Research On Microscopic Image Processing In Microassembly Robot System

With the deeper exploration of the microcosmic and the rapid development of the micro electro mechanical systems, microassembly and micromanipulation technology has been paid more attention to. Microassembly robot is the combination of robotic assembly and micromanipulation, it has become one of the research hotspots in robotic field. This dissertation mainly aim at the microscopic image processing in microassembly robot system and several topics have been discussed such as the edge detection of the micro objects and the measurement of the object's slight movement based on the subpixel edge detection arithmetic.We firstly give a brief overview of the development of the micromanipulation and microassembly robot. A deep investigation has been given to the key technology of robotic assembly in micro domain, the importance and the function of micrscopic vision system in microassembly process have also been discussed.The designed microassembly robot system which is sponsored by National 863 Plan and its structure are introduced. A detailed introduction of the human-robot task plan interaction for microassembly robot based on XML has been given.The initial process of microscopic vision is related with the detection and recognition of micro object. This dissertation proposes a edge detection arithmetic which expands the traditional Sobel template and combines non-maxima suppression and double grads threshold to get the single pixel wide consecutive edge. With the combination of region growing arithmetic, the light strip and the facula on the surface of the micro objects that result in the reflection effect can be eliminated.Because of the limited width of the CCD pixel, the traditional pixel level location arithmetic can not detect the object's slight movement. A subpixel edge detection based on the polynominal interpolation at the gradient direction of edge is proposed in this dissertation. Both the computer-simulated and the actual images are used to verify the accuracy of the arithmetic. The experimental data shows that the subpixel edge detection arithmetic has the ability to detect the objects'slight movement.