a small amount of CaS. The Zeiss Crossbeam 550 electron microscope is used to conduct a three‐dimensional reconstruction analysis of this type of composite inclusions, which once again verified that the composition of this type of inclusions includes calcium silicate, MgAl2O4, MgO, and other phases. It is also found that there is a certain gap between the inclusions and the matrix, and the gap between the inclusions with high‐melting‐points such as MgAl2O4 and MgO and the matrix is more obvious. During the rolling process of SiO2–CaO–Al2O3–MgO composite inclusions, due to differences in hardness and deformation rate between different phases, the rolling process is deformed and fragmented for extended periods. [ABSTRACT FROM AUTHOR]

NeRF can render photorealistic 3D scenes. It is widely used in virtual reality, autonomous driving, game development and other fields, and quickly becomes one of the most popular technologies in the field of 3D reconstruction. NeRF generates a realistic 3D scene by emitting light from the camera's spatial coordinates and viewpoint, passing through the scene and calculating the view seen from the viewpoint. However, when the brightness of the original input image is low, it is difficult to recover the scene. Inspired by the ambient illumination in the Phong model of computer graphics, it is assumed that the final rendered image is the product of scene color and ambient illumination. In this paper, we employ Multi-Layer Perceptron (MLP) network to train the ambient illumination tensor I , which is multiplied by the color predicted by NeRF to render images with normal illumination. Furthermore, we use tiny-cuda-nn as a backbone network to simplify the proposed network structure and greatly improve the training speed. Additionally, a new loss function is introduced to achieve a better image quality under low illumination conditions. The experimental results demonstrate the efficiency of the proposed method in enhancing low-light scene images compared with other state-of-the-art methods, with an overall average of PSNR: 20.53 , SSIM: 0.785, and LPIPS: 0.258 on the LOM dataset. [ABSTRACT FROM AUTHOR]

Three typical types of spherical fruits, namely apples, oranges, and pears, have accounted for 40.86 % of the total fruit production in China in 2023. However, there is a huge trade deficit in the import and export trade of fruits in China. Furthermore, grading technology has limited the largest production and consumption of fruits. Appearance indicators, such as size, volume and deformity index, are the main external quality traits of fruits in commercial grading lines. Alternatively, computer vision has been widely used in the field of nondestructive detection of fruits and vegetables, due to the simple operation, non-contact and low cost. The surface information can be obtained to measure the external indicators using two- dimensional images. However, some limitations are still remained to capture the detailed appearance indicators, especially in the high-accuracy measurement. Compared with traditional two-dimensional images, three-dimensional (3D) imaging with the depth information can be beneficial to the appearance indicators of fruits. Among them, binocular structured light is the promising 3D reconstruction with the cost-saving, high-precision and non-contact extraction. This study focuses on spherical fruits, selecting apple, orange, and pear as representative examples, with the aim of constructing binocular structured light imaging to obtain surface contour information. The system was also validated using standard samples. Single-view point cloud reconstruction was performed as well. Three-step phase shifting was used to determine the wrapped phase. While multi- frequency heterodyne was applied to determine the absolute phase. Afterwards, stereo matching, disparity optimization, and point cloud calculation were conducted to obtain a 3D point cloud map. A rotary table was utilized to match the point cloud images from different perspectives to the same coordinate axis. Coarse point cloud registration was also achieved. A complete point cloud was obtained to conduct more precise registration using the iterative closure point (ICP) to obtain. Finally, the 3D reconstruction of spherical fruits was realized. The deformity index of the apple was extracted from the reconstruction image of single-view point cloud using normal vector angle. According to the normal vector angle, the point cloud of the fruit stem position was projected to the bottom of the fruit point cloud image. The complete fruit point cloud was improved to calculate the convex hull of the point cloud for the volume of fruit. The maximum fruit diameter was extracted through oriented bounding box (OBB) bounding box. The experimental results show that the relative error was within 4% in the external dimensions of standard parts using 3D reconstruction. Taking manual measurement as the references, the values of coefficient of determination (R²), root mean squared error (RMSE), and mean absolute percentage error (MAPE) were 0.97, 0.755 mm, and 7.23%, respectively, for the measured apple deformity index. In the volume of spherical fruits, R² was 0.99, RMSE was 6.015 cm³, and MAPE was 1.946%, while those were 0.92, 1.823 mm, and 1.859%, respectively, for the maximum diameter. The 3D reconstruction with binocular structured light can be expected to significantly enhance the accuracy and efficiency in the appearance indicators of spherical fruits. The finding can also provide a valuable tool to improve fruit quality control and grading. [ABSTRACT FROM AUTHOR]

Computer vision techniques were employed to monitor the displacement of retaining walls using artificial markers, traditional feature detection algorithms, and photogrammetry‐based point cloud reconstruction. However, the use of artificial markers often increases both installation time and costs, whereas the performance of traditional feature matching is affected by uneven illumination, and photogrammetry techniques require multiple images for point cloud reconstruction. To overcome these limitations, a nontarget‐based displacement monitoring method for segmental retaining walls (SRWs) using a combination of deep learning and stereovision was proposed. Binocular stereovision was employed to reconstruct the geometry and surface properties of the SRW in a digital three‐dimensional (3D) model. Deep learning models were then used to extract natural features from SRW blocks, enabling displacement calculation without using artificial targets. The performance was evaluated by monitoring the behaviors of SRW experiments at both laboratory and field scales. The deep learning–based image segmentation models identified SRW block features in the experiment and real case datasets with an average F1 score from 0.910 to 0.965 under various environmental conditions. The reconstructed results of point cloud coordinates demonstrated high accuracy, ranging from 95.2% to 98.6%. Furthermore, the calculated displacement exhibited a high degree of agreement with the measured displacement. The accuracy of the calculated displacements for the laboratory and field experiments ranged from 89.5% to 99.1%. The proposed method can be used for automatic SRW displacement monitoring. [ABSTRACT FROM AUTHOR]

