Aberrometer Noise Reduction via Pixel Perimeter Averaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Accurate detection of image spot centers in aberrometers is hindered by noise and ghost images, leading to inaccurate characterization of wavefront aberrations, which is crucial for corrective ophthalmic applications.
Innovation Solution
A method that involves calculating and subtracting the average intensity value from a subset of pixels within a perimeter around each pixel, excluding those with high intensity values associated with image spots, to isolate and remove noise, thereby facilitating more accurate centroid determination of image spots.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If high band-pass filtering or linear filters are applied to reduce noise, then noise is reduced, but edge distortion and alteration of image feature size and shape occur
Solution Approach 1:
The patent segments the image processing task by dividing pixels into two distinct groups: first plurality of pixels (high intensity, associated with image spots) and second plurality of pixels (lower intensity, representing background noise). This segmentation allows different processing operations to be applied to each group independently, reducing noise in the background without affecting the edges and features of the image spots.
Solution Approach 2:
The patent applies local quality by treating different regions of the image differently based on their intensity characteristics. High intensity regions (image spots) are preserved with their original properties, while low intensity regions (background noise) are processed separately through averaging operations. This localized approach eliminates the need for global filtering that would cause edge distortion.
2Measurement precision
If centroid calculation is performed on noisy image data, then spot center location is determined, but measurement precision is reduced due to noise interference
Solution Approach 1:
The patent extracts and removes the harmful noise component from the image data by identifying pixels with intensity values below a threshold (second plurality of pixels) and replacing them with averaged intensity values from neighboring pixels. This extraction of noise allows subsequent centroid calculation to be performed on cleaned data, improving measurement precision without the interference of noise-induced ghost images.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
Methods and apparatus for facilitating determination of centroids of image spots in an image containing an array of image spots generated by an aberrometer, the image comprising a first plurality of pixels each pixel having a corresponding intensity value, the method comprising calculating an average intensity value for a second plurality of pixels in a perimeter around a pixel, the average calculated using a subset of the second plurality exclusive of at least a portion of the pixels in the perimeter.