Absolute Camera Pose Alignment Using Local 3D Scene Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vision-based navigation systems face accuracy issues due to variations in scene observation conditions and seasonal changes, leading to significant errors in determining the position of a camera-equipped aircraft or spacecraft.
Innovation Solution
A method using a monocular and passive camera on a moving aircraft or spacecraft to determine an absolute pose by constructing a local 3D model from a sequence of images, aligning it with a reference 3D model to account for scene geometry independent of observation conditions and seasonal variations, utilizing multi-view 3D reconstruction and realignment techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a single reference image is used for comparison, then the method is simple and fast, but accuracy deteriorates due to variations in scene observation conditions and seasonal changes
Solution Approach 1:
The patent divides the reference image into multiple sub-images or feature regions, allowing the system to compare the captured image with multiple segmented reference portions. This segmentation enables the system to handle variations in observation conditions and seasonal changes by matching features across different reference segments, thereby maintaining position determination accuracy while keeping the overall process efficient.
Solution Approach 2:
The patent performs preliminary processing by pre-segmenting the reference image and pre-identifying characteristic features or landmarks before actual position determination. This preliminary action allows the system to be better prepared for various lighting conditions, cloud cover, and seasonal variations, enabling more accurate matching when the captured image is compared against the prepared reference segments.
2Device complexity
If a single reference image is used, then the system complexity is low, but reliability deteriorates under varying illumination and seasonal conditions
Solution Approach 1:
The reference image is segmented into multiple portions or feature regions, each representing different areas or characteristics of the scene. This segmentation increases reliability by providing multiple reference points for comparison, making the system less susceptible to failures caused by poor lighting, clouds, or seasonal changes in any single reference area.
Solution Approach 2:
The patent creates a multi-functional reference system that can handle various observation conditions (different lighting, seasons, weather) through its segmented structure. The same reference image framework serves multiple purposes by providing different reference portions for different environmental conditions, thereby enhancing reliability without proportionally increasing system complexity.
3Measurement precision
If dense matching of all pixels is performed, then measurement precision improves, but processing time increases
Solution Approach 1:
The patent applies local quality by focusing the dense matching process on specific regions or features of interest rather than uniformly processing the entire image. By identifying and prioritizing characteristic features, landmarks, or areas with distinctive patterns, the system achieves high measurement precision for the local 3D model while reducing overall processing time by avoiding unnecessary pixel-level analysis in less informative areas.
Solution Approach 2:
The system performs preliminary feature detection and selection before executing the dense matching process. By pre-identifying characteristic features and establishing a preliminary feature map, the system can then focus computational resources on matching only these pre-selected features with their corresponding reference portions, significantly reducing processing time while maintaining or improving measurement precision through targeted detailed matching.
Data Source
AI summary
A method (20) to determine an absolute pose of a camera (11) located on a craft (10) that is able to move relative to a scene, the method includes: obtaining (S20) a sequence of images of a scene captured by the camera, from the sequence of images generating a local 3D model in a coordinate system of the camera, the local 3D model representing a portion of the scene at a target image among the sequence of images, determining (S22) the absolute pose of the camera at the target image by realigning the position and attitude of the local 3D model with a predetermined reference 3D model corresponding to the scene represented in three dimensions in the reference coordinate system.