Three-dimensional Information Research Articles

Research on visualization of environmental landscape design based on digital entertainment platform and immersive VR experience

With the rapid development of digital entertainment platforms and immersive virtual reality (VR) technology, the visualization of environmental landscape design has attracted widespread attention. This study aims to explore the visualization methods and effects of environmental landscape design based on digital entertainment platforms and immersive VR experiences. By capturing three-dimensional point information in the environment, high-precision environmental data can be obtained. Next, convert the point cloud data into a digital model with terrain features, and use voxel scene construction technology to further transform the terrain model into a voxelated digital model. In the design process of landscape design systems based on digital entertainment platforms and immersive VR experiences, various interaction methods have been implemented, such as gesture recognition, head tracking, and touch control, so that users can freely browse and operate virtual landscapes. By introducing landscape design principles and rules, users can modify and customize the landscape in a virtual environment. The research results indicate that the visualization method of environmental landscape design based on digital entertainment platforms and immersive VR experience can effectively display the characteristics and details of the environmental landscape, improve user perception and participation.

Enhancing Terrain Analysis from Digital Elevation Models Using 2-D Kalman Filtering Technique

Three-dimensional spatial information, particularly elevation, is crucial for understanding terrain characteristics essential for meaningful development, often expressed as a Digital Elevation Model (DEM). To achieve reliable and accurate DEM values for terrain analysis, modeling uncertainties is necessary. The primary objective of this study is to determine improved terrain variables from the Digital Elevation Model of the study area. The recursive 2-D Kalman filtering technique was applied four times at different orientations to 121 elevation values extracted from a 30-meter resolution ALOS DEM of the study area using QGIS Desktop 3.22.7 software of an area covering approximately 10.80 Hectares using QGIS, the process involved 144 iterations. MATLAB was used for the computations. The terrain variables (elevation, first partial derivatives along the X and Y axes) of the central point of the DEM were obtained as a linear combination of the four filtering results. The final estimated values for the central point were 26.5589m for elevation, 0.0002m and 0.0011m for partial derivatives along the X and Y directions, with standard errors of ±0.0001m, ±0.0005m, and ±0.0007m, respectively. A 3-D plot of the terrain surface of the study area using Surfer10 software showed that the recursive 2-D Kalman filtering significantly improved the quality of the terrain surface when applied over the DEM. Therefore, the adopted recursive 2-D Kalman filter is well-suited for terrain surface modeling using grid DEMs. Its use is encouraged for determining improved values of terrain topographic variables, leading to more accurate terrain interpretation. In addition, when compared with ground survey data confirmed the technique's efficiency in reducing DEM noise. These results are promising as they are necessary information for flood route modelling, land use allocation and enhance functionality of the urban space domain of the study area.

A New Paradigm in the Management of Knee Osteoarthritis and Arthroplasty with Dynamic Patient-source Outcome Measures: Comprehensive Clinical Review of the Knee Kinesiography Exam with the KneeKG® System

While knee disorders, and more specifically knee osteoarthritis, account for a significant portion of the orthopedic visits to health care providers, there are still clinical gaps that need to be addressed to achieve effective management of this disease. The absence of functional tools to objectively assess knee function in motion and weight-bearing conditions poses a challenge for clinicians to hindering their ability to design individualized conservative and surgical treatment plans aimed at achieving better patient outcomes. The Knee Kinesiography exam with the KneeKG® system provides accurate, reliable, three-dimensional, and real-time information about dynamic knee malalignment and joint dysfunction during weight-bearing activities through the assessment of kinematic biomechanical markers. This work presents a comprehensive clinical review of the materials related to this exam which have been published since its validation in 2012. The aim is to summarize literature evidence on how the Knee Kinesiography exam may enhance both non-surgical and surgical management of knee osteoarthritis and arthroplasty. The access to objective patient-source outcome measures from this exam, directly associated with both patient-reported outcomes and osteoarthritis onset and progression before and after total knee arthroplasty, can allow clinicians to identify mechanisms behind knee pain, dysfunction, and dissatisfaction. In the era of orthopedic value-based care, such patient-source data may help clinicians better understand the condition of the knee and thus assist them in their decision-making to establish the best personalized care plan possible throughout the continuum of care, from diagnosis to post-surgery management.

