Abnormality Detection Using Point Cloud Luminance Differences
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for detecting abnormalities in a space using point cloud data representing three-dimensional positions and luminance struggle with accurately distinguishing between normal and abnormal parts, particularly due to issues like moving objects and parts prone to erroneous determination.
Innovation Solution
An abnormality detection apparatus that acquires reference and inspection point cloud data, generates difference data, detects moving objects and parts prone to errors as excluded parts, and uses this information to accurately identify abnormal parts by excluding them from the detection target.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Difficulty of detecting and measuring
If point cloud data is used to detect abnormalities in a space, then detection capability is provided, but detection accuracy deteriorates due to moving objects and parts prone to erroneous determination
Solution Approach 1:
The patent extracts and removes moving objects and parts prone to erroneous determination from the set of parts to be detected. By separating these problematic parts from the target object parts, the system can focus detection accuracy on stationary parts of the target object, thereby resolving the contradiction between providing detection capability and maintaining high detection accuracy.
Solution Approach 2:
The patent segments the detection process into multiple stages: first identifying moving objects and error-prone parts, then excluding them from the detection target, and finally detecting abnormalities only in the remaining parts. This segmentation allows the system to maintain high accuracy by processing only relevant parts while still providing comprehensive detection coverage.
2Difficulty of detecting and measuring
If all parts in the space are detected for abnormalities, then detection coverage is improved, but false positives increase due to moving objects
Solution Approach 1:
The patent extracts moving objects from the detection scope by detecting their motion characteristics and excluding them from the abnormality detection target. This allows the system to maintain broad detection coverage for all parts in the space while preventing false positives by removing moving objects that would otherwise be incorrectly identified as abnormalities.
Solution Approach 2:
The patent performs preliminary detection to identify moving objects and parts prone to erroneous determination before the main abnormality detection process. This preliminary action filters out problematic parts in advance, ensuring that subsequent abnormality detection operates only on stable, reliable parts, thereby reducing false positives while maintaining comprehensive coverage.
Data Source
AI summary
Abnormality detection apparatus acquires reference point cloud data and inspection point cloud data for each of a plurality of parts in a space including a target object. The reference point cloud data includes point data representing a three-dimensional position and luminance at reference time for each of a plurality of parts. The inspection point cloud data includes point data representing a three-dimensional position and luminance at time of inspection for each of the plurality of parts. The abnormality detection apparatus generates difference point cloud data representing a difference in luminance between the reference time and the time of inspection for each of the parts, detects an excluded part to be excluded from a detection target of an abnormal part, and detects an abnormal part of the target object from the parts other than the excluded parts by using the difference point cloud data.