The Research On Precision Assembly Technology For Quartz Flexible Accelerometer Components

Quartz flexible accelerometer is the core element of inertial navigation system, and the requirement of its assembly precision is very high. The traditional manual assembly method use assembly tools and human eyes to realize alignment and assembly between the parts. However, it has low alignment accuracy, poor consistency, low efficiency and labor intensive, and it is hard to further improve the assembly accuracy and performance of the accelerometer. In this thesis, vision technology was introduced into the field of the accelerometer assembly, and an assembly alignment system was designed. This system can achieve high position and alignment precision, which is just the blank wall in traditional assembly method.The gist of this thesis are listed as follows:1) Alignment tasks are arranged by analyzing the structure of accelerometer and its assembly accuracy. Based on the missions, the alignment system was designed and built, which consists of vision module, alignment platform module and computer processing module. Vision module covers CCD camera, double telecentric lens and light source, etc, whose mainly works are image gathering and storage; Alignment platform module includes x,y axis precision motion platform, rotating platform, etc, which executes alignment operation after component image positioning; Computer processing module includes a series of image processing algorithms and interactive interface, whose tasks are image measurement and positioning.2) A series of image processing algorithm for the pendulum components are analyzed in detail, which includes image preprocessing, the image target feature edge coarse positioning and fine positioning algorithm. For the issue that previous sub-pixel edge detection algorithms will lead to uncorrected positioning due to inaccurate image coarse location, a sub-pixel positioning algorithm that is suited to accelerometer parts image was proposed——Gaussian fitting sub-pixel positioning algorithm based on sliding edge points, whose positioning accuracy is superior to 0.012 pixel in the condition of the ideal circle. The positioning method, which computers the axis of symmetry and the center of figure, was designed under the characteristics of the pendulum components, and this method exhibits higher location precision comparing with previous algorithms. The experiment results show the center positioning uncertainty of this algorithm is better than 0.021 pixel, and symmetric axis orientation deviation is not more than 10-4 °.3) The linear calibration of system was designed by analyzing the imaging model of the camera and the relationship of four coordinate system. Based on the analysis of the component alignment operation process, a software interface for the components alignment operation was designed and the pendulum components alignment experiment was conducted on this system. Experiment show that the coaxiality error is less than 7 ?m and the symmetry error is less than 0.02 °, which can meet the requirements of system alignment accuracy.