Redefining Accuracy: Underwater Depth Estimation for Irregular Illumination Scenes.

Acquiring underwater depth maps is essential as they provide indispensable three-dimensional spatial information for visualizing the underwater environment. These depth maps serve various purposes, including underwater navigation, environmental monitoring, and resource exploration. While most of the current depth estimation methods can work well in ideal underwater environments with homogeneous illumination, few consider the risk caused by irregular illumination, which is common in practical underwater environments. On the one hand, underwater environments with low-light conditions can reduce image contrast. The reduction brings challenges to depth estimation models in accurately differentiating among objects. On the other hand, overexposure caused by reflection or artificial illumination can degrade the textures of underwater objects, which is crucial to geometric constraints between frames. To address the above issues, we propose an underwater self-supervised monocular depth estimation network integrating image enhancement and auxiliary depth information. We use the Monte Carlo image enhancement module (MC-IEM) to tackle the inherent uncertainty in low-light underwater images through probabilistic estimation. When pixel values are enhanced, object recognition becomes more accessible, allowing for a more precise acquisition of distance information and thus resulting in more accurate depth estimation. Next, we extract additional geometric features through transfer learning, infusing prior knowledge from a supervised large-scale model into a self-supervised depth estimation network to refine loss functions and a depth network to address the overexposure issue. We conduct experiments with two public datasets, which exhibited superior performance compared to existing approaches in underwater depth estimation.

Integration of line-field confocal optical coherence tomography and in situ microenvironmental mapping to investigate the living microenvironment of reconstructed human skin and melanoma models

BackgroundIn tissue engineering, real-time monitoring of tumors and of the dynamics of the microenvironment within in vitro models has traditionally been hindered by the need to harvest the cultures to obtain material to analyze. Line-field confocal optical coherence tomography (LC-OCT) has proven to be useful in evaluating in vivo skin conditions, including melanoma, by capturing dynamic, three-dimensional (3D) information without the need for invasive procedures, such as biopsies. Additionally, the M-Duo Technology® developed by IMcoMET presents a unique opportunity for continuous in situ biomarker sampling, providing insights into local cellular behavior and interactions. ObjectiveThis study aimed to validate the non-destructive mapping capabilities of two advanced methodologies (LC-OCT by DAMAE Medical and M-Duo Technology® by IMcoMET) to investigate the living microenvironment of in vitro reconstructed human skin (RhS) and melanoma-RhS (Mel-RhS). MethodsLC-OCT and M-Duo Technology® were compared to conventional analysis of the RhS and Mel-RhS microenvironments. ResultsLC-OCT successfully visualized the distinct layers of the epidermis and tumor structures within the Mel-RhS, identifying keratinocytes, melanocytes, tumor nests, and fibroblasts. The M-Duo Technology® revealed differences in in situ cytokine (IL-6) and chemokine (CCL2, CXCL10, and IL-8) secretion between Mel-RhS and the control RhS. Notably, such differences were not detected through conventional investigation of secreted proteins in culture supernatants. ConclusionThe combination of LC-OCT's high-resolution imaging and M-Duo Technology®’s in situ microenvironmental mapping has the potential to provide a synergistic platform for non-invasive, real-time analysis, allowing for prolonged observation of processes within Mel-RhS models without the need for culture disruption.

