An image deblurring method based on complete defocus region segmentation
CN117611491BActive Publication Date: 2026-08-28SICHUAN UNIV
Patent Information
- Application Number
- CN202311642500.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-12-04
- Publication Date
- 2026-08-28
- Estimated Expiration
- 2043-12-04
AI Technical Summary
Technical Problem
[0003]基于反卷积去模糊算法中，有一种基于区域分割的方法，将输入图像分割成模糊量近似均匀的多个区域，然后在每个区域上分别进行去卷积恢复出清晰图像；该方法因为无法有效区分图像的结构边缘和纹理边缘，容易使获得的散焦图受到纹理边缘的干扰，从而导致过度分割或者区域分割不准确；另外该方法在散焦图估计和去卷积过程均使用了基于高斯分布的点扩散函数，该类点扩散函数会受光频率不同所带来色差的影响，导致恢复出的图像存在振铃效应
Benefits of technology
[0015]本发明提供的方法，恢复出的图像不易受振铃效应影响，更加符合实际情况，有利于后续的人类感知和计算机处理。
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN117611491B_ABST
Abstract
The application provides an image deblurring method based on full defocus image region segmentation. The method uses an improved point spread function derived from Rayleigh-Sommerfeld diffraction formula to simulate the defocus process of a focused image after propagation in the image space, obtains the blur parameters on both sides of the edge pixel points by re-blurring the original image, and thus calculates a sparse defocus image and a full defocus image. The image segmentation of the full defocus image is performed to obtain regions with different blur degrees. Different Wiener filtering parameters are used for different regions, and finally a clear image is obtained. The method can remove the texture edges while retaining the structure edges, so that the region segmentation is more accurate, and a better image deblurring effect is achieved.
Need to check novelty before this filing date? Find Prior Art
Citation Information
Patent Citations
Method for restoring weighting sparse edge regularization image
CN102147915A
Defocused image deblurring method based on boundary neighborhood gradient difference
CN115578289A