حذف آشفتگي و بازيابي تصاوير با استفاده از روش‌هاي يادگيري ماشين

Turbulence Removal an‎d Image Restoration Using Machine Learning Approaches

مهندسي هوش مصنوعي

Abstract In this thesis, a novel method for removing atmospheric turbulence from a sequence of turbulent images an‎d restoring a high-quality image is presented. Turbulence is modeled using two factors: the geometric transformation of pixel locations represents the distortion, an‎d the varying pixel brightness represents spatiotemporal varying blur. The main framework of the proposed method involves the utilization of low-rank matrix factorization, which achieves the modeling of both the geometric transformation of pixels an‎d the spatiotemporal varying blur through an iterative process. In the proposed method, the initial step involves the selec‎tion of a subset of images using the ran‎dom sample consensus method. Subsequently, estimation of the mixture of Gaussian noise parameters takes place. Following this, a window is chosen around each pixel based on the entropy of the surrounding region. Within this window, the transformation matrix is locally estimated. Lastly, by considering both the noise an‎d the estimated geometric transformations of the selec‎ted images, an estimation of a low-rank matrix is conducted. This estimation process leads to the production of a turbulence-free image. The experimental results were obtained from both real an‎d simulated datasets. These results demonstrated the efficacy of the proposed method in mitigating substantial geometrical distortions. Furthermore, the method showcased the ability to improve spatiotemporal varying blur an‎d effectively restore the details present in the original image.

5