Force And Strain Measurement Of Micro Clamping Structure Based On Vision System

With the rapid development of electronic technology and micro- and nanotechnology, the micro electro mechanical systems(MEMS) has been more and more widely used. Micro manipulation and micro assembly are important tasks in MEMS, with their key technology being the design and control of the micro clamping mechanism. In micro assembly and manipulation, the measurement of applied force and deformation is the key work in the application research of the micro clamping mechanism because it’s important to accurately position and control the micro clamping movement to avoid damaging the fragile micro component and the manipulated object. In micro and nano scale, the micro vision and imaging system is an essential tool to monitor the micro movement and obtain the feedback information for the close-loop control strategy. The efficient integration of the micro vision system, the micro manipulation mechanism and measurement system of micro force and deformation is the key to the successful application of MEMS. Compared to the rapid development of micro and nano manufacture technology and intelligent control technology, the research of micro clamping component and actuator is relatively slow. Micro cantilever is a typical structure for the clamping and actuation. Taken the components similar to micro-cantilever as research object, this paper is to make research on the force and strain measurement based on the micro vision and imaging system.First, micro cantilever is fabricated by the micro and nano lithographic mask electroforming process. Special clamping and bending mechanism is designed according to structure feature of micro cantilever. The bending process is monitored and under the optical microscope, the images of bending cantilever are captured by the CCD sensor and transmitted to the computer by the image acquisition card. The images before and after bending deformation are preprocessed to obtain the clean texture information, which can be used for the measurement of applied for and strain field of micro cantilever.Based on the texture information in the cantilever images before and after bending,the The LOG(Laplace of Gaussian) feature points are detected and a local closed feature region is constructed for each LOG feature point. The position of the LOG feature point is re calculated as the centroid of local closed feature region, and affine invariant feature descriptor is constructed based on gray value in the local closed feature region. The corresponding LOG points in images before and after bending are matched through the feature descriptor. Together with the centroids information of LOG points, the displacement of the matched points can be obtained. According to the size of micro cantilever, the bending deformation formula in the plane stress state is utilized to model the displacement error function which can be used to calculate the value and position of bending force applied on the micro cantilever.The principles of image registration based on B-spline、optical flow、digital image correlation(DIC) for the full field displacement measurement are described, and strain calculation formula is derived for three methods. The superiority of DIC is verified through the calculation of strain field of bending cantilever. Based on the calculated results of correlation coefficient around different feature points, it’s found that the gray distribution around a LOG feature point is advantageous for the convergence of iterative DIC method. The key factors influencing the DIC method are discussed. The effects of order of transformation function、radius of calculation window and initialization optimization with feature point on the calculation error are analyzed through simulated bended images. Two new weighted average method for the strain field calculation are proposed, the validity of weighted average strain method is verified by comparison with results from finite element method(FEM).The calculation model for the force and strain measurement in image sequences is discussed. The effects of random noise on the force and strain calculation of micro cantilever are shown. A kalman filtering method in image sequences is proposed to filter the centroid of LOG feature point in noise image. The error of force calculation result is deceased obviously using the filtered centroid to provide the displacement of feature points. According to the characteristics of the bending deformation of cantilever, a new DIC method based on the image sequences is proposed, the Proportional constraints between the image sequences is added to the DIC method to improve t he anti noise performance of the deformation parameter in DIC.