Inspection Montage Alignment Using Non-Sequential VIO
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for capturing and organizing visual data in environments like manufacturing, construction, and retail fail to efficiently create montages that focus on relevant elements, leading to tedious and error-prone manual processes, especially in large projects with repetitive architectural details.
Innovation Solution
A system utilizing non-sequential visual inertial odometry (VIO) to estimate the positions of a capture apparatus during a walkthrough, allowing unordered capture data to be organized into a montage that aligns with a blueprint or floorplan, using a capture device with cameras and IMUs, and applying user markings and constraints to fit the estimated path onto the blueprint.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If manual organization of inspection photos is performed, then flexibility in handling is maintained, but productivity decreases and error rate increases
Solution Approach 1:
The system automatically organizes inspection photos by having the capture device self-determine its position and orientation through VIO, automatically match photos to blueprint locations, and generate organized montages without requiring manual intervention for each photo placement
Solution Approach 2:
The patent replaces manual mechanical organization processes with an automated computational system that uses visual inertial odometry and computer vision algorithms to automatically position and organize inspection photos according to their captured locations
2Productivity
If automated montaging system is implemented, then productivity increases, but device complexity increases
Solution Approach 1:
The capture device is designed as a multi-functional integrated system that combines camera capabilities, IMU sensors, VIO processing, and automatic montage generation in a single device, eliminating the need for separate manual operations at multiple stages
3Measurement precision
If visual inertial odometry is used for position estimation, then measurement precision improves, but device complexity increases
Solution Approach 1:
The system merges visual data from cameras with inertial data from IMU sensors to create a combined VIO system that leverages the complementary strengths of both sensing modalities for more accurate position and orientation estimation
4Adaptability or versatility
If unordered capture data is processed, then adaptability improves, but loss of information increases
Solution Approach 1:
The VIO system continuously provides feedback about the capture device's position and orientation based on visual and inertial sensor data, allowing the system to maintain accurate spatial context information even when photos are captured in non-sequential orders
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
Automates the process of creating montages that accurately reflect the walkthrough path on blueprints, reducing manual effort and errors, and enabling efficient organization and visualization of captured content for various applications.
Implementation Method 1
an inertial measurement unit (IMU)
Implementation Method 2
a capture apparatus containing a camera
Data Source
AI summary
Montaging techniques are disclosed that utilize non-sequential visual inertial odometry (VIO) performed on non-sequential/unordered capture data collected by a capture apparatus. The capture apparatus is carried by an observer/inspector at a site during a walkthrough/inspection. The capture apparatus comprises one or more cameras and an inertial measurement unit (IMU). User markings are applied to portions of the capture data. Based on the non-sequential VIO, a velocity profile and subsequently a set of positions of the capture apparatus are estimated as it was carried by the observer/inspector during the walkthrough/inspection. The above is accomplished via a constrained integration that utilizes constraints conditioning the motion of the capture apparatus. A montage of the capture data is produced that suits the needs of a given application of the instant montaging technology.