MINISTRY OF EDUCATION AND TRAINING HO CHI MINH CITY UNIVERSITY OF TECHNOLOGY AND EDUCATION GRADUATION THESIS MAJOR: COMPUTER ENGINEERING TECHNOLOGY GENERATE 3D MODEL OF OBJECTS FROM IMAGES INSTRUCTOR: TRUONG NGOC SON STUDENT: TANG HIEP VY QUI Ho Chi Minh city, 7/2024 HO CHI MINH CITY UNIVERSITY OF TECHNOLOGY AND EDUCATION FALCUTY OF INTERNATIONAL EDUCATION GRADUATION PROJECT COMPUTER ENGINEERING TECHNOLOGY GENERATE 3D MODEL OF OBJECTS FROM IMAGES Student name: Tăng Hiệp Vy Quí Student ID: 19119055 Advisor: Associate professor PhD Trương Ngọc Sơn HO CHI MINH CITY – 7/2024 HO CHI MINH CITY UNIVERSITY OF TECHNOLOGY AND EDUCATION FALCUTY OF INTERNATIONAL EDUCATION GRADUATION PROJECT COMPUTER ENGINEERING TECHNOLOGY GENERATE 3D MODEL OF OBJECTS FROM IMAGES Student name: Tăng Hiệp Vy Quí Student ID: 19119055 Advisor: Associate professor PhD Trương Ngọc Sơn HO CHI MINH CITY – 7/2024 THE SOCIALIST REPUBLIC OF VIETNAM Independence – Freedom– Happiness -------- Ho Chi Minh City, June 24, 2024 GRADUATION PROJECT ASSIGNMENT Student Name: Tăng Hiệp Vy Quí Student ID: 19119055 Major: Computer Engineering Technology Class: 19119CLA Advisor: Associate professor PhD Trương Ngọc Sơn Date of assignment: / /2024 Date of submission : / /2024 1. Project title: Generate 3D model of objects from images 2. Initial materials provided by the advisor: None 3. Content of the project: - Build a system to collect data images base on photogrammetry.
- Control exposure level for high detail quality - Build a pipeline to create structure from motion then generate point cloud and create surface - Images collected and reconstructed in different Structure from motion software - Compare benchmark against existing methods. Final product: Apply the Structure from Motion pipeline to reconstruct images into a 3D model. The final product fulfils the original purpose: When a set of images is captured using photogrammetry theory and input into open-source 3D reconstruction software, the system will output a corresponding high-resolution model. CHAIR OF THE PROGRAM ADVISOR (Sign with full name) (Sign with full name) THE SOCIALIST REPUBLIC OF VIETNAM Independence – Freedom– Happiness -------- Ho Chi Minh City, June 24, 2024 ADVISOR’S EVALUATION SHEET Student Name: Tăng Hiệp Vy Quí Student ID: 19119055 Major: Computer Engineering Technology Class: 19119CLA Project title: Generate 3D model of objects from images Advisor: Associate professor PhD Trương Ngọc Sơn EVALUATION 1.
Content of the project :. Approval for oral defense? (Approved or denied). Ho Chi Minh City, ………… , 2024 ADVISOR (Sign with full name) THE SOCIALIST REPUBLIC OF VIETNAM Independence – Freedom– Happiness -------- Ho Chi Minh City, July 3, 2024 PRE-DEFENSE EVALUATION SHEET Student name: Tăng Hiệp Vy Quí Student ID: 19119055 Major: Computer Engineering Technology Project title: Generate 3D model of objects from images Name of Reviewer: Phan Van Ca EVALUATION 1. Content and workload of the project Design and implement a prototype of a photogrammetric acquisition system.
Strengths: The proposed system model in the thesis is coherent, and the system prototype functions properly. Weaknesses: The thesis has weaknesses in its design and implementation methodology, with many similar works existing in the literature. Additionally, the thesis is not well organized or well written, and it contains several paragraphs copied from the online articles without proper citation. Approval for oral defense? (Approved or denied) Approved 5.
