| Femtosecond laser two-photon polymerization processing has processing resolution beyond the limit of optical diffraction and the ability to fabricate complex three-dimensional structures,and is widely used in microelectromechanical systems,biomedical devices,micro-optical devices and other fields,which require high positioning accuracy of the laser focus center at the processing start point.Since the traditional positioning methods often have large errors and slow efficiency,in order to improve the positioning accuracy and processing efficiency,this paper conducts a microscopic vision precision positioning study for two-photon aggregation processing to achieve precise positioning of the laser focal point.The full research work includes the following aspects:(1)For the problem of weak feedback signal for real-time precise positioning of laser focus,a precision positioning method based on background differencing is proposed.The problem that polymeric elements are not easily observed when the focal point initially reaches the substrate of the glass sheet is solved by using the background differential method.The edges of the polymeric voxels are precisely extracted using a sub-pixel edge extraction algorithm based on the local area effect,which makes it possible to calculate the area of voxels in the image more accurately.The grayscale changes of the corresponding image area when the refractive index changes due to different cured voxels are investigated.The relationship between the grayscale of the center of the image and the laser power and exposure time is investigated when the voxels are fully aggregated.The relationship between the area of the body element and the position of the motion axis at the same exposure time and different laser power was investigated.Finally,the array of voxels processed by the method in this paper was inspected by white light interferometry,and the height error of the voxel array dropped from ±300nm to ±50nm after compensation,and the overall autofocus process took 40 seconds on average.Verified the effectiveness of the precise positioning method in this section.(2)For the problem that it is difficult to precisely localize the two-photon convergent laser focal point after it is detached from the substrate,an assisted localization method based on the improved Nonsubsampled shearlet transform wave transform of the image sharpness evaluation function is proposed.The energy gradient function,Tenengrad function,Laplace function algorithm,and wavelet function sharpness evaluation algorithm are compared,and it is proved that the algorithm in this paper has excellent sensitivity and robustness.The proposed line positioning method combined with sharpness evaluation solves the problem of locating the focal point within 10 μm,and the position with the highest sharpness evaluation value is the best locating point by evaluating the sharpness of the line structure intercepted on the glass substrate,and the average locating time is 25 s.The proposed surface positioning method combined with sharpness evaluation solves the problem of locating the focal point within 30 μm by processing the multilayer pyramid structure,and searching for the nearest location.Finally,2 groups of voxel arrays are processed to verify the effectiveness of the method.(3)To verify the effectiveness of the precise positioning algorithm proposed in this article in actual machining,large-scale cylindrical superhydrophobic structures were processed.Finally,the processing results were measured,with an average height error of2.01% and an automatic focusing success rate of 99.1%.In this experiment,the overall processing time of the optimized positioning method was about 10 minutes faster than that of the unoptimized one. |
