Phase Retrieval Based On Focus Positioning Of Intensity Images And Registration Progressive Compensation

Phase retrieval based on the transport of intensity equation is an important non-interference phase measurement method,which overcomes the shortcomings of the interference method,such as the harsh light source requirements and high experimental environment requirements.It is widely used in the depth detection of small objects,biomedicine and other fields.In the process of solving the equation,it is necessary to collect a focused image and obtain defocused images based on the focused image.Inaccurate positioning of the focus position will reduce the precision of the phase retrieval result;On the other hand,in the phase retrieval process of partially coherent illumination for the transport of intensity equation method,since the CCD acquisition frequency is not synchronized with the LED frequency,which case the black and white stripes appearing in the phase retrieval result,as well as the phase result error caused by the acquisition angle deviation due to the mechanical movement.Aiming at the above problems,this thesis has carried out corresponding research.The main research work and innovations are as follows:(1)A phase retrieval algorithm based on edge detection and duty ratio fusion is proposed.A series of intensity images are extracted by edges to obtain edge information points;the duty ratio evaluation standard is selected to locate the best focus position in combination with the image focusing principle;finally the defocus image is collected to solve the equation to retrieval phase.Firstly,a series of ‘subjectively determined focused images' collected by imaging system are edge-extracted to obtain corresponding edge information points.Then the best in-focus position is found by calculating the duty ratio,and the corresponding defocused images are selected according to the position of the focused image.Finally,the high-precision phase retrieval result is obtained by using these images of the accurate position to solve the equation.(2)A registration progressive compensation algorithm is proposed.First,the corner detection and cross-correlation matching are performed on the collected intensity images,and error points are removed to get the transformation matrix in order to correct the original deviation.At the same time,initial compensation is performed in the new registration imageto improve the lighting stroboscopic effect.Then,the initial phase of the sample is solved by using Transport of intensity equation.In order to resolve the halo appearing at the edge of the initial phase,the phase is compensated for the highlighting of the vital sample information again.