Overall evaluation: (Excellent, Good, Fair, Poor) Good 6.0 (in words: Eight point Zero) Ho Chi Minh City, month July day 3, year 2024 REVIEWER (Sign with full name) PHAN VAN CA THE SOCIALIST REPUBLIC OF VIETNAM Independence – Freedom– Happiness -------- Ho Chi Minh City, June 24, 2024 EVALUATION SHEET OF DEFENSE COMMITTEE MEMBER Student name:. Content of the project:. Approval for oral defense? (Approved or denied) .) Ho Chi Minh City…………………, ……. REVIEWER (Sign with full name) ACKNOWLEDGEMENT To complete this project, we would like to express our heartfelt gratitude to Associate professor PhD Trương Ngọc Sơn has enthusiastically guided and equipped us with the necessary helpful knowledge.
Furthermore, he has directly guided and created all conditions to help us during the process of the project. Also, we would like to express our sincere thanks to all the teachers in the Department of Computer Engineering - Communication, who have enthusiastically guided and equipped us with the necessary knowledge useful in the past semesters. Student Tang Hiep Vy Qui i DISCLAIMER The simulations and study findings are accurate and were carried out entirely under the direction of the instructor, Associate professor PhD Trương Ngọc Sơn. The report does not duplicate any other sources either.
Additionally, the paper includes a variety of cited and carefully labeled reference materials. Before the department, faculty, and school, I would like to fully accept responsibility for this promise. Student Tang Hiep Vy Qui ii ABSTRACT This paper introduces a 3D scanning system utilizing the photogrammetry data collection method to extract point clouds and reconstruct 3D models. Leveraging photogrammetry principles, the system captures and processes images from multiple viewpoints, extracting key features to generate a dense point cloud representation of the scanned object or scene.
Subsequently, the point cloud data is utilized to reconstruct a detailed 3D model, encompassing both the shape and intricate structure of the subject. This approach offers a versatile and efficient solution for capturing and modeling real-world objects or environments with high precision and accuracy, fostering applications across various fields such as industrial design, cultural heritage preservation, and virtual reality content creation. iii TABLES OF CONTENT ACKNOWLEDGEMENT. III LIST OF FIGURES.
VI LIST OF TABLES. viii LIST OF ABREVIATION. ix CHAPTER 1: INTRODUCTION. 3 SUBJECTS AND SCOPE OF RESEARCH.
3 CHAPTER 2: THEORETICAL BASIC .1 INTRODUCTION TO COMPUTER VISION .3 INTRODUCTION TO 3D ACQUISITION .5 BASIC CONCEPTS OF IMAGE FORMATION .1 IMAGE FORMATION AND RADIOMETRY .2 SHAPE FROM SHADING .6 CAMERA SELF-CALIBRATION .7 STRUCTURE FROM MOTION. 26 iv CHAPTER 3: SYSTEM DESIGN AND IMPLEMENTATION .1 A 3D PHOTOGRAMMETRY SCANNER SYSTEM .2 IMAGE PRE-PROCESSING WITH CAPTURE ONE SOFTWARE 34 3.3 SFM SOFTWARE : MESHROOM .4 CAMERA CALIBRATION AND TRANSMISSION .1 3D PHOTOGRAMMETRY SCANNER SYSTEM. PRE-PROCESSING IMAGE RESULT. STRUCTURE FROM MOTION RESULT.
REFINE & EXPORT IN BLENDER. 42 CHAPTER 5: CONCLUSION AND FUTURE WORK. 45 v LIST OF FIGURES Figure 2.1 Human visual system vs.2 Computer vision and related disciplines .5 Image formation process.6 Digital image formation process .7 Plane angle in 2D & Solid Angle in 3D .8 Solid angle in spherical coordinate system .9 Image formation process in a camera .10 Different shadings give a cue for the actual 3D shape .11 Geometry in 2D-image generation .12 Type of projection .14 All types of projections in graphical .15 Camera Self-Calibration Framework .16 Two camera are indicated by their centres and camera baseline intersects each image plane .17 SfM system flow chart .18 Example of a point cloud from model .19 Aerial Photogrammetry demonstration .20 Manual exposure info-graphic .21 Camera movement to capture for an object .2 Photogrammetry Scanner system diagram .3 System hardware component.6 Process multiple images in Capture One interface .7 User interface in open source Meshroom software .8 Photogrammetry pipeline in open source Meshroom software .9 Camera capture and instantly transmission .1 Sample collected from the first stage of system .2 Sample Detail & Texture adjustment before import to Meshroom .3 Result of reconstruct sparse cloud with camera position Meshroom 40 Figure 4.4 Point Cloudd result from Meshroom .4 Final Texture result from Meshroom .5 After refine in Blender. 42 vii LIST OF TABLES Table 3.3 : Feature Describer type and use case.
