A method for selecting a guided filter regularization parameter and related devices
By adaptively selecting the guided filter regularization parameters, a generalized histogram is constructed and equalized using variance-pixel histogram and gradient information. This solves the balance problem between enhancing weak details and suppressing noise in infrared image enhancement, thus improving image quality and filtering effect.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- LIAONING HONGYANHE NUCLEAR POWER
- Filing Date
- 2026-04-23
- Publication Date
- 2026-07-17
AI Technical Summary
In existing infrared image enhancement technologies, traditional methods struggle to effectively enhance subtle details while preserving image thermal distribution information. Furthermore, improper selection of guided filter regularization parameters can easily lead to image quality issues such as gradient inversion, over-sharpening, or overexposure.
By calculating the variance-pixel histogram and the squared gradient value, the regularization parameters of the guided filter are adaptively selected, a generalized histogram is constructed and equalized, and the base layer image is output, achieving the best balance in different regions.
It improves the filtering effect and quality of the image, suppresses noise in flat areas, and preserves edge and detail information, solving problems such as gradient inversion and over-sharpening.
Smart Images
Figure CN122415342A_ABST