Search Results (102)

To address the problems of traditional methods that rely on destructive sampling, the poor adaptability of fixed equipment, and the susceptibility of single-view angle measurements to occlusions, a non-destructive and portable device for three-dimensional phenotyping and biomass detection in lettuce was developed. Based on the Structure-from-Motion Multi-View Stereo (SFM-MVS) algorithms, a high-precision three-dimensional point cloud model was reconstructed from multi-view RGB image sequences, and 12 phenotypic parameters, such as plant height, crown width, were accurately extracted. Through regression analyses of plant height, crown width, and crown height, and the R² values were 0.98, 0.99, and 0.99, respectively, the RMSE values were 2.26 mm, 1.74 mm, and 1.69 mm, respectively. On this basis, four biomass prediction models were developed using Adaptive Boosting (AdaBoost), Support Vector Regression (SVR), Gradient Boosting Decision Tree (GBDT), and Random Forest Regression (RFR). The results indicated that the RFR model based on the projected convex hull area, point cloud convex hull surface area, and projected convex hull perimeter performed the best, with an R² of 0.90, an RMSE of 2.63 g, and an RMSEn of 9.53%, indicating that the RFR was able to accurately simulate lettuce biomass. This research achieves three-dimensional reconstruction and accurate biomass prediction of facility lettuce, and provides a portable and lightweight solution for facility crop growth detection. Full article

Three-dimensional (3D) reconstruction has become a fundamental technology in applications ranging from cultural heritage preservation and robotics to forensics and virtual reality. As these applications grow in complexity and realism, the quality of the reconstructed models becomes increasingly critical. Among the many factors that influence reconstruction accuracy, the lighting conditions at capture time remain one of the most influential, yet widely neglected, variables. This review provides a comprehensive survey of classical and modern 3D reconstruction techniques, including Structure from Motion (SfM), Multi-View Stereo (MVS), Photometric Stereo, and recent neural rendering approaches such as Neural Radiance Fields (NeRFs) and 3D Gaussian Splatting (3DGS), while critically evaluating their performance under varying illumination conditions. We describe how lighting-induced artifacts such as shadows, reflections, and exposure imbalances compromise the reconstruction quality and how different approaches attempt to mitigate these effects. Furthermore, we uncover fundamental gaps in current research, including the lack of standardized lighting-aware benchmarks and the limited robustness of state-of-the-art algorithms in uncontrolled environments. By synthesizing knowledge across fields, this review aims to gain a deeper understanding of the interplay between lighting and reconstruction and provides research directions for the future that emphasize the need for adaptive, lighting-robust solutions in 3D vision systems. Full article

Accurately reconstructing tree skeletons from multi-view images is challenging. While most existing works use skeletonization from 3D point clouds, thin branches with low-texture contrast often involve multi-view stereo (MVS) to produce noisy and fragmented point clouds, which break branch connectivity. Leveraging the recent development in accurate mask extraction from images, we introduce a mask-guided graph optimization framework that estimates a 3D skeleton directly from multi-view segmentation masks, bypassing the reliance on point cloud quality. In our method, a skeleton is modeled as a graph whose nodes store positions and radii while its adjacency matrix encodes branch connectivity. We use 3D Gaussian splatting (3DGS) to render silhouettes of the graph and directly optimize the nodes and the adjacency matrix to fit given multi-view silhouettes in a differentiable manner. Furthermore, we use a minimum spanning tree (MST) algorithm during the optimization loop to regularize the graph to a tree structure. Experiments on synthetic and real-world plants show consistent improvements in completeness and structural accuracy over existing point-cloud-based and heuristic baseline methods. Full article

Point clouds continue to be the main data source in 3D modeling studies with unmanned aerial vehicle (UAV) images. Structure-from-Motion (SfM) and MultiView Stereo (MVS) have high time costs for point cloud generation, especially in large data sets. For this reason, state-of-the-art methods such as Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS) have emerged as powerful alternatives for point cloud generation. This paper explores the performance of NeRF and 3DGS methods in generating point clouds from UAV images. For this purpose, the Nerfacto, Instant-NGP, and Splatfacto methods developed in the Nerfstudio framework were used. The obtained point clouds were evaluated by taking the point cloud produced with the photogrammetric method as reference. In this study, the effects of image size and iteration number on the performance of the algorithms were investigated in two different study areas. According to the results, Splatfacto demonstrates promising capabilities in addressing challenges related to scene complexity, rendering efficiency, and accuracy in UAV imagery. Full article