Multifunctionality, integration, and miniaturization are the mainstream trends in modern device development. However, optical display and depth perception, essential for innovative applications like intelligent driving and mixed reality, are typically achieved separately in complex and bulky optical systems. Metasurfaces, ultrathin artificial optical surfaces with strong light–matter interaction capabilities at the nanoscale, offer a versatile platform for creating multifunctional, highly integrated, and miniaturized devices by fully exploring spatial multiplexing and utilizing variables like location, polarization, and frequency. In this study, a multifunctional metasurface that integrates meta‐holography with Dammann gratings is presented. By engineering the phase profile, the device can display specific holographic images in the Fresnel region while simultaneously producing a high‐density spot cloud with over 37 000 spots covering the entire half space. As a proof‐of‐concept, a projection and reconstruction experiment is performed to elaborate its potential for retrieving three‐dimensional (3D) spatial information of real scenarios. The simultaneous generation of display and point cloud holds significant promise for applications in cutting‐edge imaging technologies such as intelligent driving and robot vision. [ABSTRACT FROM AUTHOR]

Facial reconstruction is challenging for plastic surgeons, as it can be difficult to decide the best approach. One technique that has been widely used is the anterolateral thigh (ALT) flap due to its numerous benefits. However, its thickness can be a drawback, especially regarding facial reconstruction. The thinning technique is not a new novel, but how to apply it to the ALT flap to get the best result hasn't been reported yet. Our study involved 117 patients, and we used 73 thinned ALT flaps to determine the best method to increase the flap's safety. After thinning, we significantly reduced the flap's thickness from an average of 22.5 mm to 5.9 mm, making it more suitable for contouring purposes. We apply a thinned ALT flap for coverage, contouring, and recreating the facial 3D structure. The 12/45 flap has the chance to make the multiple-paddle ALT flap, which helps to reconstruct difficult positions even more flexibly. The key to successfully thinning the ALT flap is understanding the perforator's structure and pathway through the fascia. With the thinning technique, we have overcome the limitations of the flap's thickness, making it suitable for use in whole-body reconstruction. The ALT flap can overcome the restriction of its thickness and can be applied even more extensively in whole-body reconstruction. [ABSTRACT FROM AUTHOR]

Introduction: Colon cancer presents significant surgical challenges that necessitate the development of precise strategies. Standardization with complete mesocolic excision (CME) is common, but some cases require extended resections. This study investigates the use of 3D Image Processing and Reconstruction (3D-IPR) to improve diagnostic accuracy in locally advanced colon cancer (LACC) with suspected infiltration and achieve R0 surgery. Methods: Single-center, prospective, observational, comparative, non-randomized study. •Participants: Patients aged > 18 years undergoing LACC surgery, as indicated by CT scans, confirmed via colonoscopy. Exclusion criteria include neoadjuvant therapy, suspected carcinomatosis on CT, and unresectable tumors. •Interventions: 3D-IPR models are used for surgical planning, providing detailed tumor and surrounding structure metrics. Surgical procedures are guided by CT scans and intraoperative findings, categorized by surgical margins as R0, R1, or R2. •Objective: The primary goal is to evaluate 3D-IPR's utility in achieving R0 resection in LACC with suspected infiltration. Secondary objectives include assessing preoperative surgical strategy, comparing CT reports, detecting adenopathy, and identifying vascularization and anatomical variants. • Outcome: The main outcome is the diagnostic accuracy of 3D-IPR in determining tumor infiltration of neighboring structures compared to conventional CT scans, using definitive pathological reports as the gold standard. Results: •Recruitment and Number Analyzed: The study aims to recruit about 20 patients annually over two years, focusing on preoperative 3D-IPR analysis and subsequent surgical procedures. •Outcome Parameters: These include loco-regional and distant recurrence rates, peritoneal carcinomatosis, disease-free and overall survival, and mortality due to oncologic progression. •Harms: No additional risks from CT scans, as they are mandatory for staging colon tumors. 3D-IPR is derived from these CT scans. Discussion: If successful, this study could provide an objective tool for precise tumor extension delimitation, aiding decision-making for radiologists, surgeons, and multidisciplinary teams. Enhanced staging through 3D-IPR may influence therapeutic strategies, reduce postsurgical complications, and improve the quality of life of patients with LACC. Trial registration: Trial is registered at ISRCTN registry as ISRCTN81005215. Protocol version I (Date 29/06/2023). [ABSTRACT FROM AUTHOR]

OBJECTIVES CALGB140503/JCOG0802 RCTs comparing lobectomy with sublobar resection in stage IA NSCLC have confirmed the non-inferiority of segmentectomy. Additional insight is needed to improve preoperative work-up and intraoperative strategies to increase safety and promote the dissemination of minimally invasive segmentectomy (MIS). A Delphi panel study assessed the level of consensus among surgeons for the planning and management of MIS. METHODS Twenty-one expert lung surgeons represented academic institutions, major teaching hospitals, and community hospitals from Europe, North America, and Asia. A 3-round Delphi methodology was used to analyse the answers of each panellist. Recognizing that questions with fewer response options have a higher consensus probability due to limited variability, weighted consensus thresholds were modified based on the number of response options. RESULTS The 21 panellists responded to all 3 rounds of questions. Based on the most robust consensus (94.4%), 3D chest CT reconstructions are recommended only when planning complex segmentectomies. Surgeons should perform 3D reconstructions chest CT scans (consensus = 83.3%). The most effective and safest technique is image-guided VATS in a hybrid operating room (consensus = 83.3%). Dyes with intravenous administration are the safest technique for identifying the intersegmental plane during MIS (consensus = 72.2%). Augmented/mixed reality will probably not immediately help reduce perioperative complications (consensus = 72.2%). CONCLUSIONS This Delphi consensus supports 3D reconstructions and preoperative pulmonary nodule localization before complex MIS. These recommendations should be considered when allocating resources to improve MIS's safety and oncologic efficacy for patients with small, early-stage lung cancers. [ABSTRACT FROM AUTHOR]

