UAV 3D Scan Planning With Iterative Model Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing unmanned aerial vehicle (UAV) systems lack the capability for autonomous, efficient, and high-resolution scanning of complex three-dimensional (3D) scan targets, such as structures with concavities, irregular surfaces, and oblique features.
Innovation Solution
The implementation of a UAV system that can autonomously scan 3D scan targets by generating a lower-resolution 3D model through initial rough scanning, iteratively refining it in real-time, and dynamically updating the scan plan to achieve high-resolution 3D reconstruction while avoiding obstacles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional UAV scanning systems are used, then the system structure is simple, but the scanning resolution is low and the system cannot handle complex 3D targets
Solution Approach 1:
The patent segments the scanning process into multiple phases: initial rough scanning to generate a low-resolution 3D model, followed by iterative refinement passes that progressively improve resolution. The scan plan is divided into multiple scan lines and passes, with each pass targeting specific unscanned or low-resolution regions. This segmentation allows the system to achieve high measurement precision through multiple coordinated scanning operations rather than requiring a single complex high-resolution scan.
Solution Approach 2:
The system performs preliminary rough scanning to generate an initial low-resolution 3D model before conducting detailed high-resolution scanning. This preliminary action provides a framework that guides subsequent scanning passes, allowing the system to identify unscanned regions and prioritize areas requiring higher resolution. The initial model serves as a foundation for iterative refinement, enabling efficient allocation of scanning resources.
2Productivity
If manual operation is used, then the system is easy to control, but the scanning efficiency is low and comprehensive scanning of large targets is slow
Solution Approach 1:
The system implements continuous feedback loops where scan data is processed in real-time to update the 3D model and dynamically adjust the scan plan. The system monitors scanning progress, identifies unscanned or low-resolution regions, and automatically generates updated flight paths to address these areas. This feedback mechanism enables autonomous operation, reducing the need for manual intervention while maintaining high scanning efficiency and comprehensive coverage.
Solution Approach 2:
The scan plan is dynamically updated during flight based on real-time analysis of scan data and current UAV position. The system adapts the scanning trajectory on-the-fly to ensure complete coverage of the target, adjusting scan lines and passes according to the evolving 3D model. This dynamic approach allows comprehensive scanning of large targets without requiring pre-planned static trajectories, improving both efficiency and completeness.
3Loss of time
If rough scanning is performed first, then the initial 3D model is obtained quickly, but the model resolution is low and requires iterative refinement
Solution Approach 1:
The scanning process is segmented into an initial rapid rough scan phase that quickly generates a low-resolution 3D model, followed by multiple refinement passes that progressively improve model resolution. Each refinement pass targets specific regions or applies enhanced scanning patterns to accumulate higher resolution data. This time-resolved segmentation allows the system to deliver a usable model quickly while continuing to improve precision in the background.
Solution Approach 2:
The system maintains continuous useful action by initiating the rough scan immediately to provide an initial model for navigation and planning, while simultaneously executing iterative refinement passes that continuously improve model resolution. The refinement process runs concurrently with or immediately following the rough scan, ensuring that model generation is an ongoing improving process rather than a discrete two-step operation. This continuity eliminates idle time and ensures constant progress toward high-resolution modeling.
Data Source
AI summary
In some examples, one or more processors of an aerial vehicle determine an update to a three-dimensional (3D) model corresponding to a scan target to scan according to a scan plan. Based at least on the update to the 3D model, the one or more processors determine a set of one or more uncovered points of the 3D model that are not covered by the scan plan. Additionally, the one or more processors determine an updated scan plan based at least on determining one or more poses to include in the scan plan for scanning the one or more uncovered points of the 3D model. The one or more processors, control the aerial vehicle to scan the scan target with the one or more image sensors according to the updated scan plan.