A Review of Enhancement Techniques for Cone Beam Computed Tomography Images

Cone Beam Computed Tomography (CBCT) has emerged as a valuable imaging modality for various medical applications due to its ability to provide three-dimensional information with minimal radiation exposure. However, CBCT images often suffer from inherent limitations, such as increased noise, artifacts, and reduced spatial resolution. This paper presents a comprehensive review of image processing techniques employed to enhance the quality of CBCT images, addressing the challenges posed by acquisition hardware and image reconstruction algorithms. The review covers a range of preprocessing and post-processing methods, including denoising, artifact correction, and resolution improvement techniques. These methods encompass various mathematical algorithms, machine learning approaches, and hybrid models, which aim to mitigate the imperfections present in CBCT data while preserving diagnostically relevant information. Additionally, this paper discusses the application of deep learning methods, convolutional neural networks, and generative adversarial networks in CBCT image enhancement. These advanced techniques have shown promise in tackling the complex nature of CBCT data and optimizing image quality.

Selection of hand features based on Random Forest algorithm and hand shape recognition

To obtain effective features applicable to hand morphology recognition, the method of obtaining effective feature indicators based on the Random Forest (RF) algorithm to downscale hand morphology parameters is proposed. Firstly, 232 female university students collected three-dimensional hand information, constructed auxiliary point, line, and surface standardised measurement methods, obtained 33 characteristic parts of human dimensions, and used k-means clustering for hand morphology subdivision. Hand morphology can be divided into three categories: short, slender and broad. The RF algorithm is used for feature index importance assessment and hand shape recognition model. The accuracy of the feature metrics determined by the RF algorithm, PCA, and VC method applied to hand shape recognition is compared and analysed to verify the effectiveness of the dimensionality reduction of the RF algorithm. The results showed that the feature indexes used for hand shape recognition were five items: hand length, thickness at the metacarpal, thenar width, the distance between the thumb and index finger, and distance from the root of the little finger to the centre of the wrist. Using the RF algorithm to reduce the dimensionality is more effective; the average recognition accuracy of the four hand shape recognition models is 93.78% on average, compared with PCA and VC reduction methods, the average accuracy of hand shape recognition models is increased by 19.17%, and 14.86% respectively. The study's results can provide methodological references for the objective selection of characteristic indicators and morphological recognition of human body parts.

Applicability of non-invasive and live-cell holotomographic imaging on fungi

The ability to acquire three-dimensional (3D) information of cellular structures without the need for fluorescent tags or staining makes holotomographic imaging a powerful tool in cellular biology. It provides valuable insights by measuring the refractive index (RI), an optical parameter describing the phase delay of light that passes through the living cell.Here, we demonstrate holotomographic imaging on industrial relevant ascomycete fungi and study their development and morphogenesis. This includes conidial germination, subcellular dynamics, and cytoplasmic flow during hyphal growth in Aspergillus niger. In addition, growth and budding of Aureobasidium pullulans cells are captured using holotomographic microscopy. Coupled to fluorescence imaging, lipid droplets, vacuoles, the mitochondrial network, and nuclei are targeted and analyzed in the 3D RI reconstructed images. While lipid droplets and vacuoles can be assigned to a specific RI pattern, mitochondria and nuclei were not pronounced. We show, that the lower sensitivity of RI measurements derives from the fungal cell wall that acts as an additional barrier for the illumination light of the microscope. After cell wall digest of hyphae and protoplast formation of A. niger expressing GFP-tagged histone H2A, location of nuclei could be determined by non-invasive RI measurements. Furthermore, we used coupled fluorescence microscopy to observe migration of nuclei in unperturbed hyphal segments and duplication during growth on a single-cell level. Detailed micromorphological studies in Saccharomyces cerevisiae and Trichoderma reesei are challenging due to cell size restrictions.Overall, holotomography opens up new avenues for exploring dynamic cellular processes in real time and enables the visualization of fungi from a new perspective.

