Tomographic Image Segmentation for Automated Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current tomographic analysis methods for mechanical parts require significant operator intervention to detect anomalies, leading to delays and potential oversight due to the reliance on human assessment of three-dimensional images.
Innovation Solution
A method involving tomographic analysis that automatically subdivides the part into sub-parts, analyzes grey level distributions, compares parameters with standard ranges, and identifies risk regions to detect anomalies without operator intervention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If manual assessment of three-dimensional images is used to detect anomalies, then detection reliability is maintained, but analysis time and operator workload increase significantly
Solution Approach 1:
The patent introduces an automated analysis system that acts as an intermediary between the tomographic imaging device and the human operator. This system processes three-dimensional images automatically, extracting features and detecting anomalies without requiring manual assessment, thereby reducing analysis time while maintaining detection reliability through algorithmic precision
Solution Approach 2:
The patent replaces the manual mechanical assessment process with an automated computational system. Instead of human operators manually examining three-dimensional images, the system uses automated image processing algorithms to detect anomalies, substituting human labor with machine-based analysis that is both faster and equally reliable
2Productivity
If automated analysis is implemented to reduce operator time, then productivity increases, but risk of missing subtle anomalies may increase
Solution Approach 1:
The patent implements feedback mechanisms where the automated analysis system continuously refines its detection algorithms based on identified anomalies and comparison with known defect patterns. This feedback loop ensures that the system maintains high detection accuracy by learning from each analysis and adjusting its parameters to minimize false negatives
Solution Approach 2:
The patent performs preliminary actions by pre-processing three-dimensional images and pre-identifying potential anomaly regions before final detection. This preliminary analysis prepares the data in advance, allowing the automated system to efficiently scan and detect anomalies with high accuracy without compromising thoroughness
3Reliability
If comprehensive inspection of entire part volume is performed, then detection completeness is ensured, but analysis complexity and time increase
Solution Approach 1:
The patent segments the three-dimensional image into multiple sections or regions of interest, analyzing each segment separately rather than processing the entire volume as a single unit. This segmentation reduces analysis complexity by breaking down the large dataset into manageable portions while maintaining comprehensive coverage through systematic processing of all segments
Solution Approach 2:
The patent extracts and isolates specific features or regions that are most likely to contain anomalies, removing unnecessary data from the analysis process. By taking out only the relevant portions for detailed examination, the system reduces overall analysis complexity while ensuring that potential anomalies are not missed
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
Enables rapid and reliable detection of anomalies with reduced operator time and minimized risk of missing defects, optimizing the analysis process.
Implementation Method 1
at least one emitting device 3, configured to emit an incident beam 5 of wave pulses towards the part 20, for example radio frequency waves, X-rays or acoustic waves
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
A method for the tomographic analysis a part in order to detect anomalies (30), the method comprising the following steps: - acquiring at least one three-dimensional image (25) of the part by means of a tomography device; - subdividing the image (25) into elementary subparts (27); - analyzing a grayscale distribution in each subpart (27) and obtaining at least one parameter representative of this grayscale distribution for each subpart (27); - comparing the one or more parameters obtained for each subpart (27) with standard values characteristic of a defect-free region and detecting abnormal subparts (31) for which the one or more parameters differ from the standard values; - determining risk regions, which comprise each abnormal subpart (31) and each subpart (33) adjacent to at least one abnormal subpart; and - analyzing the risk regions to detect the anomalies (30) in the part.