Abnormal phase image restoration method

• Main Authors: LIU SHENG, LEI CHENG, MEI LIYE, CHEN SIYAN, WENG YUEYUN
• Format: Patent
• Language: Chinese; English
• Published: 09.03.2021
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention provides an abnormal phase image restoration method. The method aims at solving problems that phase unwrapping cannot be well completed through a traditional algorithm, phase jump exists, and consequently flaws appear on an image. According to the method, abnormal phase unrolling restoration is taken as a research target, modeling is carried out by using a deep learning method, and an effective defect phase image restoration method is provided by research. Experimental results show that the algorithm provided by the invention achieves a better effect in the aspects of subjectivevisual effect and objective evaluation standard. 本发明提出了一种异常相位图像的修复方法。针对于传统算法并不能很好地完成相位解包裹，存在相位跳变，导致图像出现瑕点等问题。本发明以修复异常相位解卷为研究目标，利用深度学习方法进行建模，研究提出了有效的修复瑕疵相位图像的方法。本发明的实验结果表明，本发明的算法在主观视觉的效果和客观评价的标准上都取得了较好效果。