Streamlining BIM models for GIS integration: an advanced approach to building envelope extraction

ABSTRACT Building Information Modelling (BIM) models are crucial for urban building modeling in three-dimensional (3D) Geographic Information Systems (GIS); however, their intricate nature requires lightweight design for seamless GIS integration. Envelope extraction is a primary method for achieving lightweight BIM; nevertheless, it encounters challenges related to robustness and simplification. To tackle this issue, this study proposes a more effective envelope extraction method. Initially, BIM elements are categorized into three types based on their semantics: space-occupying, portal, and other facility elements. Subsequently, for space-occupying elements, a 3D alpha wrapping algorithm is employed to obtain a tightly enveloped triangular mesh and then simplified into a lightweight polygonal mesh. Finally, portal elements are replaced with simplified proxies and embedded into a polygonal mesh through extrusion operations to refine the model. The refined model is combined with other facility elements to achieve the final simplification result. Experiments were conducted using publicly available datasets and manually created BIM models of various types and complexities. The experimental results demonstrate that our method significantly reduces the data volume of BIM models while maintaining their accuracy. Furthermore, this approach does not necessitate strict geometric requirements for BIM building elements and allows for gaps between them, thereby exhibiting higher robustness.

Development and implementation of the course entitled "Virtual Simulation Experiment of Recombinant Human Erythropoietin Manufacturing Process"

Considering the issues present in traditional learning methods of manufacturing process for biotechnology majors, this paper presents the development and implementation process of the course entitled "Virtual Simulation Experiment of Recombinant Human Erythropoiesis Manufacturing Process". The experiment combines modern biological manufacturing technology and three-dimensional information technology, with recombinant human erythropoiesis drug serving as the focal point. This paper elaborates on the teaching concepts, objectives, contents, implementation methods, experimental procedures, interactive steps, and assessment criteria used in the experiment. Through innovative experimental scheme design, teaching methodologies, and evaluation systems, this course aims to cultivate students' analytical and problem-solving skills in the field of biopharmaceutical engineering, while also broadening students' perspective and expanding their vision.

Object–Environment Fusion of Visual System for Automatic Pear Picking

A visual system is a key tool for automatic fruit harvesting. It faces the challenges of varied occlusion and illumination in the field, resulting in difficulties in fruit recognition and picking. At present, there are many studies that lack a comprehensive analysis of the impact of the environment on harvesting. This study proposes an object–environment fusion visual system. It comprises modules for object perception, environment perception, and picking pose estimation. The object perception module aims to identify and locate pears. The environment perception module is deployed to analyze the three-dimensional (3D) information of objects and obstacles. Finally, the picking pose estimation module fuses the information of the objects and environment to calculate the collision-free picking position and orientation. Additionally, specific implementations are employed for each module. It compares three networks to identify pears for object perception. Voxel-based representation is used to simplify point clouds for environmental perception. A sampler and evaluator are applied for picking pose estimation. The S1 and S2 datasets were acquired in a laboratory pear tree model and the orchard of Zhejiang Academy of Agricultural Sciences, respectively, to evaluate the proposed method. In the S2 dataset, the success rate for picking pose estimation reached 87.11% within the distance range of 30∼50 cm. All of the results demonstrate that the proposed method can be applied to visual perception for automatic pear harvesting.

A new method for individual treetop detection with low-resolution aerial laser scanned data

In the past decade, the use of three-dimensional forest information from airborne Light Detection and Ranging (LiDAR) has become widespread in forest inventories. Accurate Individual Treetop Detection (ITD) and crown boundary delineation using LiDAR data are critical for obtaining precise inventory metrics. To address this need, we introduced a novel growing tree region (GTR)-driven ITD method that utilizes canopy height models (CHM) derived from very low-resolution airborne LiDAR data. The GTR algorithm consists of three key stages: (i) preserving all height layers through incremental cutting and stacking of CHM; (ii) employing a three-layer concept to identify individual treetops; and (iii) refining the detected treetops using a distance-based filter. Our method was tested in five temperate forests across Central Europe and was compared against the widely-used local maxima (LM) search combined with an optimized variable window filtering (VWF) technique. Our results showed that the GTR method outperformed LM with VWF, particularly in forests with high canopy density. The achieved root mean square accuracies were 74% for the matching rate, 19% for commission errors, and 27% for omission errors. In comparison, the LM with the VWF method resulted in a matching rate of 71%, commission errors of 20%, and omission errors of 31%. To facilitate the application of our algorithm, we developed an R package called TREETOPS, which seamlessly integrates with the lidR package, ensuring compatibility with existing treetop-based segmentation methods. By introducing TREETOPS, we provide the most accurate open-source tool for detecting treetops using low-resolution LiDAR-derived CHM.

