Image Processing Device for Accurate Metal Piece Extraction in Radiation Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing methods inaccurately assign metal pieces in radiation images, leading to erroneous recognition and adverse impact on color tone correction and tomographic image generation, especially when distinguishing between metal pieces and surrounding materials like cement.
Innovation Solution
An image processing device that employs binarization, edge extraction, image synthesis, profile trimming, and graph cut processing to accurately discriminate metal pieces from other regions, using median and Laplacian filters to enhance precision and exclude metal pieces from color tone correction, thereby improving visual recognition and preventing false images in tomographic images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If color tone correction is executed on the entire image including metal piece regions, then processing speed is improved, but measurement precision of metal piece regions deteriorates due to erroneous recognition
Solution Approach 1:
The image is segmented into metal piece regions and non-metal piece regions using graph cut processing. This segmentation allows different processing strategies to be applied to different regions, ensuring accurate metal piece identification while maintaining efficient overall processing.
Solution Approach 2:
Metal piece regions are extracted from the image using binarization and graph cut processing. By separating metal piece regions, the system can exclude them from color tone correction processing, preventing erroneous recognition while maintaining processing efficiency for non-metal regions.
2Device complexity
If binarization processing is used to extract metal pieces, then device complexity is reduced, but measurement precision of metal piece boundaries deteriorates due to inaccurate threshold selection
Solution Approach 1:
Graph cut processing is introduced as an intermediary step between binarization and final metal piece identification. The graph cut algorithm uses the binarization result as initial input but refines the metal piece region extraction by considering spatial relationships and intensity gradients, thereby improving boundary accuracy without significantly increasing system complexity.
Solution Approach 2:
Binarization processing is performed as a preliminary step to create an initial segmentation of the image. This preliminary action provides a starting point for the more accurate graph cut processing, allowing the system to benefit from both the simplicity of binarization and the precision of graph cut methods.
3Measurement precision
If color tone correction excludes metal piece regions, then measurement precision of non-metal regions is improved, but productivity decreases due to additional processing steps
Solution Approach 1:
Metal piece regions are extracted and excluded from color tone correction processing. This extraction allows the color tone correction to be applied only to non-metal regions, improving visual recognition of these areas while minimizing the impact on processing speed by limiting the correction scope.
Solution Approach 2:
Different processing qualities are applied to different regions: metal piece regions are precisely identified and excluded, while non-metal regions receive color tone correction to improve visual recognition. This local differentiation optimizes both precision and efficiency for each region type.
Data Source
AI summary
According to the image processing device of the present invention, the binarization image having increasing assuredness can be generated by extracting the metal piece from the original image with the graph cut processing. The image processing device of the present invention is the system that executes an image trimming from near the center of the intermediate region after the metal piece is divided relative to the image of the roughly extracted binarization image near the center of the intermediate region in that it is difficult to decide whether it belongs to the metal piece or not. Following such steps, the intermediate region can be assuredly trimmed while executing the image trimming in the region as small as possible.