The Design And Development Of Machine Vision System For Flipchip Pickup

With the rapid development of electronic science and technology, LCD (Liquid Crystal Display) seems more and more manifest in all our lives. With the growth of the pixel and pin density in LCD, the connecting precision becomes higher. At present, main domestic COG (Chip on Glass) equipments depend on import so that the development of high-precision COG equipments is significant. The machine vision technology is one of the most important parts of COG equipments because it improves the precision and efficiency. For that reason, the research on the machine vision system for flip-chip pickup is meaningful. The main research work is as follows:1) According to the workflow and process requirement of chip-pickup module in COG equipment, the relationship among the machine vision system, host computer and bottom movement system was analyzed and the framework of the vision system for chip-pickup module was established.2) According to the requirement of the chip-pickup vision system and based on the former COG equipment hardware, the hardware for this system was selected and the platform was established with reference to the mainstream hardware in the market.3) The original picture is pre-treated by threshold and edge-detecting. A novel chip detecting and positioning system was developed. The detecting precision and efficiency can meet the manufacturing requirement after a series of optimized treatment.4) By using the C# language, the chip-pickup vision system software, Chip-Detecting, was developed and a friendly human-computer interface and easy operating workflow was designed. The system outputs the position and angle data for chip-obtaining. Also, this novel machine vision system can possibly replace the FVX developing pack and decrease the developing cost.5) By comparing the novel algorithm with the FVX vision function library and general layering searching algorithm, the advantage and disadvantage of these three algorithms was proposed.6) By introducing the statistical process control methods in industry and qualifying the matching precision, the control chart for Chip-Detecting can be created according to the knowledge of statistics and the data for chip detecting can be controlled.