DAAL-WS: A weakly-supervised method integrated with data augmentation and active learning strategies for MLS point cloud semantic segmentation

Mobile laser scanning (MLS) point clouds have increasingly been a significant data source for acquiring accurate three-dimensional (3D) semantic information from complex scenes. However, most current point cloud semantic segmentation methods heavily depend on a huge number of manually labeled training samples, which is labor-intensive and time-consuming. To address the above challenge, a weakly-supervised MLS point cloud semantic segmentation method integrated with data augmentation and active learning strategies is proposed (termed DAAL-WS). By taking advantage of our previously presented multi-branch weakly-supervised network (WSPointNet) in weakly supervisory signals and ensemble predictions, the DAAL-WS integrates WSPointNet with two essential components, i.e., an elevation-calibrated Mix3D (EC-Mix3D) data augmentation strategy and a point-level ensemble prediction-based (PLEP) active learning strategy. Specifically, the EC-Mix3D data augmentation strategy leverages elevation information to calibrate sub-point clouds and generates new contextual scenes by mixing the elevation-calibrated sub-point clouds, thereby augmenting the training point cloud distribution. Designed to reduce labeling redundancy, the PLEP active learning strategy selects the most important labeled points for model training. This strategy first measures the uncertainty for each unlabeled point by ensemble predictions and then employs a feature-distance suppression module to select the significant and discriminating unlabeled points for manual labeling. The proposed DAAL-WS method was evaluated on three public MLS datasets, including Toronto3D, Paris-Lille-3D, and WHU-MLS datasets, on which DAAL-WS obtained a competitive performance over fully-supervised baselines using only 0.015 %, 0.03 %, and 0.03 % labeled points, with mean Intersection over Union (mIoU) scores of 81.91 %, 82.59 %, and 60.36 %, respectively.

Compact multispectral light field camera based on an inkjet-printed microlens array and color filter array.

With emerging advanced optical sensing technologies and their wide-ranging applications, gathering comprehensive optical data from real-world scenes is becoming increasingly crucial for their accurate reconstruction and analysis. In order to capture both three-dimensional (3D) spatial and spectral information from a scene, multiple devices or time-intensive scanning processes are often involved. Here, we demonstrate a multispectral light field camera that allows for the simultaneous acquisition of 3D information and spectral data in a single snapshot. By utilizing inkjet printing as the fabrication technology, the miniaturized optical components in the camera were manufactured with high precision and can be integrated into a standard camera housing. Furthermore, the microlens arrays and the color filter arrays were fabricated on the same substrate, and a precise alignment between the two arrays was achieved. The compact multispectral camera opens the door to a multitude of possibilities for mobile applications, ranging from autonomous driving and consumer electronics such as smartphones to medical technology such as endoscopes.