To address the challenge of precise picking point localization in morphologically diverse safflower plants, this study proposes PointSafNet—a novel three-stage 3D point cloud analysis framework with distinct architectural and methodological innovations. In Stage I, we introduce a multi-view reconstruction pipeline integrating Structure from Motion (SfM) and Multi-View Stereo (MVS) to generate high-fidelity 3D plant point clouds. Stage II develops a dual-branch architecture employing Star modules for multi-scale hierarchical geometric feature extraction at the organ level (filaments and frui balls), complemented by a Context-Anchored Attention (CAA) mechanism to capture long-range contextual information. This synergistic feature learning approach addresses morphological variations, achieving 86.83% segmentation accuracy (surpassing PointNet++ by 7.37%) and outperforming conventional point cloud models. Stage III proposes an optimized geometric analysis pipeline combining dual-centroid spatial vectorization with Oriented Bounding Box (OBB)-based proximity analysis, resolving picking coordinate localization across diverse plants with 90% positioning accuracy and 68.82% mean IoU (13.71% improvement). The experiments demonstrate that PointSafNet systematically integrates 3D reconstruction, hierarchical feature learning, and geometric reasoning to provide visual guidance for robotic harvesting systems in complex plant canopies. The framework’s dual emphasis on architectural innovation and geometric modeling offers a generalizable solution for precision agriculture tasks involving morphologically diverse safflowers. Full article

Accurate dietary assessment is essential for effective health management and disease prevention. However, conventional methods that rely on manual food logging and nutritional lookup are often time consuming and error prone. This study proposes an image-based nutritional advisory system that integrates multimodal deep learning to automate food classification, volume estimation, and dietary recommendation to address these limitations. The system employs a fine-tuned CLIP model for zero-shot food recognition, achieving high accuracy across diverse food categories, including unseen items. For volume measurement, a learning-based multi-view stereo (MVS) approach eliminates the need for specialized hardware, yielding reliable estimations with a mean absolute percentage error (MAPE) of 23.5% across standard food categories. Nutritional values are then calculated by referencing verified food composition databases. Furthermore, the system leverages a large language model (Llama 3) to generate personalized dietary advice tailored to individual health goals. The experimental results show that the system attains a top 1 classification accuracy of 91% on CNFOOD-241 and 80% on Food 101 and delivers high-quality recommendation texts with a BLEU-4 score of 45.13. These findings demonstrate the system’s potential as a practical and scalable tool for automated dietary management, offering improved precision, convenience, and user experience. Full article

Non-collaborative (i.e., reflective, transparent, metallic, etc.) surfaces are common in industrial production processes, where 3D reconstruction methods are applied for quantitative quality control inspections. Although the use or combination of photogrammetry and photometric stereo performs well for well-textured or partially textured objects, it usually produces unsatisfactory 3D reconstruction results on non-collaborative surfaces. To improve 3D inspection performances, this paper investigates emerging learning-based surface reconstruction methods, such as Neural Radiance Fields (NeRF), Multi-View Stereo (MVS), Monocular Depth Estimation (MDE), Gaussian Splatting (GS) and image-to-3D generative AI as potential alternatives for industrial inspections. A comprehensive evaluation dataset with several common industrial objects was used to assess methods and gain deeper insights into the applicability of the examined approaches for inspections in industrial scenarios. In the experimental evaluation, geometric comparisons were carried out between the reference data and learning-based reconstructions. The results indicate that no method can outperform all the others across all evaluations. Full article

Feature matching is pivotal when using multi-view stereo (MVS) to reconstruct dense 3D models from calibrated images. This paper proposes PAC-MVSNet, which integrates perspective-aware convolution (PAC) and metadata-enhanced cost volumes to address the challenges in reflective and texture-less regions. PAC dynamically aligns convolutional kernels with scene perspective lines, while the use of metadata (e.g., camera pose distance) enables geometric reasoning during cost aggregation. In PAC-MVSNet, we introduce feature matching with long-range tracking that utilizes both internal and external focuses to integrate extensive contextual data within individual images as well as across multiple images. To enhance the performance of the feature matching with long-range tracking, we also propose a perspective-aware convolution module that directs the convolutional kernel to capture features along the perspective lines. This enables the module to extract perspective-aware features from images, improving the feature matching. Finally, we crafted a specific 2D CNN that fuses image priors, thereby integrating keyframes and geometric metadata within the cost volume to evaluate depth planes. Our method represents the first attempt to embed the existing physical model knowledge into a network for completing MVS tasks, which achieved optimal performance using multiple benchmark datasets. Full article

