Implementation of the Machine Vision System of Inspecting Nozzle

In this paper, we implemented machine vision system of inspecting nozzle to be used a part of SMD equipment. To inspect nozzle, the grey image is converted to binary image using bimodal technique. Its system first remove noise using morphological operation such as opening and closing, and then extract feature using template matching technique to identify type of nozzle. Finally, its system extract another feature to calculate the size of region of interest using projection technique. A projection is the sum of the run-lengths of 1s along a straight line oriented at angle, which is useful in shape analysis as well as image segmentation. We implemented its system and obtained satisfied results.