Relevant Technology Research On The Micro Vision Of The Microrobot Assembly System

This thesis describes relevant technologies of the micro vision of the microrobot assembly system and focuses on the research of auto focusing and depth extraction, which has a wide application prospect and important research significance in the field of roboticized microparts assemblage, micro manipulation and integrate circuit packaging.The microrobot assembly auto focusing system, combined coarse with fine adjustment method, applies variance operator combined with optimized hill climbing searching algorithm to ascertain the focalizing object area in the coarse focusing. The system then uses gradient square operator to evaluate the microscopic image's definition value in the fine adjustment and utilizes least square method fitting to gain the focal plane position. The experimental results show that the auto-focusing system is characterized by speediness and high definition, which satisfies the focusing requirement of the microassembly.The microrobot assembly depth extraction system uses a innovative'depth from defocus'method. After gathering microarm images in the micro- assembly process, the system firstly uses Frame different with threshold to inhibit the background interference in these images and then applies gray random- distribution variance operator to compute their defocus image features. The experiment acquires the microarm depth location through the depth and defocus feature curve. And this is the foundation of microrobot assembly visual servo along Z axis.