The revelation of the internal structure of objects through computed tomography (CT scan) contributes to a more comprehensive understanding of their creation, the assessment of their preservation status, and the prediction of their future behavior. Consequently, in the case of Yiannis Pappas' collection, this knowledge aids in the perpetuation of the models it hosts, which are made from malleable materials, such as wax, plasticine, and mazut, on metallic armature. This publication presents the complete methodology for extracting three-dimensional (3D) models (reconstructions) of the individual construction materials of the figurines, with the aim of subsequently utilizing them in research, as well as in their digital preservation and restoration. The 3D reconstructions were obtained by automatic segmentation algorithms based on the absorption measurements of the materials of the specific figurines, and were furthered edited (post-processing) to obtain the final models. [ABSTRACT FROM AUTHOR]

Expansion Microscopy (ExM) and Light‐Sheet Fluorescence Microscopy (LSFM) are forefront imaging techniques that enable high‐resolution visualization of biological specimens. ExM enhances nanoscale investigation using conventional fluorescence microscopes, while LSFM offers rapid, minimally invasive imaging over large volumes. This review explores the joint advancements of ExM and LSFM, focusing on the excellent performance of the integrated modality obtained from the combination of the two, which is refer to as ExLSFM. In doing so, the chemical processes required for ExM, the tailored optical setup of LSFM for examining expanded samples, and the adjustments in sample preparation for accurate data collection are emphasized. It is delve into various specimen types studied using this integrated method and assess its potential for future applications. The goal of this literature review is to enrich the comprehension of ExM and LSFM, encouraging their wider use and ongoing development, looking forward to the upcoming challenges, and anticipating innovations in these imaging techniques. [ABSTRACT FROM AUTHOR]

DP-MVSNet, a learning-based multi-view stereo (MVS) network utilizing deformable convolutions, addresses the challenge of weak textures in MVS reconstruction. Surpassing the baseline MVSNet architecture, DP-MVSNet introduces deformable convolution modules for enhanced feature extraction and incorporates deformable 3D convolutions into the view weighting network to optimize cross-view informative feature aggregation. DP-MVSNet achieves advanced results on the DTU (0.338 vs. PatchMatchNet's 0.352) and Tanks & Temples (55.07 vs. PatchMatchNet's 53.15) datasets, generating denser point clouds with finer details and higher scene completeness. Ablation studies confirm the effectiveness of the deformable convolution-based multiscale feature extraction (DMFE) and deformable 3D convolution-based learning-based patch matching (DLP) modules, reducing overall error by 4.0 % compared to the baseline while maintaining competitive runtime. Demonstrating remarkable robustness in reconstructing realistic landscapes without ground truth calibration, DP-MVSNet represents a significant advancement in accurate and efficient 3D reconstruction of complex environments, particularly those with weakly textured regions. [ABSTRACT FROM AUTHOR]

The increasing demand for accurate and detailed 3D modeling in fields such as cultural heritage preservation, industrial inspection, and scientific research necessitates advanced techniques to enhance model quality. This paper addresses this necessity by incorporating spectral imaging data to improve the surface detail and reflectivity of 3D models. The methodology integrates spectral imaging with traditional 3D modeling processes, offering a novel approach to capturing fine textures and subtle surface variations. The experimental results of this paper underscore the advantages of incorporating spectral imaging data in the creation of 3D models, particularly in terms of enhancing surface detail and reflectivity. The achieved experimental results demonstrate that 3D models generated with spectral imaging data exhibit significant improvements in surface detail and accuracy, particularly for objects with intricate surface patterns. These findings highlight the potential of spectral imaging in enhancing 3D model quality. This approach offers significant advancements in 3D modeling, contributing to more precise and reliable representations of complex surfaces. [ABSTRACT FROM AUTHOR]

This paper presents a novel framework for 3D face reconstruction from single 2D images and addresses critical limitations in existing methods. Our approach integrates modified adversarial neural networks with graph neural networks to achieve state-of-the-art performance. Key innovations include (1) a generator architecture based on Graph Convolutional Networks (GCNs) with a novel loss function and identity blocks, mitigating mode collapse and instability; (2) the integration of facial landmarks and a non-parametric efficient-net decoder for enhanced feature capture; and (3) a lightweight GCN-based discriminator for improved accuracy and stability. Evaluated on the 300W-LP and AFLW2000-3D datasets, our method outperforms existing approaches, reducing Chamfer Distance by 62.7% and Earth Mover's Distance by 57.1% on 300W-LP. Moreover, our framework demonstrates superior robustness to variations in head positioning, occlusion, noise, and lighting conditions while achieving significantly faster processing times. [ABSTRACT FROM AUTHOR]

