Image deblurring method based on multi-parameter regular optimization model
An optimized model and multi-parameter technology, which is applied in image enhancement, image data processing, instrumentation, etc., can solve problems such as unsuitable optical defocus image deblurring processing, etc., and achieve the effect of rich detail texture, less distortion, and no artifacts
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0028] An embodiment of the present invention provides an image deblurring method based on a multi-parameter regular optimization model, see figure 1 , the method includes the following steps:
[0029] 101: By combining the Tychonoff regularization term with the Huber function, construct a multi-parameter regularization optimization model, which is used to deblur the image;
[0030] 102: Convert the multi-parameter regular optimization model into the form of augmented Lagrangian function;
[0031] 103: Minimization Algorithms by Alternating Directions [11] Solving the above augmented Lagrange function An all-in-focus image corresponding to the original defocused image is reconstructed.
[0032] In summary, the image deblurring method proposed in the embodiment of the present invention can deblur grayscale images, text images, non-text images and low-light images, and the restored image contains richer detail textures, less distortion, and no Artifact phenomenon, clearer a...
Embodiment 2
[0034] The scheme in embodiment 1 is introduced in detail below in conjunction with specific calculation formulas, see the following description for details:
[0035] 201: Construct a multi-parameter regular optimization model, which is used to deblur the image:
[0036] Among them, the image blurring process is expressed as:
[0037] d=hf+r (1)
[0038] in, is the existing defocused blurred image, h represents the two-dimensional Gaussian blur kernel, is a potential clear image, hf means convolving h and f, and r means additive noise; represents an n 2 dimensional real number space.
[0039] Given a defocused image A multi-parameter regular optimization model is proposed to reconstruct the clear image f, which can be expressed as:
[0040]
[0041] in, is the fidelity item, α is the weight parameter of the balanced fidelity item, μ 1 and μ 2 is the weight parameter of the regular term. is the discrete form of the Tychonoff regularization term, Represents...
Embodiment 3
[0079] The method in embodiment 1 and 2 is verified below in conjunction with specific accompanying drawing, and experimental data, see the following description for details:
[0080] The embodiments of the present invention verify the effectiveness of the algorithm by comparing the subjective experimental results and the evaluation scores of the objective LR criterion. The parameters of the regular optimization model constructed in the embodiment of the present invention are set to: α=22, μ 1 =0.9,μ 2 = 0.1 and γ = 10, Indicates the step size of each update of κ, which is used to ensure the convergence of the regularized optimization model.
[0081] 1. Subjective experiment
[0082] The method proposed in the embodiment of the present invention is respectively compared with the fast image deblurring algorithm [1] , an image deblurring algorithm based on spectral characteristics [3] , an image deblurring algorithm based on total variation [4] , a natural image deblurrin...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com