Research On Key Technologies Of Automatic Assembly System Of Laser Gyro Dithering-wheel

Since laser gyros have advantages of simple structure, long service life, highreliability, wide working temperature range, good repeatability, they are widely used inthe field of aviation science and technoly. Dithering-wheel is an important componentof laser gyro, and it is assembled by manual method, which is low efficiency and pooronsistency. Therefore, this paper developed an automatic assembly system of laser gyrodithering-wheel.General planning of this system was determined by analysing of the assemblyprocess of the dithering-wheel. According to function requuirement of the system, thedesigned system has a mechanism including9modules. Glue dispensing and thedetection of the thickness of glue solution are key technologies of this assemble system,so detail mechanism design and selection of some correlative parts were presented.Hardware and software of control system of the auto-assembly system were designedbased on the modular design method, which make the control system had goodscalability and reliability. Control software which achieved human-computer interactionof this system was programed with VC++on PC. Multi-threading was used forcontrolling progress, the realized multi-task parallel execution could drastic reductionwork time, optimize work progress and improve the efficiency of the system.Limited by factors, such as space and glue solution viscosity, the process of gluespread is devided into two steps, and the first one is fluid dispesing, the second is tospread the glue solution and make it well-distributed. The mathermatical models for theglue spread were etabilshed in this paper, and simulation experiment was excuted bymodule discretization. Through comparing the resules of five different kinds ofdispensing path, the route of three discrete point type dispensing has the bestcomprehensive advantages, and it was chosen as the final dispensing path, thesimulation experiment results showed that its largest coverage ratio of solution reached89.34%, which meet the requirements.For the detection of the thickness of glue layer, this paper designed an automaticfocusing system and selected focusing evaluation fuctions to obtain clear images. Afterfiltering out noise in the image by median filter, the improved Canny edge detectionalgorithm was used to detect edge information. The concept of gravitational fieldintensity was introduce to replace image gradient, and the traditional22operator wasreplaced by33gravitational field intensity operator. An adaptive threshold selectionmethod was proposed based on the mean of gradient magnitude and standard deviation.Morphological closing operation to image made the edge line more coherent. Select theedge line according to the geometric feature extraction, and use linear least squaremethod to fit the synthesis. The distance between the fitting line segments in the image represents the thickness of the glue layer. Through experiment, the actual size of imagein view was calibrated. Finally, this paper designed image processing to complete thetask of the non-contact detection of glue layer thickness, which was proved byexperience and can meet the design requirements.