In the production of high-definition maps, it is necessary to achieve the three-dimensional instantiation of road furniture that is difficult to depict on traditional maps. The development of mobile laser measurement technology provides a new means for acquiring road furniture data. To address the issue of traffic marking extraction accuracy in practical production, which is affected by degradation, occlusion, and non-standard variations, this paper proposes a 3D reconstruction method based on energy functions and template matching, using zebra crossings in vehicle-mounted LiDAR point clouds as an example. First, regions of interest (RoIs) containing zebra crossings are obtained through manual selection. Candidate point sets are then obtained at fixed distances, and their neighborhood intensity features are calculated to determine the number of zebra stripes using non-maximum suppression. Next, the slice intensity feature of each zebra stripe is calculated, followed by outlier filtering to determine the optimized length. Finally, a matching template is selected, and an energy function composed of the average intensity of the point cloud within the template, the intensity information entropy, and the intensity gradient at the template boundary is constructed. The 3D reconstruction result is obtained by solving the energy function, performing mode statistics, and normalization. This method enables the complete 3D reconstruction of zebra stripes within the RoI, maintaining an average planar corner accuracy within 0.05 m and an elevation accuracy within 0.02 m. The matching and reconstruction time does not exceed 1 s, and it has been applied in practical production. [ABSTRACT FROM AUTHOR]

To decipher the mechanical response mechanisms of coal seams with multi-scale porefracture structure to supercritical CO2 (ScCO2) injection, two coal samples from different mines of the Ordos Basin, North China, were first selected for conducting ScCO2-watercoal reaction experiments. Subsequently, the pore-fracture structure, macroscopic and microscopic mechanical behaviors of samples with different reaction times were analyzed, and the evolution patterns of pore-fracture parameters and relationships between the macroscopic and microscopic mechanical parameters were finally elucidated. The results showed that the ScCO2-water-coal reaction modifies the pore-fracture structure in coal. Originally filled fractures re-open, original micro-fractures expand, new fractures form, and pores evolve from small- to large-sized. After the ScCO2-water-coal reaction, the evolution of the compaction stage, the macroscopic mechanical parameters and the energy dissipation during loading corroborate the weakening effect of the ScCO2-water-coal reaction on coal. The changes observed in the microscopic mechanical parameters align with those in the macroscopic mechanical parameters; however, due to the strong heterogeneity of coal and the inability of microscopic parameters to reflect the component and porefracture distribution, certain characteristics of the change amplitude of macroscopic and microscopic mechanical parameters of coal are inconsistent. The ScCO2 extraction effect, the chemical dissolution, the different-sized pore-fracture evolution, the coupling effect of geostress, reservoir pressure, and swelling stress are the main factors to consider during the process of ScCO2 sequestration in deep coal seams at the micro-, meso- and macro-scales, as they are responsible for potential safety issues. [ABSTRACT FROM AUTHOR]

Purpose: The aim of this study was to find an alternative method to meet traditional human anatomy teaching and clinical needs in order to solve the problem of cranial specimen attrition and specimen resource shortage due to long-term use. Methods: We performed a computed tomography (CT) scan of a well-preserved male cranial specimen and used Mimics 19.0 software for 3D reconstruction and cranial block separation. Subsequently, we compared the recognition ability of the processed cranial digital model with that of the 3D body digital model and used 3D printing to create the cranial model and compare it with the physical specimen. Results: Twenty-two cranial bone block models were obtained, excluding the hyoid bone. Their 3D reconstructed digital models had better bony landmark recognition than the 3D body human digital models, and the differences between the 3D printed models and the physical specimens were minimal. In addition, only one stereolithography (STL) file was required to produce the cranial models, which facilitates repetitive printing at any time. Conclusion: By isolating cranial bone blocks through 3D reconstruction techniques and preparing high-quality cranial models in combination with 3D printing techniques, this study solves the problem of shortage of cranial teaching specimens for the sustainable development of clinical and medical schools. [ABSTRACT FROM AUTHOR]

In recent years, learning-based MVS methods have achieved excellent performance compared with traditional methods. However, these methods still have notable shortcomings, such as the low efficiency of traditional convolutional networks and simple feature fusion, which lead to incomplete reconstruction. In this research, we propose a lightweight network for low-texture scene reconstruction (LLR-MVSNet). To improve accuracy and efficiency, a lightweight network is proposed, including a multi-scale feature extraction module and a weighted feature fusion module. The multi-scale feature extraction module uses depth-separable convolution and point-wise convolution to replace traditional convolution, which can reduce network parameters and improve the model efficiency. In order to improve the fusion accuracy, a weighted feature fusion module is proposed, which can selectively emphasize features, suppress useless information and improve the fusion accuracy. With rapid computational speed and high performance, our method surpasses the state-of-the-art benchmarks and performs well on the DTU and the Tanks & Temples datasets. The code of our method will be made available at . [ABSTRACT FROM AUTHOR]

Robots in tomato greenhouses need to perceive the plant and plant parts accurately to automate monitoring, harvesting, and de-leafing tasks. Existing perception systems struggle with the high levels of occlusion in plants and often result in poor perception accuracy. One reason for this is because they use fixed cameras or predefined camera movements. Next-best-view (NBV) planning presents an alternate approach, in which the camera viewpoints are reasoned and strategically planned such that the perception accuracy is improved. However, existing NBV-planning algorithms are agnostic to the task-at-hand and give equal importance to all the plant parts. This strategy is inefficient for greenhouse tasks that require targeted perception of specific plant parts, such as the perception of leaf nodes for de-leafing. To improve targeted perception in complex greenhouse environments, NBV planning algorithms need an attention mechanism to focus on the task-relevant plant parts. In this paper, the role of attention in improving targeted perception using an attention-driven NBV planning strategy was investigated. Through simulation experiments using plants with high levels of occlusion and structural complexity, it was shown that focusing attention on task-relevant plant parts can significantly improve the speed and accuracy of 3D reconstruction. Further, with real-world experiments, it was shown that these benefits extend to complex greenhouse conditions with natural variation and occlusion, natural illumination, sensor noise, and uncertainty in camera poses. The results clearly indicate that using attention-driven NBV planning in greenhouses can significantly improve the efficiency of perception and enhance the performance of robotic systems in greenhouse crop production. • Method to plan the next-best viewpoint to reconstruct the nodes of tomato plants. • Adding attention mechanism improves the efficiency of next-best-view planning. • Performed systematic evaluations with both simulation and real-world tomato plants. • Effectively handles occlusion in complex greenhouse environments. [ABSTRACT FROM AUTHOR]

