Research On Micro-nano Depth Information Recovery Method Based On Defocus Images

Depth information recovery refers to the use of the relationship between the features of the two-dimensional image and the depth of the target object to obtain the depth information of the target object from the two-dimensional image,which is widely used in three-dimensional reconstruction,robot navigation,industrial detection,medical imaging and other fields.Depth information recovery method based on defocus images has attracted extensive attention due to its lack of required images,overcoming matching and occlusion between images,non-existence of characteristic correlation.Especially in micro-nano observations,it is not allowed to change any of the parameters in any form after the observation of the microscope.Therefore,by extracting the defocus images with the microscope and restoring the depth information of the target object,it is possible to achieve better observation of the sample.Based on the defocus images,this paper studied the micro-nano depth information recovery.The main research contents are as follows:Firstly,research on the depth from defocus with fixed camera paraters.The fuzzy imaging principle,relative ambiguity and depth information recovery process of the defocus depth recovery method were introduced.The depth from defocus based on the thermal radiation equation was studied.The advantages and disadvantages of the traditional depth from defocs dwere analyzed by the depth information recovery experiment of the standard 500 nm scale grid.Secondly,research on the accuracy of depth information recovery of defocus images.In order to solve the problem that the objective function of the traditional depth from defocus is ill-posed by the Tikhonov regularization method and the accuracy of depth information recovery is not good,the TV regularization method and the L-curve method were introduced in the objective function of the traditional depth from defocus.A depth information recovery algorithm(L-TV)based on TV regularization and L-curve was proposed,which effectively improved the accuracy of depth information recovery.The experiments of depth information recovery through the standard 500 nm scale grid show that compared with Tikhonov regularization methods and TSVD regularization methods,the L-TV algorithm proposed in this paper can avoid the excessive punishment of the restored depth information,tend to be smooth,which can effectively improve the accuracy of deep information recovery.Finally,research on the effectiveness of depth information recovery of defocus images.Aiming at the problem that the depth of information recovery is not efficient by using the iterative shrinkage threshold algorithm(ISTA)to solve the dynamic solution process of depth from defocus.based on the ISTA algorithm,an ISTA algorithm based on accelerated operator gradient estimation and secant linear search was proposed.(FL-ISTA),speeding up the dynamic solution process of depth from defocus,and effectively improving the efficiency of depth information recovery.The experiments of depth information recovery through the standard 500 nm scale grid,conductive probe and triangular probe,it can be seen that compared with the iterative shrinkage threshold algorithm(ISTA)and the fast iterative shrinkage threshold algorithm(FISTA),the FL-ISTA algorithm proposed in this paper can speed up the dynamic solution process of depth from defocus,and effectively improve the efficiency of depth information recovery.