With the rapid advancements in computer vision and deep learning, multi-view stereo (MVS) based on conventional RGB cameras has emerged as a promising and cost-effective tool for botanical research. However, existing methods often struggle to capture the intricate textures and fine edges of plants, resulting in suboptimal 3D reconstruction accuracy. To overcome this challenge, we proposed Edge_MVSFormer on the basis of TransMVSNet, which particularly focuses on enhancing the accuracy of plant leaf edge reconstruction. This model integrates an edge detection algorithm to augment edge information as input to the network and introduces an edge-aware loss function to focus the network’s attention on a more accurate reconstruction of edge regions, where depth estimation errors are obviously more significant. Edge_MVSFormer was pre-trained on two public MVS datasets and fine-tuned with our private data of 10 model plants collected for this study. Experimental results on 10 test model plants demonstrated that for depth images, the proposed algorithm reduces the edge error and overall reconstruction error by 2.20 ± 0.36 mm and 0.46 ± 0.07 mm, respectively. For point clouds, the edge and overall reconstruction errors were reduced by 0.13 ± 0.02 mm and 0.05 ± 0.02 mm, respectively. This study underscores the critical role of edge information in the precise reconstruction of plant MVS data. Full article

This paper presents a novel methodology to generate Digital Elevation Models (DEMs) in flat areas, incorporating river channels from relatively coarse initial data. The technique primarily utilizes filtered dense point clouds derived from SfM-MVS (Structure from Motion-Multi-View Stereo) photogrammetry of available crewed aerial imagery datasets. The methodology operates under the assumption that the aerial survey was carried out during low-flow or drought conditions so that the dry (or almost dry) riverbed is detected, although in an imprecise way. Direct interpolation of the detected elevation points yields unacceptable river channel bottom profiles (often exhibiting unrealistic artifacts) and even distorts the floodplain. In our Fluvial Domain Method, channel bottoms are represented like “highways”, perhaps overlooking their (unknown) detailed morphology but gaining in general topographic consistency. For instance, we observed an 11.7% discrepancy in the river channel long profile (with respect to the measured cross-sections) and a 0.38 m RMSE in the floodplain (with respect to the GNSS-RTK measurements). Unlike conventional methods that utilize active sensors (satellite and airborne LiDAR) or classic topographic surveys—each with precision, cost, or labor limitations—the proposed approach offers a more accessible, cost-effective, and flexible solution that is particularly well suited to cases with scarce base information and financial resources. However, the method’s performance is inherently limited by the quality of input data and the simplification of complex channel morphologies; it is most suitable for cases where high-resolution geomorphological detail is not critical or where direct data acquisition is not feasible. The resulting DEM, incorporating a generalized channel representation, is well suited for flood hazard modeling. A case study of the Ranchería river delta in the Northern Colombian Caribbean demonstrates the methodology. Full article

Phenotypic traits of fungi and their automated extraction are crucial for evaluating genetic diversity, breeding new varieties, and estimating yield. However, research on the high-throughput, rapid, and non-destructive extraction of fungal phenotypic traits using 3D point clouds remains limited. In this study, a smart phone is used to capture multi-view images of shiitake mushrooms (Lentinula edodes) from three different heights and angles, employing the YOLOv8x model to segment the primary image regions. The segmented images were reconstructed in 3D using Structure from Motion (SfM) and Multi-View Stereo (MVS). To automatically segment individual mushroom instances, we developed a CP-PointNet++ network integrated with clustering methods, achieving an overall accuracy (OA) of 97.45% in segmentation. The computed phenotype correlated strongly with manual measurements, yielding R² > 0.8 and nRMSE < 0.09 for the pileus transverse and longitudinal diameters, R² = 0.53 and RMSE = 3.26 mm for the pileus height, R² = 0.79 and nRMSE = 0.12 for stipe diameter, and R² = 0.65 and RMSE = 4.98 mm for the stipe height. Using these parameters, yield estimation was performed using PLSR, SVR, RF, and GRNN machine learning models, with GRNN demonstrating superior performance (R² = 0.91). This approach was also adaptable for extracting phenotypic traits of other fungi, providing valuable support for fungal breeding initiatives. Full article

Aiming to address the issue that existing multi-view stereo reconstruction methods have insufficient adaptability to the repetitive and weak textures in multi-view images, this paper proposes a three-dimensional (3D) reconstruction algorithm based on Feature Enhancement and Weight Optimization MVSNet (Abbreviated as FEWO-MVSNet). To obtain accurate and detailed global and local features, we first develop an adaptive feature enhancement approach to obtain multi-scale information from the images. Second, we introduce an attention mechanism and a spatial feature capture module to enable high-sensitivity detection for weak texture features. Third, based on the 3D convolutional neural network, the fine depth map for multi-view images can be predicted and the complete 3D model is subsequently reconstructed. Last, we evaluated the proposed FEWO-MVSNet through training and testing on the DTU, BlendedMVS, and Tanks and Temples datasets. The results demonstrate significant superiorities of our method for 3D reconstruction from multi-view images, with our method ranking first in accuracy and second in completeness when compared to the existing representative methods. Full article