This work presents AstrobeeCD, a system for 3D scene change detection toward near-real-time environmental awareness of space outposts using the Astrobee free-flying robot in microgravity. Assistive free-flyer robots autonomously caring for future crewed space habitats must be able to detect day-to-day interior changes to track inventory, detect and diagnose faults, and monitor the outpost status. A set of image and depth data from one time step is used to reconstruct a 3D model of the environment. The 3D model is used as the basis for comparison for free-flyer environment surveys at future time steps, where an image-based change detection algorithm identifies inconsistencies against the 3D model. Change detection is demonstrated using real image and pose data collected by an Astrobee robot in a test environment on Earth at NASA Ames Research Center and from microgravity aboard the International Space Station. Change detection computation time and performance are quantitatively evaluated on the test data captured on Earth, and it identifies scene changes more quickly than a point cloud clustering-based algorithm applied to data from the same surveys. [Display omitted] • Free-flying robotic assistants in space need to efficiently detect scene changes. • AstrobeeCD demonstrates near-real-time 3D scene change detection for space habitats. • AstrobeeCD processes Astrobee's images and poses for image-based change detection. • AstrobeeCD is validated on Astrobee data collected on Earth and in space on the ISS. • Data and code are publicly released to promote further development. [ABSTRACT FROM AUTHOR]

The body size, shape, weight, and scoring of goats are crucial indicators for assessing their growth, health, and meat production. The application of computer vision technology to measure these parameters is becoming increasingly prevalent. However, in real farm environments, obstacles, such as fences, ground conditions, and dust, pose significant challenges for obtaining accurate goat point cloud data. These obstacles lead to difficulties in rapid data extraction and result in incomplete reconstructions, causing substantial measurement errors. To address these challenges, we developed a system for real-time, non-contact acquisition, extraction, and reconstruction of goat point clouds using three depth cameras. The system operates in a scenario where goats walk naturally through a designated channel, and bidirectional distributed triggering logic is employed to ensure real-time acquisition of the point cloud. We also designed a noise recognition and filtering method tailored to handle complex environmental interferences found on farms, enabling automatic extraction of the goat point cloud. Furthermore, a distributed point cloud completion algorithm was developed to reconstruct missing sections of the goat point cloud caused by unavoidable factors such as railings and dust. Measurements of body height, body slant length, and chest circumference were calculated separately with deviation of no more than 25 mm and an average error of 3.1%. The system processes each goat in an average time of 3–5 s. This method provides rapid and accurate extraction and complementary reconstruction of 3D point clouds of goats in motion on real farms, without human intervention. It offers a valuable technological solution for non-contact monitoring and evaluation of goat body size, weight, shape, and appearance. [ABSTRACT FROM AUTHOR]

This review discusses recent advances in robotic surgery for urologic tumors, focusing on three key areas: robotic systems, assistive technologies, and artificial intelligence. The Da Vinci SP system has enhanced the minimally invasive nature of robotic surgeries, while the Senhance system offers advantages such as tactile feedback and eye-tracking capabilities. Technologies like 3D reconstruction combined with augmented reality and fluorescence imaging aid surgeons in precisely identifying the anatomical relationships between tumors and surrounding structures, improving surgical efficiency and outcomes. Additionally, the development of artificial intelligence lays the groundwork for automated robotics. As these technologies continue to evolve, we are entering an era of minimally invasive, precise, and intelligent robotic surgery. [ABSTRACT FROM AUTHOR]

Large-scale indoor 3D reconstruction with multiple robots faces challenges in core enabling technologies. This work contributes to a framework addressing localization, coordination, and vision processing for multi-agent reconstruction. A system architecture fusing visible light positioning, multi-agent path finding via reinforcement learning, and 360° camera techniques for 3D reconstruction is proposed. Our visible light positioning algorithm leverages existing lighting for centimeter-level localization without additional infrastructure. Meanwhile, a decentralized reinforcement learning approach is developed to solve the multi-agent path finding problem, with communications among agents optimized. Our 3D reconstruction pipeline utilizes equirectangular projection from 360° cameras to facilitate depth-independent reconstruction from posed monocular images using neural networks. Experimental validation demonstrates centimeter-level indoor navigation and 3D scene reconstruction capabilities of our framework. The challenges and limitations stemming from the above enabling technologies are discussed at the end of each corresponding section. In summary, this research advances fundamental techniques for multi-robot indoor 3D modeling, contributing to automated, data-driven applications through coordinated robot navigation, perception, and modeling. [ABSTRACT FROM AUTHOR]

The present study addresses the segmentation and the 3D reconstruction of the corpus callosum from MRI scans. Accurate segmentation of the corpus callosum is essential in order to enable its reconstruction and 3D visualization to facilitate early diagnosis. In fact, many studies have established a strong correlation between the shape of the corpus callosum and several pathological conditions. However, the segmentation is made difficult by regions of similar intensity within the MRI images. To overcome this challenge, we propose an automated method that relies mainly on a probabilistic neural network applied to superpixels. The proposed scheme involves segmenting the corpus callosum within the MRI scans, followed by the application of the marching cubes technique in order to generate 3D volumes. The effectiveness of the proposed method has been extensively validated on four challenging datasets (OASIS, ABIDE, MIRIAD, and SBD), and the obtained results demonstrate its superior performance compared to other state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Due to its rapid projection and capture of fringe patterns, coded structured light measurement technology is widely utilised for acquiring three-dimensional information in welding areas. To achieve real-time segmentation of the welding area for T-joint butt welding workpieces with various noise interference, this paper has developed a lightweight dual-resolution semantic segmentation network (LDRNet). This paper designed a lightweight feature extraction module that provides efficient feature representations for the network using fewer parameters and computational costs. To enhance the network’s robustness to complex environmental noise, this paper introduced a multi-scale adaptive feature extraction module that can capture information at different scales of the environment. To further improve segmentation accuracy, this paper reconstructed the traditional pyramid pooling module and combined it with the CBAM attention mechanism to enhance the focus on important features. This paper proposes a local feature constraint method to improve the phase matching accuracy. Experimental results show that this paper algorithm significantly reduces Params and FLOPs by 62.02% and 65.78%, respectively, compared to the original network DDRNet-23-Slim. Moreover, it leads to an improvement in the IOU of the welding area from 74.55% to 77.43%. Additionally, this paper’s proposed algorithm effectively reduces the generation of noise points through phase matching by approximately 93.88%. Consequently, this algorithm satisfactorily meets the requirements of practical production processes. [ABSTRACT FROM AUTHOR]

