2-D 3-D Data Fusion via Morphology Extraction for UAV Geolocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for geolocation of unmanned aerial vehicles (UAVs) using 2-D and 3-D data fusion are complex and prone to errors, particularly in aligning optical and LiDAR data for accurate surface reconstruction and feature extraction.
Innovation Solution
A system and method that register 2-D and 3-D images by scanning a geographical region with a LIDAR system to generate a 3-D point cloud, and a camera to capture 2-D images, using morphology and shape extraction techniques to find matching points, estimate camera pose, and determine geo-location, enabling fusion of image pixels and LIDAR points for accurate geolocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 2-D and 3-D data fusion is used for geolocation of UAVs, then geolocation capability is improved, but system complexity increases and error potential arises
Solution Approach 1:
The patent segments the complex 2-D and 3-D data fusion process into distinct modular components: feature extraction module, feature matching module, and pose estimation module. This segmentation allows each module to handle specific tasks independently, reducing overall system complexity while maintaining geolocation accuracy through specialized processing at each stage.
Solution Approach 2:
The patent introduces key point features as an intermediary element between 2-D aerial images and 3-D LiDAR point clouds. These extracted features serve as a common language for matching between different data types, simplifying the fusion process by providing explicit correspondence points rather than attempting direct pixel-to-point mapping.
2Measurement precision
If feature extraction and matching is performed between 2-D aerial images and 3-D LiDAR point clouds, then correspondence accuracy is improved, but processing complexity increases
Solution Approach 1:
The patent extracts only the most salient geometric features (corners, edges, keypoints) from the complete 2-D aerial images and 3-D LiDAR point clouds. By taking out only these critical features rather than processing all data points, the system achieves accurate correspondence matching while significantly reducing processing complexity and computational load.
Solution Approach 2:
The patent applies different processing strategies to different regions and feature types within the data. For example, corner features are extracted and matched with specific geometric primitives in the point cloud, while edge features are handled differently. This local quality approach optimizes matching accuracy for each feature type without uniformly increasing processing complexity across all data.
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 provides accurate camera pose estimation and geo-location of UAVs, even without GPS, by matching 2-D and 3-D image features, enhancing object detection and surface reconstruction with improved accuracy and robustness.
Implementation Method 1
scanning the geographical region by a LIDAR system attached to a LIDAR drone to produce a 3-dimensional point cloud
Data Source
Figure 1
Figure 2
Figure 3
AI summary
According to an embodiment, a 2-dimensional (2-D) image of a geographical region is transformed via a regional maxima transform (RMT) or an edge segmenting and boundary filling (ESBF) transform to produce a filtered 2-D image. The filtered 2-D image is iteratively eroded and opened to produce a processed EO 2-D image, 2-D object shape morphology is extracted from the processed EO 2-D image, and 2-D shape properties are extracted from the 2-D object shape morphology. A height slice of a 3-dimensional (3-D) point cloud comprising 3-D coordinate and intensity measurements of the geographical region is generated, and slice object shape morphology is extracted from the height slice. Slice shape properties from the slice object shape morphology are extracted, and the 2-D image is constellation matched to the height slice based on the 2-D shape properties and the slice shape properties.