Digital Image Noise Reduction via Selective Edge Sharpening
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Digital gray value images recorded with high-speed electron beams suffer from noise and reduced resolution due to mechanical and electronic disturbances, while slow scanning results in blurred images due to long integration times.
Innovation Solution
A method that generates a binary edge image to identify edges, applies a sharpness algorithm within edge regions using a tangent hyperbolic function, and smooths outside regions to produce a sharpened and smoothed image without global noise suppression filters, selectively filtering noise only in areas with small gradients.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If a noise suppression filter is applied to the entire image, then noise is reduced, but image resolution deteriorates
Solution Approach 1:
The patent applies different processing treatments to different regions of the image: edge regions undergo sharpening processing while non-edge regions undergo noise suppression processing. This local differentiation allows noise reduction in appropriate areas without degrading the resolution of edge regions, thereby resolving the contradiction between noise suppression and resolution preservation.
Solution Approach 2:
The image is segmented into edge regions and non-edge regions based on gradient calculation. By dividing the image processing into region-specific operations, the patent enables selective application of noise suppression and sharpening algorithms, avoiding the resolution deterioration that would result from global noise filtering.
2Productivity
If the electron beam is guided at high speed, then productivity is improved, but noise and granulation increase
Solution Approach 1:
The patent converts the harmful noise and granulation effects into useful information by detecting gray value gradients. Regions with high gradients are identified as edges and subjected to sharpening, while regions with low gradients are treated as noise and subjected to suppression. This transforms the noise problem into a basis for selective image enhancement.
Solution Approach 2:
The patent changes the processing parameters dynamically based on local image characteristics. By calculating gradient magnitude at each pixel, the system adapts the processing intensity and type (sharpening vs. noise suppression) to local conditions, enabling effective noise reduction without compromising overall image quality even when acquired at high scanning speeds.
3Object-affected harmful factors
If the electron beam is guided at slow speed, then noise is reduced, but image elements become blurred
Solution Approach 1:
The patent applies sharpening processing specifically to edge regions where gradient magnitude exceeds a threshold, while applying noise suppression only to non-edge regions. This local differentiation ensures that image sharpness is enhanced at edges without introducing blurring, while noise suppression is applied only where it will not affect edge definition.
Data Source
AI summary
A method for processing a digital gray value image includes the steps of: generating a binary edge image from the gray value image so that edges present in the gray value image are determined as line areas around the edges; applying a sharpness algorithm in the gray value image within regions which correspond to the line areas; and, carrying out a smoothing process in the gray value image within regions which lie outside of the line areas so that an additional smoothed, sharpened gray value image is generated.