Early detection and timely intervention are essential to prevent liver fibrosis from progressing to cirrhosis or hepatocellular carcinoma. Herein, utilizing the enhanced ferritin‐manganese interaction by nanoplatinum growth, a novel ferritin‐platinum‐manganese magnetic resonance nanoplatform with RGD grafting and metformin loading (FNMMR) is developed. RGD can enhance the targeting ability of the nanoplatform toward integrin αVβ3 on activated hepatic stellate cells (aHSCs) in liver fibrosis. Systemic delivery of FNMMR shows clear degree‐dependent magnetic resonance contrast enhancement in liver fibrosis. 3D reconstruction techniques and histogram‐based features are achieved to qualitatively and quantitatively analyze the inhomogeneous liver fibrosis areas. FNMMR with catalase‐like activity can catalyze the generation of O2 to alleviate the liver fibrosis hypoxia and inhibit the expression of HIF‐1α, blocking the TGF‐β1/Smad signaling pathway. In addition, metformin shows synergy with HIF‐1α reduction in blocking the TGF‐β1/Smad pathway, effectively inhibiting the activation of HSCs and reducing collagen formation. Furthermore, FNMMR can achieve real‐time anti‐fibrotic therapy monitoring by magnetic resonance imaging. Importantly, no obvious side effects can be observed in both histological and hematology examinations. Therefore, this work presents a novel nanoplatform for accurate liver fibrosis diagnosis and synergistic anti‐fibrotic therapy with real‐time monitoring. [ABSTRACT FROM AUTHOR]

Background: This study aimed to precisely predict the size and silicone oil injection of a foldable capsular vitreous body (FCVB) via computerized three-dimensional (3D) ocular reconstruction in the treatment of severe retinal detachment in China. Methods: The 3D software Unigraphics NX was applied to determine the volume of the inner cavity with 16–30 mm axial length, assigning the anterior and posterior chambers, the FCVB sizes, and the silicone oil injection volume, and modeling the data between the axial length and the FCVB size. In clinical practice, IOL Master was applied to accurately measure the axial length of the contralateral healthy eye to anchor the anterior-posterior and horizontal diameters of the operated eye in horizontal position CT, and compared with the model to recommend the FCVB size and silicone oil amount, and the clinical effect was validated in cases across five hospitals in China. Results: For the axial length of 16–30 mm, the volume of the inner cavity is 1.2 ml-8.4 ml. FCVB size and silicone oil volume were recommended based on this volume of the inner cavity. Of 253 cases, we noted 11 cases implanted with AV-10P and 1.05 ± 0.21 ml of silicone oil, 41 with AV-12P and 1.58 ± 0.18 ml of silicone oil, 163 with AV-13.5P and 2.48 ± 0.29 ml of silicone oil, 31 with AV-15P and 3.57 ± 0.39 ml of silicone oil, and 7 with AV-17P and 5.71 ± 0.81 ml of silicone oil. There was no significant difference in postoperative visual acuity scores compared with preoperative (P = 0.097), postoperative IOP(10.29 ± 0.57mmHg)was slightly higher than preoperative IOP (9.76 ± 0.48 mmHg), but there was still no statistically significant difference between the two comparisons (P = 0.405). Conclusion: Three-dimensional reconstruction prediction is a good solution for eyeballs with obvious individualized changes in severe retinal detachment, and this method helps doctors standardize FCVB size selection and the silicone oil amount for patients. [ABSTRACT FROM AUTHOR]

Objective: To create a deep‐learning automatic segmentation model for esophageal cancer (EC), metastatic lymph nodes (MLNs) and their adjacent structures using the UperNet Swin network and computed tomography angiography (CTA) images and to improve the effectiveness and precision of EC automatic segmentation and TN stage diagnosis. Methods: Attention U‐Net, UperNet Swin, UNet++ and UNet were used to train the EC segmentation model to automatically segment the EC, esophagus, pericardium, aorta and MLN from CTA images of 182 patients with postoperative pathologically proven EC. The Dice similarity coefficient (DSC), sensitivity, and positive predictive value (PPV) were used to assess their segmentation effectiveness. The volume of EC was calculated using the segmentation results, and the outcomes and times of automatic and human segmentation were compared. All statistical analyses were completed using SPSS 25.0 software. Results: Among the four EC autosegmentation models, the UperNet Swin had the best autosegmentation results with a DSC of 0.7820 and the highest values of EC sensitivity and PPV. The esophagus, pericardium, aorta and MLN had DSCs of 0.7298, 0.9664, 0.9496 and 0.5091. The DSCs of the UperNet Swin were 0.6164, 0.7842, 0.8190, and 0.7259 for T1‐4 EC. The volume of EC and its adjacent structures between the ground truth and UperNet Swin model were not significantly different. Conclusions: The UperNet Swin showed excellent efficiency in autosegmentation and volume measurement of EC, MLN and its adjacent structures in different T stage, which can help to T and N stage diagnose EC and will save clinicians time and energy. [ABSTRACT FROM AUTHOR]

