Real-Object 3D Scanning With Virtual Camera Guidance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D scanning methods are tedious, uncertain in quality, and inconvenient due to the need for specialized equipment or markers, and are not user-friendly for scanning large objects.
Innovation Solution
A method using a standard camera-equipped device with augmented reality guidance, where a virtual 3D box is superimposed on the object to assist in positioning and a structured virtual dome is used for efficient image capture, leveraging machine learning and computer vision for accurate 3D reconstruction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If photogrammetry technique is used with manual picture acquisition, then 3D scanning can be performed with standard cameras, but the process becomes tedious and time-consuming requiring user to take many pictures manually
Solution Approach 1:
The patent replaces the manual mechanical process of taking pictures with an automated computer vision system. The system automatically captures images from multiple viewpoints using the camera device, processes them through algorithms to identify valid pictures, and constructs the 3D model without requiring manual intervention for each picture capture step.
Solution Approach 2:
The patent changes the operational parameters by implementing automatic picture acquisition and processing. The system automatically adjusts capture parameters, selects valid pictures based on overlap and quality criteria, and proceeds with 3D reconstruction without user intervention, thereby reducing the time loss while maintaining compatibility with standard cameras.
2Ease of operation
If automated 3D scanning algorithm is used, then picture sorting and processing is simplified, but the quality outcome becomes uncertain and user control is lost
Solution Approach 1:
The patent implements feedback mechanisms where the system evaluates captured pictures against predefined quality criteria (overlap, clarity, lighting) and provides user feedback on which pictures are valid and useful. This allows the system to maintain automation while ensuring quality control through iterative validation and user confirmation of key processing decisions.
3Productivity
If AR technology with tracking marker is used, then scanning process becomes faster with algorithm resolution, but specialized equipment and printed markers are required which is inconvenient
Solution Approach 1:
The patent extracts and removes the tracking marker component from the AR scanning system. Instead of requiring printed markers or specialized AR equipment, the system uses the camera device alone to capture images and automatically processes them through computer vision algorithms, thereby eliminating the need for specialized equipment while maintaining scanning functionality.
4Measurement precision
If dome structure with many segments is used, then complete object coverage is achieved, but the height is limited by tracking marker dimensions making it hard to scan larger objects
Solution Approach 1:
The patent implements a dynamic picture acquisition system that adapts to objects of various sizes. Instead of a fixed dome structure constrained by marker dimensions, the system automatically determines the number and arrangement of pictures needed based on the object's dimensions, adjusting the scanning strategy dynamically to accommodate anything from small objects to large structures like mountain chains.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
The invention relates to a computer-implemented method for 3D scanning of a real object (OBJ) with a camera (CAM) having a 3D position, comprising the steps of: a) receiving, from the camera (CAM), an image (IM) of the real object (OBJ); b) displaying on a screen, in an augmented reality view, the image (IM) of the real object (OBJ) enclosed within a virtual 3D box (VBO) and, superimposed to the real object (OBJ), a virtual structure (VST) made of a set of planar tiles (TIL), and being anchored to the virtual 3D box (VBO), each tile (TIL) corresponding to a predetermined pose of the camera (CAM); c) detecting that a tile (TIL) is pointed at with the camera (CAM); d) acquiring, from the camera (CAM), a frame of the virtual 3D box (VBO), thereby validating said tile (TIL), said frame being a projection of the virtual 3D box (VBO) on the image (IM); steps a) to d) being iterated for different 3D positions of the camera (CAM), until a sufficient number of tiles is validated for scanning the real object (OBJ); e) implementing a 3D reconstruction algorithm with all the captured frames.