Insulator Extraction from UAV LiDAR Point Cloud Based on Multi-Type and Multi-Scale Feature Histogram

Insulators are key components to ensure the normal operation of power facilities in transmission corridors. Existing insulator identification methods mainly use image data and lack the acquisition of three-dimensional information. This paper proposes an efficient insulator extraction method based on UAV (unmanned aerial vehicle) LiDAR (light detection and ranging) point cloud, using five histogram features: horizontal density (HD), horizontal void (HV), horizontal width (HW), vertical width (VW) and vertical void (VV). Firstly, a voxel-based method is employed to roughly extract power lines and pylons from the original point cloud. Secondly, the VV histogram is used to categorize the pylons into suspension and tension types, and the HD histogram is used to locate the tower crossarm and further refine the roughly extracted powerlines. Then, for the suspension tower, insulators are segmented based on the HV histogram and HD difference histogram. For the tension tower, the HW histogram is used to recognize the jumper conductor (JC) and transmission conductor (TC) from the power line. The HW histogram and VW histogram are used to extract the tension insulator in the TC and suspension insulator in the JC, respectively. Finally, considering the problem of setting a suitable grid width when constructing the feature histogram, an adaptive method of multi-scale histograms is proposed to refine the extraction result. Two 220 kV long transmission lines are used for the validation, and the overall object-based accuracy for suspension and tension towers are 100% and 97.3%, respectively. Compared with the point feature-based method, the mean F1 score of the proposed method improved by 0.3, and the runtime for each tower is within 2 s.

PP-ISEA: An Efficient Algorithm for High-Resolution Three-Dimensional Geometry Reconstruction of Space Targets Using Limited Inverse Synthetic Aperture Radar Images.

As the variety of space targets expands, two-dimensional (2D) ISAR images prove insufficient for target recognition, necessitating the extraction of three-dimensional (3D) information. The 3D geometry reconstruction method utilizing energy accumulation of ISAR image sequence (ISEA) facilitates superior reconstruction while circumventing the laborious steps associated with factorization methods. Nevertheless, ISEA's neglect of valid information necessitates a high quantity of images and elongated operation times. This paper introduces a partitioned parallel 3D reconstruction method utilizing sorted-energy semi-accumulation with ISAR image sequences (PP-ISEA) to address these limitations. The PP-ISEA innovatively incorporates a two-step search pattern-coarse and fine-that enhances search efficiency and conserves computational resources. It introduces a novel objective function 'sorted-energy semi-accumulation' to discern genuine scatterers from spurious ones and establishes a redundant point exclusion module. Experiments on the scatterer model and simulated electromagnetic model demonstrate that the PP-ISEA reduces the minimum image requirement from ten to four for high-quality scatterer model reconstruction, thereby offering superior reconstruction quality in less time.

Quantum dot-enabled infrared hyperspectral imaging with single-pixel detection.

Near-infrared (NIR) hyperspectral imaging is a powerful technique that enables the capture of three-dimensional (3D) spectra-spatial information within the NIR spectral range, offering a wide array of applications. However, the high cost associated with InGaAs focal plane array (FPA) has impeded the widespread adoption of NIR hyperspectral imaging. Addressing this challenge, in this study, we adopt an alternative approach-single-pixel detection for NIR hyperspectral imaging. Our investigation reveals that single-pixel detection outperforms conventional FPA, delivering a superior signal-to-noise ratio (SNR) for both spectral and imaging reconstruction. To implement this strategy, we leverage self-assembled colloidal quantum dots (CQDs) and a digital micromirror device (DMD) for NIR spectral and spatial information multiplexing, complemented by single-pixel detection for simultaneous spectral and image reconstruction. Our experimental results demonstrate successful NIR hyperspectral imaging with a detection window about 600 nm and an average spectral resolution of 8.6 nm with a pixel resolution of 128 × 128. The resulting spectral and spatial data align well with reference instruments, which validates the effectiveness of our approach. By circumventing the need for expensive and bulky FPA and wavelength selection components, our solution shows promise in advancing affordable and accessible NIR hyperspectral imaging technologies, thereby expanding the range of potential applications.