In dietary assessment using a single-view food image, an object of known size, such as a checkerboard, is often placed manually in the camera's view as a scale reference to estimate food volume. This traditional scale reference is inconvenient to use because of the manual placement requirement. Consequently, utensils, such as plates and bowls, have been suggested as alternative references. Although these references do not need a manual placement procedure, there is a unique challenge when a dining bowl is used as a reference. Unlike a dining plate, whose shallow shape does not usually block the view of the food, a dining bowl does obscure the food view, and its shape may not be fully observable from the single-view food image. As a result, significant errors may occur in food volume estimation due to the unknown shape of the bowl. To address this challenge, we present a novel method to premeasure both the size and shape of the empty bowl before it is used in a dietary assessment study. In our method, an image is taken with a labeled paper ruler adhered to the interior surface of the bowl, a mathematical model is developed to describe its shape and size, and then an optimization method is used to determine the bowl parameters based on the locations of observed ruler makers from the bowl image. Experimental studies were performed using both simulated and actual bowls to assess the reliability and accuracy of our bowl measurement method. [ABSTRACT FROM AUTHOR]

Recent single-view reconstruction methods have sought to reconstruct 3D point clouds from images and corresponding silhouette collections alone. However, merely utilizing input images as supervision without any auxiliary methods amplifies the matching ambiguity. To address this issue, we propose a self-supervised 3D point cloud reconstruction method based on generative adversarial network (GAN) inversion. Three novel components are introduced to solve the intrinsic challenges of cross-dimensional inversion. First, we develop a uniform loss to enhance the uniformity of the point clouds generated by the GAN. Second, we devise a coarse-to-fine differentiable point cloud renderer to facilitate accurate projections. Third, we design a pseudo ground-truth pose predictor that can estimate the precise viewpoints of the input images. Experimental results on both synthetic datasets and real-world datasets demonstrate that our approach outperforms existing state-of-the-art 2D supervised reconstruction methods and is comparable to 3D supervised approaches. [ABSTRACT FROM AUTHOR]

To address the problem of incomplete perception of limited viewpoints of ore blocks in future remote and intelligent shoveling-dominated mining scenarios, a method of using new view generation technology to predict ore blocks with limited view based on a latent diffusion model is proposed. Initially, an ore block image-pose dataset is created. Then, based on prior knowledge, the latent diffusion model undergoes transfer learning to develop an intelligent ore block shape prediction model (IOBSPM) for rock blocks. During training, structural similarity loss is innovatively introduced to constrain the prediction results and solve the issue of discontinuity in generated images. Finally, neural surface reconstruction is performed using the generated multi-view images of rock blocks to obtain a 3D model. Experimental results show that the prediction model, trained on the rock block dataset, produces better morphological and detail generation compared to the original model, with single-view generation time within 5 s. The average PSNR, SSIM, and LPIPS values reach 23.02 dB, 0.754, and 0.268, respectively. The generated views also demonstrate good performance in 3D reconstruction, highlighting significant implications for future research on remote and autonomous shoveling. [ABSTRACT FROM AUTHOR]

Flame temperature field measurement has always been a key topic in combustion research, which is of great significance for combustion state diagnosis and fuel combustion optimization. Deep learning technology exploits its superior nonlinear ability to rapidly reconstruct the three-dimensional (3D) temperature field of flames from flame light-field images. However, existing algorithms have problems with complex networks, poor noise resistance and low reconstruction accuracy. This paper proposes a lightweight 3D flame temperature field reconstruction algorithm based on an improved MobileNet that combines residuals and spatial attention mechanisms. This method basically achieves a balance between low complexity and high accuracy and has good noise resistance performance. Simulation results show that the average relative error of temperature field reconstruction on the noise-free unimodal flame dataset is only 0.022%, and its computational complexity is only one-tenth of the existing CNN. The noise simulation experiment shows that this method has good noise resistance performance. The maximum relative error and the average relative error at Gaussian white noise standard deviation σ = 0.15 are only 2.91% and 0.27%, respectively. [ABSTRACT FROM AUTHOR]

Numerical simulation studies of water–rock interaction mechanisms and pore-scale multiphase flow properties often require high computational efficiency and realistic geometries to enable a fast and accurate description of hydrodynamic behavior. In this paper, we have chosen to use deep learning models to achieve these requirements, firstly by using encoder structures to refine the image segmentation of void-solid structures on complex geometries of scanning electron microscopy (SEM) images of porous media through few-shot learning (FSL), not only obtaining an accuracy of 0.97, but also reducing the amount of annotation work. We then focus on pore-scale three-dimensional (3D) structural reconstruction using the unpaired image-to-image translation method, optimizing the cycle-consistent adversarial network (cycle-GAN) model via sliced Wasserstein distance (SWD) to transfer marine sedimentary sandstone features to the initial image, and the geometric stochastic reconstruction problems are transformed into optimization problems. Subsequently, the computational efficiency was improved by a factor of 21 by implementing the lattice Boltzmann simulation method (LBM) accelerated by GPU through compute-unified device architecture (CUDA). The flow field distribution and absolute permeability of the extracted 2D samples and the reconstructed 3D porous media structure were simulated. The results showed that our method could rapidly and accurately reconstruct the 3D structures of a given feature, ensuring statistical equivalence between the 3D reconstructed structures and 2D samples. We solve the problem of extrapolation-based 3D reconstruction of porous media and significantly reduce the time spent on structure extraction and numerical calculations. [ABSTRACT FROM AUTHOR]

