Image Sharpening via Edge-Preserving Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image sharpening methods often result in noise amplification and overshoot, leading to low image quality, particularly in mobile phone images due to hardware limitations.
Innovation Solution
An image sharpening processing method that uses edge-preserving filtering to separate image edge information from noise, determining a sharpening gain based on image edge information and metadata like light sensitivity and exposure time, to prevent noise amplification and overshoot.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional high-pass filter or low-pass filter is used to decompose image into base and detail images, then image definition is improved through sharpening processing, but noise amplification and overshoot phenomenon occur
Solution Approach 1:
The patent changes the filtering approach from traditional high-pass/low-pass filters to edge-preserving filters, fundamentally altering the parameter of filter type. This allows the decomposition process to maintain edge information while separating noise, thereby improving image definition without causing noise amplification or overshoot artifacts
Solution Approach 2:
The patent applies different processing strategies to different regions of the image based on edge detection results. By identifying edge areas versus non-edge areas, the sharpening gain is selectively applied - stronger enhancement in edge regions for definition improvement, and suppressed enhancement in non-edge regions to avoid noise amplification
2Manufacturing precision
If sharpening gain is increased to improve image definition, then image quality is improved, but noise amplification and overshoot phenomenon occur
Solution Approach 1:
The patent implements spatially varying sharpening gain based on local image characteristics. Edge-preserving filters and edge detection enable the system to identify which regions should receive enhanced sharpening (edge areas) and which should receive reduced or no sharpening (non-edge areas with noise), thereby improving definition locally without amplifying noise in other regions
Solution Approach 2:
The patent uses edge detection results and filter output as feedback to dynamically adjust the sharpening gain. The base image from edge-preserving filtering provides feedback about edge locations, which then controls the amount of sharpening applied in subsequent processing stages, preventing noise amplification while maintaining definition improvement
Data Source
AI summary
Embodiments of the present invention provide an image sharpening processing method and apparatus, and a shooting terminal, where the method includes: acquiring to-be-processed image information; performing edge-preserving filtering processing on the image information to obtain base image information and detail image information, where the base image information includes image edge information; determining, according to the image edge information, a sharpening gain required for performing sharpening processing on the detail image information; and performing sharpening processing on the detail image information by using the sharpening gain. It is implemented that the sharpening gain required for sharpening processing is determined by the image edge information, thereby suppressing noise amplification and avoiding generating a phenomenon of overshoot for sharpened image information and improving quality of a sharpened image.