Introduction and Motivation
- 3D reconstruction determines the 3D profile of objects and the 3D coordinates of points on the profile.
- It is a core technology in various fields such as Computer Aided Geometric Design (CAGD), computer graphics, computer vision, medical imaging, and virtual reality.
- It offers a new and accurate approach in diagnosis, particularly in presenting lesion information in 3D for clinical value.
- Digital elevation models can be reconstructed using methods like airborne laser altimetry and synthetic aperture radar.
Active and Passive Methods
Active methods:
- Use range data to reconstruct the 3D profile of objects by numerical approximation.
- Interfere with the object using techniques like structured light, laser range finders, and time-of-flight lasers.
- Acquire depth maps by actively measuring distance or emitted and reflected radiance from the object.
- Applicable in various scenarios such as computer stereo vision and 3D ultrasound.
- Enable accurate reconstruction but may require complex setups and computations.
Passive methods:
- Do not interfere with the object and rely on sensors to measure the radiance reflected or emitted by the object's surface.
- Use image sensors in cameras sensitive to visible light and process digital images or videos.
- Can be applied in a wider range of situations compared to active methods.
- Used for image-based reconstruction and output 3D models.
- Commonly used in computer vision and computer graphics.
Monocular Cues and Stereo Vision
Monocular cues methods:
- Use one or more images from a single viewpoint to reconstruct 3D shapes.
- Measure 3D shape using 2D characteristics such as silhouettes, shading, and texture.
- Techniques include shape-from-shading, photometric stereo, and shape-from-texture.
Stereo vision:
- Obtains 3D geometric information of an object from multiple images.
- Uses two cameras simultaneously or a single camera at different times to restore the object's 3D profile and location.
- Requires careful calibration and matching of image points to calculate depth information from disparity.
Techniques and Algorithms for 3D Reconstruction
- Various techniques and algorithms are used for 3D reconstruction, including real-time non-rigid reconstruction, 3D building model reconstruction, investigating landslides, estimating pavement roughness, and shape-from-X.
- Algorithms for 3D reconstruction include Delaunay triangulation, marching cubes, shape from shading, and photometric methods.
- Challenges include loss of geometry precision and complexity of post-processing techniques.
Applications of 3D Reconstruction
- 3D reconstruction has various applications, including 360 depth estimation for virtual reality, car shape reconstruction, translation with conditional vector-quantised code diffusion, learning cultural heritage, and digital archaeological exhibition.
- It is also used in computer vision, robotics, medical imaging, and surface reconstruction.
- Applications in medical imaging facilitate clinical routine, surgical planning, patient follow-up, and various clinical areas such as radiotherapy planning and neurointerventions.
This article needs to be updated. (October 2019) |
In computer vision and computer graphics, 3D reconstruction is the process of capturing the shape and appearance of real objects. This process can be accomplished either by active or passive methods. If the model is allowed to change its shape in time, this is referred to as non-rigid or spatio-temporal reconstruction.
