Mapping Close-up Images to 3D Structure Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing imaging methods using unmanned aerial vehicles (UAVs) struggle to accurately locate close-up images of structural damage within the overall structure, as these images often show only a small portion and vary in orientation, making it difficult for users to determine their location without additional information.
Innovation Solution
The method involves mapping close-up images to a 3D model of the structure, allowing the location of these images to be indicated on an overview image, using techniques such as structure from motion and triangulation to generate a 3D point cloud and associate the images with specific portions and positions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If close-up photos are taken to obtain high-resolution images, then detailed textures are provided, but the overall location of the images on the structure becomes difficult to identify
Solution Approach 1:
The patent projects 2D close-up images onto a 3D model of the structure, adding spatial context. By mapping images to specific locations and orientations on the 3D model, the system preserves both high-resolution details and location information simultaneously, resolving the contradiction between image quality and location identifiability
Solution Approach 2:
The 3D model serves as an intermediary between the close-up images and the overall structure. It provides a framework that connects detailed images to their global positions, allowing users to locate images on the structure without losing either resolution or location context
2Area of stationary object
If many close-up images are captured to cover the entire structure, then comprehensive coverage is achieved, but the time and computational effort to locate specific images increases
Solution Approach 1:
The system pre-processes images during capture by embedding location data, orientation information, and timestamps directly into each image file. This preliminary organization allows for rapid retrieval and location identification later, eliminating the need for manual comparison and significantly reducing time consumption
Solution Approach 2:
The patent creates a virtual copy of the structure as a 3D model that mirrors the physical structure's geometry and layout. This digital twin serves as a reference framework where all captured images can be systematically organized and located, enabling efficient navigation and retrieval without manually examining numerous images
3Adaptability or versatility
If images are taken from varying orientations to capture all structure details, then comprehensive viewing is achieved, but determining image orientation and location becomes more difficult
Solution Approach 1:
By mapping images from various orientations onto a 3D model, the system preserves orientation information in the spatial coordinates of the model. Each image's position, angle, and rotation are captured as part of its 3D location data, making orientation information readily available without increasing complexity
Solution Approach 2:
The system transforms orientation parameters (roll, pitch, yaw angles) into standardized 3D coordinate system parameters. This parameter transformation allows varying orientations to be systematically represented and compared within a unified reference frame, simplifying orientation determination
Data Source
AI summary
Certain aspects of the present disclosure relate to methods and apparatus for implementing image locating relative to a global structure. The method generally includes mapping one or more close-up images of a structure to a 3D model of the structure, and indicating, on an overview image of the structure, the location of the one or more close-up images based on the mapping.