LiDAR Abnormality Detection Using Dynamic Point Group Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
LiDAR devices face challenges in accurately identifying abnormal parts of objects due to potential displacement during measurement, leading to wrongful detection of non-abnormal parts as abnormal.
Innovation Solution
A processing device with units for calculating differences between reference and inspection point groups, extracting dynamic point groups indicating changes, removing these groups from differential point groups, and identifying abnormal parts based on the cleaned differential point groups.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If simple difference calculation between inspection point group and reference point group is used, then detection process is simple and fast, but displacement-related parts are wrongly detected as abnormal parts
Solution Approach 1:
The processing device segments the differential point group into two distinct categories: dynamic point groups (representing displacement) and abnormal point groups (representing genuine abnormalities). This segmentation is achieved by comparing multiple reference point groups to identify dynamic regions, then separating these from the inspection differential results. The segmentation enables accurate identification of true abnormalities while excluding displacement-related false positives.
Solution Approach 2:
The system performs preliminary action by acquiring and processing multiple reference point groups before the actual inspection. By calculating differences between reference point groups in advance and identifying dynamic point groups beforehand, the system prepares a baseline for distinguishing displacement from true abnormalities during inspection, preventing wrongful detection.
2Reliability
If multiple reference point groups are acquired and processed, then accuracy of abnormal part identification is improved, but measurement time and processing complexity increase
Solution Approach 1:
The system applies partial action by selectively processing only the necessary reference point groups. Instead of uniformly processing all possible reference data, the system identifies dynamic point groups from reference comparisons and applies this information specifically to filter differential point groups. This partial processing approach maintains high accuracy while reducing unnecessary computational overhead and time consumption.
3Measurement precision
If dynamic point groups are extracted and removed from differential point groups, then false detection is reduced, but processing steps increase
Solution Approach 1:
The processing device segments the differential point group into dynamic point groups and abnormal point groups through systematic comparison and identification processes. By dividing the detection task into distinct segmentation steps - identifying dynamic regions from reference groups, then separating them from inspection differentials - the system achieves precise detection while organizing complexity into manageable functional units.
Data Source
AI summary
Provided is a processing device capable of accurately identifying an abnormal part. A processing device (1) according to the present disclosure includes a first difference calculation unit (11) that calculates a difference between a plurality of reference point groups, a dynamic point group extraction unit (15) that extracts a dynamic point group being a point group involving a change from the reference point groups on the basis of a calculation result in the first difference calculation unit (11), a second difference calculation unit (12) that calculates a difference between an inspection point group acquired after the reference point group and the reference point group and generates a differential point group, a point group removal unit (16) that removes a point group corresponding to the dynamic point group from the differential point group generated in the second difference calculation unit (12), and an abnormal part identification unit (17) that identifies an abnormal part of the object to be measured on the basis of a differential point group from which the point group corresponding to the dynamic point group is removed.