Dense 3D Mapping via Multi-Sensor Fusion and vSLAM
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current three-dimensional mapping technologies face challenges in creating fully automatic and real-time high-quality maps due to scale factor cumulative drift, loop closure problems, and sensitivity to illumination and weather conditions, especially in outdoor environments.
Innovation Solution
A system that combines successive scans from a range sensor and images from a camera to generate a dense three-dimensional map, using multi-view geometry and vSLAM to refine depth estimates and account for translational and rotational movements, thereby improving accuracy and robustness across varying environmental conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If vision-based vSLAM solutions are used for three-dimensional reconstruction, then spatial resolution and cost are improved, but scale factor cumulative drift and loop closure problems occur leading to inaccurate results
Solution Approach 1:
The patent combines multiple sensing modalities (camera images, lidar point clouds, and radar depth measurements) into a unified mapping system. By merging vision-based high-resolution imaging with lidar's accurate geometric data and radar's robust depth measurements, the system achieves both high spatial resolution and reliability without suffering from cumulative drift or loop closure problems
2Reliability
If MMW radar is used for three-dimensional mapping, then reliability independent of illumination and weather conditions is improved, but depth output becomes very sparse and elevation/shape/size recognition fails
Solution Approach 1:
The patent integrates millimeter-wave radar with lidar and camera systems. The radar provides reliable all-weather depth information that complements the high-resolution but illumination-sensitive lidar and camera data. This fusion allows the system to maintain accurate three-dimensional mapping under varying weather and lighting conditions while preserving fine geometric details
3Measurement precision
If lidar is used for three-dimensional reconstruction, then accurate three-dimensional points are provided, but processing algorithms become memory- and time-consuming and data alignment is heavy
Solution Approach 1:
The patent fuses lidar point cloud data with radar depth measurements and camera images. The radar provides additional depth constraints that help align and register lidar data more efficiently, reducing the computational burden of processing large lidar datasets. The multi-sensor fusion approach enables accurate three-dimensional reconstruction with reduced processing complexity through complementary information from different sensing modalities
4Measurement precision
If point cloud-based methods are used for scene reconstruction, then accurate three-dimensional points are obtained, but unstructured representation prevents direct representation as connected surfaces
Solution Approach 1:
The patent combines lidar point cloud data with radar depth information and camera imagery to create a unified three-dimensional representation. The multi-sensor fusion provides both accurate geometric points and additional contextual information that enables the construction of connected surface models from the point cloud data, bridging the gap between unstructured point representations and structured surface models
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
The system generates a dense, accurate three-dimensional map that can be continually updated, enhancing applications such as autonomous navigation and obstacle detection by leveraging the strengths of both range sensors and cameras.
Implementation Method 1
a range sensor configured to receive signals reflected from objects in an environment and generate two or more successive scans of the environment at different times
Data Source
AI summary
In some examples, a system includes a range sensor configured to receive signals reflected from objects in an environment and generate two or more successive scans of the environment at different times. The system also includes a camera configured to capture two or more successive camera images of the environment, wherein each of the two or more successive camera images of the environment is captured by the camera at a different location within the environment. The system further includes processing circuitry configured to generate a three-dimensional map of the environment based on the two or more successive scans and the two or more successive camera images.