36 viii LIST OF ABREVIATION Abbreviation Description AI Artificial Intelligence AKAZE Accelerated KAZE FAST Features from Accelerated Segment Test KAZE Non-linear Scale Space Feature Detection and Description (Named after the Japanese word "Kaze" which means "wind") LIDAR Light Detection and Ranging MVS Multiview Stereo OCR Optical Character Recognition PSF Point Spread Function SFM Structure from motion SIFT Scale-Invariant Feature Transform SURF Speeded-Up Robust Features ix CHAPTER 1: INTRODUCTION INTRODUCTION In recent years, the field of computer vision has witnessed remarkable advancements, enabling the development of sophisticated algorithms and techniques for extracting valuable information from images. One of the fascinating applications of computer vision is the generation of three-dimensional (3D) models from two-dimensional (2D) images. This capability holds immense potential in various domains, including augmented reality, virtual reality, robotics, archaeology, and cultural heritage preservation. The process of generating 3D models from images, often referred to as "photogrammetry" involves the reconstruction of the spatial geometry and appearance of real-world objects or scenes using computational methods.
Traditional photogrammetry techniques rely on manual measurements, stereo vision, or structured light scanning, which can be labor-intensive, time-consuming, and often require specialized hardware setups. This graduation thesis aims to explore and investigate the process of generating high-quality 3D models of objects from images using state-of-the-art computer vision techniques. Specifically, the focus will be on leveraging workflow for accurate and efficient reconstruction of object geometry and appearance from a collection of 2D images. The project will delve into the following key aspects: Image Acquisition and Preprocessing: The acquisition of image data and preprocessing steps to enhance the quality and suitability of input images for 3D reconstruction.
Feature Extraction and Matching: Techniques for extracting distinctive features from images and establishing correspondences across multiple views, essential for robust reconstruction. 1 Surface Reconstruction: Methods for generating a coherent 3D surface mesh from the estimated depth maps, incorporating techniques for mesh regularization and refinement. Texture Mapping and Rendering: Strategies for mapping texture information from input images onto the reconstructed 3D model and rendering realistic visualizations. The project will also include practical implementation, experimentation, and evaluation of the proposed methodologies using real-world datasets and benchmarking against existing approaches.
The goal is to contribute to the advancement of 3D reconstruction techniques and facilitate their adoption in various applications requiring accurate and efficient generation of 3D models from images. In conclusion, this graduation project endeavors to provide insights into the exciting realm of 3D reconstruction from images using computer vision, showcasing the potential of deep learning approaches to revolutionize this field and pave the way for novel applications and advancements in technology. OBJECTIVES - Build a system to collect data images base on photogrammetry. - Control exposure level for high detail quality - Build a pipeline to create structure from motion then generate point cloud and create surface - Images collected and reconstructed in different Structure from motion software - Compare benchmark against existing methods.
2 TOPIC LIMITATION The topic only focuses on researching the application of capturing images at many different angles of an object to create points in space and form a plane connected to each other and then exporting into corresponding formats. compatible with 3D editing software. RESEARCH METHODS - Data Collection: Gather diverse image datasets and ground truth data. - Preprocessing: Enhance image quality and consistency.
- Feature Extraction and Matching: Evaluate algorithms for identifying key points and establishing correspondences. - Surface Reconstruction: Implement algorithms for generating 3D surface meshes and refine surface quality. - Texture Mapping and Rendering: Develop methods for texture extraction, mapping, and realistic rendering. - Evaluation: Assess reconstruction accuracy quantitatively and qualitatively.
- Performance Analysis: Measure computational efficiency and scalability. - Comparison with Baselines: Benchmark against existing methods. SUBJECTS AND SCOPE OF RESEARCH - Pipeline and workflow for data collection, pre-processing raw data, feature extraction, depth analysis, plane reconstruction, REPORT LAYOUT Chapter 1: Introduction: Provides an overview of the topic, research objectives, scope, research methods, and the target audience. Chapter 2: Theoretical Basis: Introduces the current state of research, research directions, and relevant services being utilized.
Chapter 3: System Design and Implementation: Presents the general model of the entire system, the system components, detailed design of each component, and the devices used in the components. 3 Chapter 4: Results: Presents the execution results of the system model. Compare with other different method Chapter 5: Conclusion and Future Development: Draws conclusions, highlights strengths and weaknesses, and suggests directions for future development of the model.