Road Surface 3D Modeling from Multi-Vehicle Image Parallax
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies struggle to generate highly accurate three-dimensional models of small objects like road surfaces or structures on roads, as they rely on imaging data that is insufficient for detailed analysis.
Innovation Solution
A three-dimensionalization system that acquires images from multiple moving bodies, selects suitable images based on predetermined conditions, and generates a three-dimensional model using these images, employing methods like parallax calculation and image matching to create precise depth maps of road surfaces and structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple images from multiple moving bodies are acquired and processed, then the accuracy of three-dimensional model generation is improved, but the complexity of the system increases
Solution Approach 1:
The system segments the image processing task by having each moving body independently capture images and identify features, then combining results from multiple sources. This divides the complex task of three-dimensional reconstruction into manageable independent units that can be processed separately and integrated.
Solution Approach 2:
The system uses universal image processing algorithms that can handle images from any moving body regardless of position or orientation. The feature identification and matching methods are designed to work with images captured from multiple different viewpoints, making the system adaptable to various imaging scenarios without requiring specialized processing for each case.
2Reliability
If images are captured by multiple moving bodies to improve measurement accuracy, then the reliability of the three-dimensional model increases, but the difficulty of detecting and measuring features increases
Solution Approach 1:
The system introduces feature points as intermediary elements that mediate between the multiple image sources and the final three-dimensional model. By identifying and matching specific feature points across images from different moving bodies, the system creates a reliable correspondence relationship that simplifies the integration process and improves measurement reliability.
Solution Approach 2:
The system transitions from two-dimensional image data to three-dimensional spatial information by utilizing parallax effects. Images captured from different positions and angles provide additional dimensional information that enables accurate three-dimensional reconstruction, transforming the problem from planar feature detection to spatial feature matching.
3Manufacturing precision
If a large number of images are processed to generate accurate three-dimensional models of small objects, then the manufacturing precision improves, but the loss of time increases
Solution Approach 1:
The system processes only the necessary subset of images and features required for accurate three-dimensional reconstruction rather than exhaustively processing all available data. By selectively identifying and matching key feature points, the system achieves high precision while reducing processing time through optimized selective action.
Solution Approach 2:
The system performs preliminary feature identification and image filtering before the main three-dimensional reconstruction process. By pre-processing images to identify salient features and establish correspondence relationships in advance, the system reduces the computational burden during the actual model generation phase, thereby reducing overall processing time while maintaining precision.
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
Enables the generation of highly accurate three-dimensional models of road surfaces and structures, supporting effective repair planning while reducing costs by using monocular cameras instead of expensive stereo systems.
Implementation Method 1
an imaging region of an existing three-dimensional model is specified using supplementary information regarding imaging included in imaging data
Implementation Method 2
employing methods like parallax calculation and image matching to create precise depth maps of road surfaces and structures
Data Source
AI summary
A three-dimensionalization system according to an aspect of the present disclosure includes: at least one memory storing instructions; and at least one processor configured to execute the instructions to acquire a plurality of images captured by imaging devices each installed on each of a plurality of moving bodies, select at least two images obtained by capturing a road surface of a road or a structure on the road from among the plurality of acquired images based on a predetermined condition, and generate a three-dimensional model of the captured road surface of the road or the captured structure on the road using the selected images.