Single-Pass 3D Object Scanning With Motion-Defect Frame Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for generating 3D models of physical environments using mobile devices are inaccurate and inefficient, particularly in real-time environments, and fail to provide sufficient object detection.
Innovation Solution
A method involving a mobile device that selects a subset of images and depth data for 3D model generation, excluding motion-based defects, using a user interface for feedback and a meshing algorithm to create a 3D model based on keyframes, ensuring accurate and efficient object detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all captured images are used to generate 3D model, then more complete object coverage is achieved, but motion-based defects reduce model accuracy
Solution Approach 1:
The system extracts and removes images containing motion-based defects from the captured image set. By identifying and excluding defective images through quality assessment, the system prevents motion-blurred or distorted images from degrading the 3D model reconstruction accuracy, thus resolving the contradiction between using all images for completeness and filtering out defective ones for quality.
Solution Approach 2:
The system applies different quality standards and selection criteria to different regions or portions of the captured images. By assessing image quality locally and selecting only high-quality portions for model generation, the system maintains high accuracy in critical areas while still utilizing sufficient image coverage, thereby balancing model accuracy with image data quality.
2Productivity
If a subset of images is selected for 3D model generation, then computational efficiency is improved, but model completeness may be reduced
Solution Approach 1:
The system performs preliminary assessment and selection of images before the actual 3D model generation process. By pre-evaluating image quality, identifying keyframes, and selecting the optimal subset of images in advance, the system ensures that only high-quality, relevant images are processed during model generation. This preliminary action maintains model completeness while significantly improving computational efficiency by avoiding processing of defective or redundant images.
Solution Approach 2:
The system changes parameters such as image selection criteria, keyframe identification thresholds, and quality assessment metrics to optimize the balance between model completeness and processing speed. By adjusting these parameters based on the specific scanning scenario and object characteristics, the system achieves efficient generation of complete and accurate 3D models.
3Measurement precision
If high-resolution depth data is captured throughout the environment, then measurement precision is improved, but data processing complexity increases
Solution Approach 1:
The system segments the captured environment into distinct regions, objects, or zones and processes depth data separately for each segment. By dividing the large-scale depth data into manageable segments, the system maintains high measurement precision for each local region while reducing the overall processing complexity through localized analysis and reconstruction, avoiding the need to process all depth data simultaneously.
Data Source
AI summary
Various implementations disclosed herein include devices, systems, and methods that generates a three-dimensional (3D) model based on a selected subset of the images and depth data corresponding to each of the images of the subset. For example, an example process may include acquiring sensor data during movement of the device in a physical environment including an object, the sensor data including images of a physical environment captured via a camera on the device, selecting a subset of the images based on assessing the images with respect to motion-based defects based on device motion and depth data, and generating a 3D model of the object based on the selected subset of the images and depth data corresponding to each of the images of the selected subset.