Research On Micro Drill Measuring System Based On Machine Vision

With the growth of electronics and information industry, the size of micro drill is becoming smaller than ever. In traditional manufacturing process, workers must sample frequently and measure the dimension of micro drill with microscope, image measuring instrument and web thickness measuring tool, then adjust the parameter or mechanical parts of manufacturing equipment according to the measurement results. The measuring instrument’s accuracy is low and the measuring data cannot be analyzed easily, manual inspection can be tiresome. In order to improve the quality of manufacturing process, a micro drill measuring system is designed. All the blade defects, longitudinal parameter and web thickness can be measured in one machine. The results can be sent to server via intranet and analyzed to implement automatic parameter tuning and process control of manufacturing equipment.In this paper, the manufacturing process of micro drill is analyzed. According to the structure characteristics, proper illuminant, lens and cameras are selected for blade defects, longitudinal parameter and web thickness measurement. Then the mechanical system and motion control system are analyzed. In order to inspect blade defects for micro drills of different specifications, digital light controller and binary search algorithm are used to get proper light intensity. OpenCV is used as image processing library, hough line detection algorithm is used to find the incline angle of center line. After rotation all micro drills will have the same orientation, so ROI can be easily setup. Then segmentation and Zernike moment are used to locate the subpixel edge, and fit algorithm is used to calculate blade defects.Back lights are used for both longitudinal parameter and web thickness measurement, Canny edge detection and segmentation edge detection are used to calculate point angle and relief angle respectively. Connected component labeling algorithm is used to measure flute length. UC diameter is an important parameter, in order to improve accuracy, three kinds of subpixel algorithms are compared, gray level moment algorithm is used, blank is then used to calibrate measure system, result shows maximum deviation is only 0.5μm. A novel web thickness measuring method is then proposed, micro drills whose outer diameter are less than 0.5mm can be accurately measured, repeatability can achieve 1μm.Finally, the architecture and implementation of software are discussed. Then analysis of variance is used to measure the repeatability and reproducibility of the measuring system. Experiments show the system can meet the proposed requirement. It can play an important role in micro drills process control and improving the quality.