Research On Vision-based Automatic Microassembly Oriented To Microfluidic Chip

Plastic microfluidic chip is a novel kind of micro total analysis system. It can be used in chemical analysis with respect to small amounts of samples, high analysis speed and good sensitivity. It can employ many popular detection methods including UV absorption, laser induced fluorescence and electrochemistry detection. Among them UV absorption detection is the most popular one with low cost. However detection precision will be destroyed when UV radiation is absorbed by polymer PMMA material of microfluidic chip. To solve this problem one method is to join a quartz capillary to the end of the microchannel of plastic chip. In this way detection can be carried out on the quartz capillary and detection precision is assuring. Miniature function device and interface device are applied to microfluidic chip, which has accelerated the industrialization and commercialization process of microfluidic chip. Automatic microassembly system is the key equipment to realize the assembly of miniature function device and interface device.The dissertation aims to realize exact butt joint between capillary and microchannel by assembling the capillary to the end of microfluidic chip. System built consists of three degree-of-freedom stages, microgripper, visual device (microscope, frame grabber, and CCD camera), automatic adhesive dispenser and UV curing lamp for microjoining with adhesives. Software was designed with the tool of Visual C++ for this microassembly task in combination with the microassembly system. With the visual feedback the relative positions between capillary and microhole can be detected. Accordingly the positions three degree-of-freedom stages are controlled. Problems and difficulties are summarized in this paper. Several algorithms about image processing, detection, identification and depth recover techniques are compared. And suitable visual treatment methods are decided. The relationship between image coordinate system and world coordinate system is confirmed by the calibrations of microassembly and visual systems. Edge images of identifying objects captured by frame grabber are got after image pretreatment, binarization and edge detection. Then the key point's coordinates of capillary and microhole are assured. Two methods automatic focusing and stereo- visual are used here for depth recover. And advantages and disadvantages between them are expatiated. Moreover research experiments about the microassembly of capillary and microfluidic chip were done, which have proved that the assembly assignment can be finished successfully.