Parameter Extraction of Accelerated Moving Targets under Non-Quasi-Axial Incidence Conditions Based on Vortex Electromagnetic Wave Radar

Vortex electromagnetic wave radar carrying orbital angular momentum can compensate for the deficiency of planar electromagnetic wave radar in detecting motion parameters perpendicular to the direction of electromagnetic wave propagation, thus providing more information for target recognition, which has become a hot research field in recent years. However, existing research makes it difficult to obtain the acceleration and rotation centers of targets under non-quasi-axial incidence conditions of vortex electromagnetic waves. Based on this, this article proposes a variable speed motion target parameter extraction method that combines single element and total element echoes. This method can achieve three-dimensional information extraction of radar targets based on a uniform circular array (UCA). Firstly, we establish a non-quasi-axis detection echo model for variable-speed moving targets and extract echoes from different array elements. Then, a single element echo is used to extract the target’s range slow time profile and obtain the target’s rotation center z coordinate. We further utilize the target linear, angular Doppler frequency shift extremum, and median information to extract parameters such as target acceleration, tilt angle, rotation radius, and rotation center x and y coordinates. We analyzed the impact of different signal-to-noise ratios and motion states on parameter extraction. The simulation results have verified the effectiveness of the proposed algorithm.

Gram matrix: an efficient representation of molecular conformation and learning objective for molecular pretraining.

Accurate prediction of molecular properties is fundamental in drug discovery and development, providing crucial guidance for effective drug design. A critical factor in achieving accurate molecular property prediction lies in the appropriate representation of molecular structures. Presently, prevalent deep learning-based molecular representations rely on 2D structure information as the primary molecular representation, often overlooking essential three-dimensional (3D) conformational information due to the inherent limitations of 2D structures in conveying atomic spatial relationships. In this study, we propose employing the Gram matrix as a condensed representation of 3D molecular structures and for efficient pretraining objectives. Subsequently, we leverage this matrix to construct a novel molecular representation model, Pre-GTM, which inherently encapsulates 3D information. The model accurately predicts the 3D structure of a molecule by estimating the Gram matrix. Our findings demonstrate that Pre-GTM model outperforms the baseline Graphormer model and other pretrained models in the QM9 and MoleculeNet quantitative property prediction task. The integration of the Gram matrix as a condensed representation of 3D molecular structure, incorporated into the Pre-GTM model, opens up promising avenues for its potential application across various domains of molecular research, including drug design, materials science, and chemical engineering.

Organ Segmentation and Phenotypic Trait Extraction of Cotton Seedling Point Clouds Based on a 3D Lightweight Network

Cotton is an important economic crop; therefore, enhancing cotton yield and cultivating superior varieties are key research priorities. The seedling stage, a critical phase in cotton production, significantly influences the subsequent growth and yield of the crop. Therefore, breeding experts often choose to measure phenotypic parameters during this period to make breeding decisions. Traditional methods of phenotypic parameter measurement require manual processes, which are not only tedious and inefficient but can also damage the plants. To effectively, rapidly, and accurately extract three-dimensional phenotypic parameters of cotton seedlings, precise segmentation of phenotypic organs must first be achieved. This paper proposes a neural network-based segmentation algorithm for cotton seedling organs, which, compared to the average precision of 75.4% in traditional unsupervised learning, achieves an average precision of 96.67%, demonstrating excellent segmentation performance. The segmented leaf and stem point clouds are used for the calculation of phenotypic parameters such as stem length, leaf length, leaf width, and leaf area. Comparisons with actual measurements yield coefficients of determination R2 of 91.97%, 90.97%, 92.72%, and 95.44%, respectively. The results indicate that the algorithm proposed in this paper can achieve precise segmentation of stem and leaf organs, and can efficiently and accurately extract three-dimensional phenotypic structural information of cotton seedlings. In summary, this study not only made significant progress in the precise segmentation of cotton seedling organs and the extraction of three-dimensional phenotypic structural information, but the algorithm also demonstrates strong applicability to different varieties of cotton seedlings. This provides new perspectives and methods for plant researchers and breeding experts, contributing to the advancement of the plant phenotypic computation field and bringing new breakthroughs and opportunities to the field of plant science research.