基于点扩散函数估计长曝光大气湍流图像复原

Restoration of the Long Exposure Atmospheric Turbulence Image Based on the Point Diffusion Function Estimation

在大气介质中工作的光学成像设备因为受到大气湍流的影响,获取的图像存在抖动、畸变、模糊以及光照不均等影响。在遥感、天文观测以及远程监视等领域需要对这一类的退化图像进行处理以求获得清晰图像。利用大气湍流退化的点扩散函数可以对模糊图像进行复原,但在实际的情况下,点扩散函数通常无法精确获得。结合研究背景,先针对长曝光大气湍流退化图像使用APEX算法,并通过结合图像的频谱特征对APEX算法中的点扩散函数(PSF)估计过程进行改进,通过该方法可以较为精确地获得大气湍流模型的点扩散函数(PSF),然后优化拟合算法,恢复出的图像可以拥有较为良好的视觉效果,通过对比改进前后图像的灰度平均梯度值和拉普拉斯梯度模这两个客观评价标准,进一步验证该方法的有效性。

The optical imaging equipment working in atmospheric media is affected by the turbulence of the atmosphere,and the images obtained are affected by jitter,distortion,ambiguity and uneven illumination.This kind of degraded image needs to be processed in remote sensing,astronomical observation and remote monitoring to obtain a clear image.The point diffusion function of atmospheric turbulence degradation can be used to reconstruct the fuzzy image,but in the actual situation,the point diffusion function can not be precisely obtained.Based on the research background,uses APEX algorithm for the long exposure of atmospheric turbulence degradation images.And by combining image spectrum feature of APEX algorithm of Point Spread Function(PSF)estimation process is improved,by the method can more accurately obtain the PSF of atmospheric turbulence model,and then optimize the fitting method,to restore the image can have a good visual effect,by comparing the average gray level gradient value of the image before and after improvement and Laplace gradient mode the two objective evaluation standard,further verifies the effectiveness of the proposed method.