计算机科学
自适应光学
图像质量
波前
计算机视觉
像素
算法
变形镜
人工智能
光学
图像(数学)
物理
作者
Huizhen Yang,Xinyang Li,Wenhan Jiang
摘要
The conventional adaptive optics (AO) technique based on phase conjugate principle requires the information about a reference wave front that corresponds to the undistorted image when it is applied to extended source image correction. The AO technique based on the image function appears more promising than the conventional AO for this application. The success of this technique depends on an appropriate image quality criterion and a suitable control algorithm. We tried to find a convenient image performance metric not from additional optics elements, but only from the distorted image which is a gray level image. Simulation results showed the variance function of gray level presents more potential than other ways, such as ways based on the information about the gray level gradient of each pixel. Based on SPGD algorithm, an AO system was simulated with a 61-element deformable mirror and the above resultant image quality criterion. The correction capability of the AO system was investigated through different images and different level aberrations. Numerical simulation results verified the performance metric we used is effective. The system can correct different level disturbed images successfully when parameters of the algorithm were appropriate for static distorted wave-fronts.
科研通智能强力驱动
Strongly Powered by AbleSci AI