Research On Image Restoration Method Based On Atmospheric Turbulence Related Parameters

Atmospheric turbulence is a form of irregular motion,due to changes in temperature and wind speed,the refractive index of the atmosphere appears unevenly distributed,resulting in a decline in image quality,which has a great impact on long-range target observation.Therefore,the restoration of turbulence-degraded images has research value and is a very challenging problem.Image restoration methods based on atmospheric turbulence-degraded images have been extensively studied at home and abroad,but existing methods have less consideration of the effects of atmospheric turbulence parameters.In order to realize the measurement of atmospheric related parameters and the restoration of turbulence-degraded images,this paper studies the image restoration method based on atmospheric turbulence related parameters based on the long-range binocular ranging project.The main work of this paper is as follows:The influence of different atmospheric turbulence parameters on degraded images is studied.Based on the classical Kolmogrov turbulence degradation model,an improved turbulence degradation model is proposed,the improved model includes parameters such as atmospheric coherence length,turbulence external scale and spatial cut-off frequency,which makes the prior constraint conditions more perfect.Using this model to study the effect of atmospheric turbulence parameters on image degradation,the experimental results show that: atmospheric coherence length has great influence on turbulence image.The smaller the atmospheric coherence length,the worse the image quality.The traditional measurement method of atmospheric coherence length and its shortcomings are studied and analyzed;Combining the theory of atmospheric turbulence,SIFT feature extraction and optical flow method,an improved Lucas-Kanade optical flow method for measuring atmospheric coherence length is proposed,the practical problems of atmospheric optics is solved by image processing algorithm,the numerical value and variation trend of atmospheric coherence length are calculated,the result obtained accord with the variation range of atmospheric coherence length.Based on the improved atmospheric turbulence degradation model,the Richardson-Lucy algorithm based on the Taylor series second-order vector extrapolation acceleration algorithm is used to recover degenerate images.The experimental results show that,the traditional Richardson-Lucy algorithm is compared,and the algorithm has better recovery effect,moreover,the convergence speed is fast and the time is short.