图像质量评价是地基光电探测任务中的基础性的研究工作之一, 与其他事后处理任务密切相关。重点研究了无参、客观自适应光学图像的质量表示与指标提取方法, 总结了目前对图像质量评价研究的基本框架及研究现状。提出了一种基于非线性能量归一化的自适应光学图像质量感知模型, 并结合自适应光学系统与人眼视觉感知的特性, 提取了两个无参、客观像质评价指标, 该指标可用于帧选、复原等关键任务中, 实现对图像质量的自动评价或图像内容变化的自动检测, 评估结果与人眼主观评价呈现出良好的一致性。

Image quality assessment is the basic research work in the field of ground-based photoelectric detection, which is closely related to other post-processing tasks. The quality representation and index extraction of no-reference and objective adaptive optics images were focused on, and the basic framework and current research status on image quality evaluation were summarized. An adaptive optics image quality perception model based on non-linear energy normalization was proposed. Based on the characteristics of adaptive optics system and human visual perception, two no-reference and objective quality indices were then extracted, which could be used in the frame selection, recovery and other key tasks to achieve automatic image quality evaluation or automatic detection of image content change, the evaluation results and subjective evaluation of human eyes show a favorable consistency.