In the field of spinal pathology, sagittal balance of the spine is usually judged by the spatial structure and morphology of pelvis, which can be represented by pelvic parameters. Pelvic parameters, including pelvic incidence, pelvic tilt and sacral slope, are therefore essential for the diagnosis and treatment of spinal disorders, however, it is a time-consuming and laborious procedure to measure these parameters by traditional methods. In this paper, an automatic measurement framework for pelvic CT images was proposed to calculate three-dimensional (3D) pelvic parameters with the support of deep learning technology. Pelvic images were first preprocessed, and 3D reconstruction was then performed to obtain 3D pelvic model by the Visualization Toolkit. DRINet was trained to segment the femoral head region in the pelvic images, and 3D sphere fitting was performed to locate the femoral heads. In addition, VGG16 was adopted to recognize images containing superior sacral endplate, and the plane growth algorithm was used to fit the plane so that the midpoint and normal vector of the superior sacral endplate could be obtained. Finally, 3D pelvic parameters were automatically calculated, and compared with manual measurements for 15 patients. The proposed framework automatically generated 3D pelvic models, and calculated two-dimensional (2D) and 3D pelvic parameters from continuous CT images. Experiments demonstrated that the framework can greatly speed up the calculation of pelvic parameters, and these parameters are accurate when compared with the manual measurements. In conclusion, the proposed framework demonstrates good performance on automatic pelvimetry measurement by incorporating deep learning technology, and can well replace the traditional methods for pelvic parameter measurement. [ABSTRACT FROM AUTHOR]

Objectives: Efficient match pair selection and image feature matching directly affect the efficiency of structure from motion (SfM) -based 3D reconstruction for unmanned aerial vehicle (UAV) images. This paper combines the inverted and direct index structure of the vocabulary tree to achieve the speedup of match pair selection and feature matching for UAV images. Methods: First, for match pair selection, vocabulary tree-based image retrieval has been the commonly used technique. However, it depends on the fixed number or fixed ratio threshold for match pair selection, which may cause many redundant match pairs. An adaptive vocabulary tree-based retrieval algorithm is designed for match pair selection by using the word-image index structure and the spatial distribution of similarity scores, and it can avoid the drawback of depending on fixed thresholds. Second, for feature matching, the nearest neighbor searching method attempts to compute the Euclidean distance exhaustively between two sets of feature descriptors, which causes high computational costs and generates high outlier ratios. Thus, a guided feature matching (GFM) algorithm is presented which casts the explicit closest descriptor searching as the direct assignment by using the image-word index structure of the vocabulary tree. Combining the match pair selection and GFM algorithm, an integrated workflow is finally presented to achieve feature matching of both ordered and unordered UAV images with high precision and efficiency. Results: The proposed workflow is verified using four UAV datasets and compared comprehensively with classical nearest neighbor searching algorithms and commercial software packages. Conclusions: The experimental results verify that the proposed method can achieve efficient match pair selection and avoid the problem of retrieving too many or too few match pairs that are usually caused by traditional methods using fixed threshold or number strategies. Without the sacrifice of matching precision, the speedup ratio of direct assign based feature matching ranges from 156 to 228, and competitive accuracy is also obtained from 3D reconstruction compared with the nearest neighbor searching method. [ABSTRACT FROM AUTHOR]

Rapid localization of ROI (Region of Interest) for tomographic medical images (TMIs) is an important foundation for efficient image reading, computer-aided education, and well-informed rights of patients. However, due to the multimodality of clinical TMIs, the complexity of anatomy, and the deformation of organs caused by diseases, it is difficult to have a universal and low-cost method for ROI organ localization. This article focuses on actual concerns of TMIs from medical students, engineers, interdisciplinary researchers, and patients, exploring a universal registration method between the clinical CT/MRI dataset and CVH (Chinese Visible Human) to locate the organ ROI in a low-cost and lightweight way. The proposed method is called Two-step Progressive Registration (TSPR), where the first registration adopts "eye–nose triangle" features to determine the spatial orientation, and the second registration adopts the circular contour to determine the spatial scale, ultimately achieving CVH anatomical knowledge automated mapping. Through experimentation with representative clinical TMIs, the registration results are capable of labeling the ROI in the images well and can adapt to the deformation problem of ROI, as well as local extremum problems that are prone to occur in inter-subject registration. Unlike the ideal requirements for TMIs' data quality in laboratory research, TSPR has good adaptability to incomplete and non-thin-layer quality in real clinical data in a low-cost and lightweight way. This helps medical students, engineers, and interdisciplinary researchers independently browse images, receive computer-aided education, and provide patients with better access to well-informed services, highlighting the potential of digital public health and medical education. [ABSTRACT FROM AUTHOR]

High-quality observations of the real world are crucial for a variety of applications, including producing 3D printed replicas of small-scale scenes and conducting inspections of large-scale infrastructure. These 3D observations are commonly obtained by combining multiple sensor measurements from different views. Guiding the selection of suitable views is known as the Next Best View (NBV) planning problem. Most NBV approaches reason about measurements using rigid data structures (e.g., surface meshes or voxel grids). This simplifies next best view selection but can be computationally expensive, reduces real-world fidelity and couples the selection of a next best view with the final data processing. This paper presents the Surface Edge Explorer (SEE), a NBV approach that selects new observations directly from previous sensor measurements without requiring rigid data structures. SEE uses measurement density to propose next best views that increase coverage of insufficiently observed surfaces while avoiding potential occlusions. Statistical results from simulated experiments show that SEE can attain similar or better surface coverage with less observation time and travel distance than evaluated volumetric approaches on both small- and large-scale scenes. Real-world experiments demonstrate SEE autonomously observing a deer statue using a 3D sensor affixed to a robotic arm. [ABSTRACT FROM AUTHOR]