Image Distortion Correction via Segmented Geometric Transformation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image processing methods struggle to achieve a balance between line reproducibility and magnification distortion reduction, often resulting in noticeable distortions or inadequate representation of image areas.
Innovation Solution
An image processing method that employs a geometric transformation characteristic switching between tan θ for central areas and tan (θ/2)^(κp) for peripheral areas, with θp as a threshold, to maintain line reproducibility and minimize magnification distortion, using equations to calculate new pixel coordinates and perform interpolation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If geometric transformation processing uses tan θ characteristic for high line reproducibility, then line reproducibility is improved, but magnification distortion increases in peripheral areas
Solution Approach 1:
The image is divided into a central area and a peripheral area based on a threshold subject angle θp. Different geometric transformation characteristics are applied to each area: tan θ is used for the central area to maintain line reproducibility, while tan (θ/2)κp is used for the peripheral area to reduce magnification distortion. This segmentation allows each region to be optimized independently according to its specific requirements.
Solution Approach 2:
Different geometric transformation characteristics are assigned to different regions of the image based on their local requirements. The central area, where line reproducibility is critical, receives the tan θ transformation. The peripheral area, where magnification distortion is more problematic, receives the tan (θ/2)κp transformation. This local quality approach ensures that each region's specific needs are addressed without compromising the other.
2Manufacturing precision
If geometric transformation processing uses tan (θ/2)κp characteristic for magnification distortion reduction, then magnification distortion is reduced, but line reproducibility deteriorates
Solution Approach 1:
The image processing is segmented into two distinct transformation zones. The peripheral area, identified by subject angles greater than θp, applies the tan (θ/2)κp transformation to minimize magnification distortion. The central area, with subject angles less than or equal to θp, applies the tan θ transformation to preserve line reproducibility. This segmentation ensures that the trade-off is optimized for each region rather than applying a single transformation globally.
Solution Approach 2:
The geometric transformation characteristic is made local rather than global. Each area of the image receives a transformation characteristic suited to its specific requirements. The peripheral area gets the distortion-reducing tan (θ/2)κp transformation, while the central area gets the line-preserving tan θ transformation. This local quality approach resolves the contradiction by allowing different qualities in different locations.
3Device complexity
If a single geometric transformation characteristic is used for the entire image, then processing is simplified, but noticeable distortion occurs in certain areas
Solution Approach 1:
Rather than using a single geometric transformation for the entire image, the processing is segmented into two areas with different transformation characteristics. The threshold subject angle θp serves as the segmentation criterion. This segmentation increases processing complexity slightly but dramatically improves image quality by preventing noticeable distortion in both central and peripheral areas.
Solution Approach 2:
The geometric transformation parameter is changed based on the subject angle θ. When θ ≤ θp, the transformation follows tan θ; when θ > θp, the transformation follows tan (θ/2)κp. This parameter change approach allows the system to adapt the transformation characteristic to the specific region being processed, improving overall image quality while maintaining a relatively simple processing framework.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
A proposition of the invention is to provide an image processing method, an image processing program, an image processing apparatus, and an imaging apparatus capable of attaining line reproducibility and magnification distortion reduction in a well-balanced manner in an image. The image processing method according to the invention is an image processing method for performing predetermined geometric transformation processing (h(θ)) on an image to be processed, in which the predetermined geometric transformation processing (h(θ)) includes geometric transformation processing for magnification distortion reduction (h=αtan (θ/2)^(κp)) that reduces discrepancy between circumferential magnification and radial magnification of the image to be processed, and at least one parameter (θp) that determines the strength or content of the predetermined geometric transformation processing (h=αtan (θ/2)^(κp)) is set according to the structure of the image to be processed.