Transparent Object 3D Reconstruction via First-Reflection Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional 3D reconstruction technologies face challenges in reconstructing transparent objects due to complex laser transmission scenarios within these objects.
Innovation Solution
The method involves using a system with a structure light generation module, image acquisition module, control module, and computing module. The system emits a laser that scans across the object's surface, and the image acquisition module captures feedback image pairs. The computing module processes these images to calculate three-dimensional positions and performs a refinement process to extract first-reflection points, specifically removing secondary points and retrieving missing points based on camera views.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If laser scanning is used to reconstruct transparent objects, then three-dimensional position data can be collected, but complex laser transmission scenarios cause measurement errors and ambiguity points
Solution Approach 1:
The patent segments the reflection points into two categories: first reflection points (direct reflection from object surface) and second reflection points (refraction through transparent object). By separating these types and applying different handling rules to each, the system resolves the measurement ambiguity caused by laser transmission through transparent materials.
Solution Approach 2:
The patent extracts and removes second reflection points (caused by laser refraction through transparent objects) from the measurement data, keeping only first reflection points for reconstruction. This extraction eliminates the harmful effect of refraction-induced measurement errors while preserving accurate surface geometry data.
2Quantity of substance
If multiple cameras are used to capture reflection points, then more data points can be collected, but ambiguity points increase due to multiple viewing angles
Solution Approach 1:
The patent uses feedback from multiple cameras to identify and eliminate ambiguity points. By comparing reflection point data across different camera views, the system determines whether a point is a valid first reflection point or a spurious second reflection point, thereby improving overall measurement reliability.
Solution Approach 2:
The patent converts the harmful effect of multiple viewing angles (which create ambiguity points) into a beneficial filtering mechanism. Ambiguity points that would normally corrupt the data actually serve as indicators that help the system identify and remove incorrect measurements, thereby improving reconstruction accuracy.
3Ease of manufacture
If traditional 3D reconstruction methods are used on transparent objects, then the process is simple, but reconstruction accuracy deteriorates due to laser transmission effects
Solution Approach 1:
The patent performs preliminary classification of reflection points into first and second types before the actual 3D reconstruction process. This preliminary action identifies and isolates problematic refraction points, ensuring they do not contaminate the final reconstruction model, thereby maintaining both process simplicity and high accuracy.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach effectively reconstructs the exterior surface of transparent objects by accurately extracting first-reflection points, overcoming the challenges of laser transmission and ambiguity points within the objects.
Implementation Method 1
collecting reflection points from the surface
Implementation Method 2
complex situations of laser transmission inside the transparent object
Data Source
AI summary
The present application relates to methods, devices and systems for transparent object three-dimensional reconstruction. The system comprises a structure light generation module, an image acquisition module, a control module and a computing module. The computing module acquires image pairs from the image acquisition module; calculates three-dimensional positions of points according to the image pairs; and performs refinement process to extract first-reflection points. The refinement process comprising: when reflection points are obtained by a first camera, the points include a first point and a second point, the first point is closer to reflection spot of laser on galvanometer mirror than the second point, remove the second point; when the first point is not obtained by a second camera, remove the first point; when the second point is obtained by the second camera, retrieve the second point, and when a discrete external virtual contour is formed, remove the discrete external virtual contour.