Endoscopic Image Processor Scale Correction via Insertion Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Endoscopic inspection methods, particularly those using visual simultaneous localization and mapping (SLAM), face challenges in accurately reconstructing the three-dimensional shape of objects within industrial settings due to difficulties in measuring displacement and scale information, especially when the insertion portion deflects during operations.
Innovation Solution
An endoscopic image processing apparatus and system that includes a processor to estimate the self-position of the image pickup device, calculate displacement amounts based on detection signals from an insertion/removal state detection device, and generate scale information to create accurate three-dimensional shape models of objects by associating these displacement amounts, thereby improving the accuracy of object inspection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If visual SLAM method is used to reconstruct three-dimensional shape, then physical components for distance measurement can be eliminated, but measurement precision of displacement and scale information deteriorates
Solution Approach 1:
The patent introduces an insertion/removal state detection device as an intermediary component that detects the insertion/removal state of the insertion portion. This detection device provides reference information about the actual displacement of the image pickup device, which is then used to correct scale information calculated from visual SLAM. This intermediary detection mechanism bridges the gap between the component-less visual SLAM approach and the need for accurate measurement, allowing the system to maintain simplicity while improving measurement precision through external reference data.
2Ease of operation
If insertion portion deflects during insertion or removal, then ease of operation is improved, but measurement precision of displacement amount deteriorates
Solution Approach 1:
The patent implements a feedback mechanism where the insertion/removal state detection device continuously monitors the insertion/removal state of the insertion portion and provides detection signals to the processor. This feedback loop allows the system to detect when deflection occurs and use the detection signals to correct or compensate for measurement errors in the visual SLAM data. The feedback ensures that even when the insertion portion deflects for ease of operation, the system can identify and correct the resulting measurement inaccuracies.
3Device complexity
If visual SLAM is used for three-dimensional reconstruction, then device complexity is reduced, but reliability of scale information deteriorates
Solution Approach 1:
The patent merges two different measurement approaches: visual SLAM (which provides three-dimensional reconstruction without physical measurement components) and insertion/removal state detection (which provides reliable reference displacement information). By combining these two methods, the system maintains the simplicity of component reduction while improving reliability. The detection signals from the insertion/removal state detection device are integrated with the visual SLAM data to generate corrected scale information, achieving both simplicity and reliability simultaneously.
Data Source
AI summary
An endoscopic image processing apparatus is configured to create a three-dimensional shape model of an object by performing processing on an endoscopic image group of an inside of the object, and includes a processor. The processor estimates a self-position of the image pickup device based on the endoscopic image group, calculates a first displacement amount corresponding to a displacement amount of the image pickup device based on an estimation result of the self-position of the image pickup device obtained by the estimation, calculates a second displacement amount corresponding to a displacement amount in a direction parallel to a longitudinal axis direction of the insertion portion, based on a detection signal outputted from an insertion/removal state detection device that detects an insertion/removal state of an insertion portion inserted into the object, and generates scale information in which the first displacement amount and the second displacement amount are associated with each other.