基于指导滤波的单次曝光强度关联成像研究

针对低采样条件下单次曝光强度关联成像重构图像质量不理想问题,提出一种基于指导滤波的单次曝光强度关联成像重构算法。首先,通过快速投影Landweber正则化获得较低质量的初始估计图像,然后将该图像作为指导图像,采用指导滤波进行多次联合迭代,最后获得去噪后的高质量重构图像。在不同采样率条件下与基于压缩感知全变差低秩约束算法的重构图像进行了对比分析,数值模拟和实验结果均证明了本文算法的有效性。

Single Exposure Intensity Correlation Imaging Based on Guided Filtering

A single exposure intensity correlation imaging restructuring method was proposed in the paper below as a solution to unsatisfactory quality of single exposure intensity correlation imaging restructuring under low sampling conditions,wherein a relatively low-quality initial estimated image was obtained by means of rapid projection Landweber regularization,which was then taken as guided image subject to multiple joint iteration with guided filtering to obtain the final high-quality restructured image after noise reduction.Through a comparison between it with the restructured image obtained through total variation low rank constraint algorithm based on compressed sensing,the result of numerical simulation and experiments shows that the new algorithm is effective.