On-Site Image Quality Evaluation Using Edge Scores
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Visual inspection of structures, particularly wind turbine generators, often results in capturing images of insufficient quality, leading to the need for costly and time-consuming re-inspection due to the high running costs and risks of insufficient data collection, especially in offshore environments.
Innovation Solution
A method for on-site evaluation of image quality involves receiving images, dividing them into sub-images, calculating edge scores, sorting into top and bottom sub-groups based on edge scores, and evaluating image quality using coordinates and regression angles to determine focus and centering, allowing immediate retakes of low-quality images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If visual inspection collects data by taking images of the structure, then inspection data is obtained, but the running cost increases and there is a risk of insufficient data quality requiring repeat inspection
Solution Approach 1:
The patent applies preliminary action by evaluating image quality immediately after capture using automated algorithms (edge detection, focus assessment, composition analysis) before the inspection process completes. This allows immediate identification of insufficient quality images, enabling on-the-spot retakes rather than discovering data insufficiency after returning from inspection, thus preventing time loss from repeat trips.
Solution Approach 2:
The system implements feedback by providing real-time quality assessment results to the inspector during the inspection process. The automated evaluation metrics (edge sharpness, focus quality, composition scores) are fed back immediately, allowing the inspector to adjust capture parameters or retake images on-site, ensuring data quality requirements are met without requiring post-inspection review and repeat trips.
2Reliability
If visual inspection collects data by taking images of the structure, then inspection data is obtained, but the running cost increases
Solution Approach 1:
The patent reduces operational costs by performing preliminary quality evaluation immediately after image capture using automated computer vision algorithms. This prevents the waste of resources (fuel, time, personnel) associated with returning to the structure for repeat inspections, thereby reducing energy consumption and operational expenses while ensuring data quality.
Solution Approach 2:
The system applies self-service by using automated algorithms to evaluate image quality without requiring manual review of each image. The computer vision system independently assesses edge sharpness, focus quality, and composition metrics, enabling the inspection process to self-correct by triggering retakes only when necessary, thus minimizing unnecessary operational costs.
3Measurement precision
If the image is out of focus, then the part of the structure cannot be properly inspected, but retaking images increases time and cost
Solution Approach 1:
The patent implements feedback by providing immediate focus quality assessment through automated algorithms (analyzing edge sharpness, gradient magnitude, and frequency domain characteristics). This real-time feedback enables the inspector to adjust focus settings or retake images on-site, ensuring measurement precision requirements are met without delaying the inspection workflow, thus maintaining productivity.
Solution Approach 2:
The system replaces manual focus assessment with automated computer vision algorithms that objectively measure focus quality through image processing techniques (edge detection, gradient analysis, frequency domain transformation). This substitution eliminates subjective judgment and enables rapid, consistent evaluation of focus quality, maintaining inspection efficiency while ensuring measurement precision.
4Reliability
If thousands of images are taken to properly inspect each wind turbine generator, then complete inspection coverage is achieved, but the data processing burden and time consumption increase
Solution Approach 1:
The patent applies preliminary action by filtering and evaluating images immediately after capture using automated quality metrics. Images that fail to meet quality thresholds (poor focus, insufficient edge sharpness, bad composition) are identified and flagged for retake before leaving the inspection site. This preliminary filtering reduces the total number of images requiring detailed processing and analysis, thereby reducing data processing complexity while maintaining inspection coverage.
Solution Approach 2:
The system extracts and evaluates specific quality metrics (edge sharpness, focus quality, composition scores) from each captured image using computer vision algorithms. By extracting these key quality indicators and using them to filter images, the system separates high-quality images suitable for analysis from low-quality images requiring retake, thereby reducing the data processing burden while ensuring complete inspection coverage of relevant features.
Data Source
AI summary
A method for on-site evaluating quality of an image of a part of a structure, optionally a part of a wind turbine generator, the method comprising acts of: —receiving an image from a visual inspection system with a field of view about a line of sight towards the part of the structure; —dividing the image into sub-images; —calculating edge scores of the sub-images; —sorting the sub-images into a top sub-group having an edge-score above a pre-set edge score and a bottom sub-group having an edge-score below the pre-set edge score; —evaluating the quality of the image as a function of coordinates of the top sub-group.