Image Processing Method Of The Stereo Light Microscope

With the continuous development of the science and technology, in the field of theengineering, medical and biological, the operation objects become smaller and smaller, the stereovision system of macro-world can’t meet people’s needs, therefore urgently require thedevelopment of some methods and techniques which are applicable to micro-world objects forthe operation of the micro-world objects, such as the micro-operations, micro-injection,micro-surgery and micro-measuring and so on. The common feature of the all techniques aresmall-scale objects and can’t be manual operation directly, In order to facilitate observation andoperation we must be by means of the digital microscope, charge coupled device(ChargeCoupled Device, CDC) and other equipment to amplify the micro-object operation. As thedemand of the people, microscope stereo vision system which is based on the stereo lightmicroscope is came into being.First of all, in order to obtain accurate3D measurement of microscope, high-precisioncalibration method is indispensable. Regarding the difficult to move electronic eyepiece of stereolight microscope, we use traditional camera calibration method to obtain accurate calibrationresults in this paper. Additionally, because the target objects which are obtained by electroniceyepiece are several times larger than the original ones, the difficulty making of traditionalcheckerboard can’t be apply to the micro camera calibration, in this paper, we use the gridcalibration board. Although the calibration board is solved, the existing corner detectionalgorithm is no longer applicable to the grid corner coordinates extraction. In this paper,considering the specificity of microscopic image, we propose an improved corner detectionmethod to obtain the actual corner coordinates information of the grid. In this paper presented amicroscope camera calibration method based on grid corner detection for the particularapplication area of microscopic measurement. The method considered not only the radialdistortion, but also other non-linear factors, such as centrifugal distortion and thin-prismdistortion. First of all, the actual corner coordinates information of the grid was obtained throughthe improved corner detection method. Then, the matrixes of lens distortion parameters and camera internal parameters were gotten according to the camera imaging model. Finally, throughthe established non-linear camera model, the average error between fore-projection andre-projection grid corner coordinates was obtained by re-projecting the grid corner coordinates.Experimental results show that the corner detection algorithm is accurate, which is applicable tomicroscopic camera calibration.Secondly, because in this paper the use of grid is different from traditional checkerboard, theexisting sub-pixel corner detection algorithm is no longer applicable to the grid sub-pixel cornercoordinates extraction. This paper presented a microscope camera calibration method based ongrid sub-pixel corner detection for the particular application area of microscopic measurementand improved Harris corner detection method. First of all, the actual corner coordinatesinformation of the grid was obtained through the improved Harris corner detection method. Then,considering the distribution law of corner coordinates in the microscopic image, the paperobtained the sub-pixel corner coordinates by combining the quadratic surface and linear fitting.Finally, through the established non-linear camera model, the average error betweenfore-projection and re-projection grid corner coordinates was obtained by re-projecting the gridcorner coordinates. Experimental results show that the sub-pixel corner detection algorithm isaccurate. Compared with improved Harris corner detection, which is applicable to microscopiccamera calibration.Then, the key technology of microscope stereo vision system is the microscopic stereocalibration which is the process of to calculate the spatial geometric relationship of the twoelectronic eyepieces, in the basis of traditional single-channel micro-camera calibration methodwhich has been established, the stereo micro-camera calibration method is established bysingle-channel micro-camera calibration imaging model extended to microscope imaging model.Using the calibration results of the microscopic stereo calibration to correction the distortion lens,re-projection the two cameras imaging plane so that they fall exactly on the same plane, andthree-dimensional images of the polar aligned completely parallel structure, making the stereomatching more accurate and simple.Finally, in the process of collection image by stereo light microscope, as the impact of thelight, the exposure of electronic eyepiece and camera noise, the follow-up a series of work moredifficulties to process. In order to reduce the color deviation of the microscopicthree-dimensional image, in the final chapter we describe the color correction technology of thethree-dimensional microscopic images. The global mutual compensation color correctionalgorithm which is based on mean and variance information is being considered the differencepixel values of microscopic three-dimensional image between the left and right channel, to strike the mean and variance information of the two channels, according to those information toachieve the purpose of color correction; The global mutual compensation color correctionalgorithm which is based on mean information is being considered the difference pixel values ofmicroscopic three-dimensional image between the left and right channel, to strike the meaninformation of the two channels, according to those information to establish the color mappingand to achieve the purpose of color correction; The global mutual compensation color correctionalgorithm which is based on mean difference factor information is being considered thedifference pixel values of microscopic three-dimensional image between the left and rightchannel, to strike the mean and variance information of the two channels, according to thoseinformation to calculate mean difference factor, using the factor to establish the color mapping,then by controlling the number of iterations can correct the mean of the two channels similarly.