Plant genomics have progressed significantly due to advances in information technology, but phenotypic measurement technology has not kept pace, hindering plant breeding. As maize is one of China’s three main grain crops, accurately measuring plant height is crucial for assessing crop growth and productivity. This study addresses the challenges of plant segmentation and inaccurate plant height extraction in maize populations under field conditions. A three-dimensional dense point cloud was reconstructed using the structure from motion–multi-view stereo (SFM-MVS) method, based on multi-view image sequences captured by an unmanned aerial vehicle (UAV). To improve plant segmentation, we propose a column space approximate segmentation algorithm, which combines the column space method with the enclosing box technique. The proposed method achieved a segmentation accuracy exceeding 90% in dense canopy conditions, significantly outperforming traditional algorithms, such as region growing (80%) and Euclidean clustering (75%). Furthermore, the extracted plant heights demonstrated a high correlation with manual measurements, with R² values ranging from 0.8884 to 0.9989 and RMSE values as low as 0.0148 m. However, the scalability of the method for larger agricultural operations may face challenges due to computational demands when processing large-scale datasets and potential performance variability under different environmental conditions. Addressing these issues through algorithm optimization, parallel processing, and the integration of additional data sources such as multispectral or LiDAR data could enhance its scalability and robustness. The results demonstrate that the method can accurately reflect the heights of maize plants, providing a reliable solution for large-scale, field-based maize phenotyping. The method has potential applications in high-throughput monitoring of crop phenotypes and precision agriculture. Full article

This study addresses the prevalent challenges of inefficiency and suboptimal quality in indoor 3D scene generation and rendering by proposing a parameter-tuning strategy for 3D Gaussian Splatting (3DGS). Through a systematic quantitative analysis of various performance indicators under differing resolution conditions, threshold settings for the average magnitude of spatial position gradients, and adjustments to the scaling learning rate, the optimal parameter configuration for the 3DGS model, specifically tailored for indoor modeling scenarios, is determined. Firstly, utilizing a self-collected dataset, a comprehensive comparison was conducted among COLLI-SION-MAPping (abbreviated as COLMAP (V3.7), an open-source software based on Structure from Motion and Multi-View Stereo (SFM-MVS)), Context Capture (V10.2) (abbreviated as CC, a software utilizing oblique photography algorithms), Neural Radiance Fields (NeRF), and the currently renowned 3DGS algorithm. The key dimensions of focus included the number of images, rendering time, and overall rendering effectiveness. Subsequently, based on this comparison, rigorous qualitative and quantitative evaluations are further conducted on the overall performance and detail processing capabilities of the 3DGS algorithm. Finally, to meet the specific requirements of indoor scene modeling and rendering, targeted parameter tuning is performed on the algorithm. The results demonstrate significant performance improvements in the optimized 3DGS algorithm: the PSNR metric increases by 4.3%, and the SSIM metric improves by 0.2%. The experimental results prove that the improved 3DGS algorithm exhibits superior expressive power and persuasiveness in indoor scene rendering. Full article

Reflectance anisotropy in remote sensing images can complicate the interpretation of spectral signature, and extracting precise structural information under these pixels is a promising approach. Low-altitude unmanned aerial vehicle (UAV) systems can capture high-resolution imagery even to centimeter-level detail, potentially simplifying the characterization of leaf anisotropic reflectance. We proposed a novel maize point cloud generation method that combines an advanced UAV cross-circling oblique (CCO) photography route with the Structure from the Motion-Multi-View Stereo (SfM-MVS) algorithm. A multi-spectral point cloud was then generated by fusing multi-spectral imagery with the point cloud using a DSM-based approach. The Rahman–Pinty–Verstraete (RPV) model was finally applied to establish maize leaf-level anisotropic reflectance models. Our results indicated a high degree of similarity between measured and estimated maize structural parameters (R² = 0.89 for leaf length and 0.96 for plant height) based on accurate point cloud data obtained from the CCO route. Most data points clustered around the principal plane due to a constant angle between the sun and view vectors, resulting in a limited range of view azimuths. Leaf reflectance anisotropy was characterized by the RPV model with R² ranging from 0.38 to 0.75 for five wavelength bands. These findings hold significant promise for promoting the decoupling of plant structural information and leaf optical characteristics within remote sensing data. Full article