Autonomous Video Crack Detection for Nuclear Reactor Surfaces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automated crack detection methods for nuclear power plant components, particularly underwater reactors, are inefficient due to high temperatures and radiation hazards, leading to unreliable crack identification and high false positive rates, especially with small cracks and surface texture features like scratches and welds.
Innovation Solution
A system utilizing a video camera and light source to capture surface texture features, combined with machine learning classifiers like LBP/SVM and CNN architectures, to analyze video frames, track motion, and filter out non-crack features, enhancing crack detection reliability and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated crack detection algorithms based on edge detection, thresholding, or morphological operations are used, then the speed of inspection is improved, but the reliability of crack detection deteriorates due to high false positive rates from non-crack surface features
Solution Approach 1:
The inspection process is segmented into multiple independent stages: candidate feature detection, feature classification, and verification. This multi-stage segmentation allows each stage to focus on specific tasks, improving overall reliability while maintaining speed through parallel processing capabilities.
Solution Approach 2:
A machine learning classifier acts as an intermediary between raw image data and final crack detection results. This intermediary layer processes candidate features through learned patterns, filtering out false positives from non-crack surface features while preserving true crack detections.
2Reliability
If machine learning classifiers are used to improve crack detection accuracy, then the false positive rate from non-crack features is reduced, but the complexity of the detection system increases
Solution Approach 1:
The system segments complexity into manageable modules: preprocessing, candidate detection, machine learning classification, and verification. Each module has a specific function and can be independently optimized or replaced, making the overall complex system maintainable and understandable.
Solution Approach 2:
The system performs preliminary actions by pre-processing images and pre-identifying candidate features before applying the machine learning classifier. This preliminary filtering reduces the input complexity to the classifier, allowing it to focus on discrimination rather than basic feature extraction.
3Device complexity
If traditional automated detection methods are used, then the system simplicity is maintained, but the ability to detect small cracks with low contrast deteriorates
Solution Approach 1:
The system applies partial action by using simple preprocessing and candidate detection methods for most of the image processing, then applies excessive action through machine learning classification specifically for discriminating difficult-to-detect small cracks with low contrast, achieving high precision without overwhelming complexity.
Data Source
AI summary
Systems and methods suitable for capable of autonomous crack detection in surfaces by analyzing video of the surface. The systems and methods include the capability to produce a video of the surfaces, the capability to analyze individual frames of the video to obtain surface texture feature data for areas of the surfaces depicted in each of the individual frames, the capability to analyze the surface texture feature data to detect surface texture features in the areas of the surfaces depicted in each of the individual frames, the capability of tracking the motion of the detected surface texture features in the individual frames to produce tracking data, and the capability of using the tracking data to filter non-crack surface texture features from the detected surface texture features in the individual frames.