A Study of Passive Component Recognition By Using Neural Network

• Main Authors: Kuo-Feng Huang, 黃國峰
• Other Authors: 陳澤民
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/82353075070441809046

碩士 === 國立中興大學 === 生物產業機電工程學系所 === 99 === In this research, an automatic passive components recognition and classification system was established by using image process technology and artificial neural network. The MATLAB Language integrates the image and, recognition processing simultaneously. This system is able to take place of the traditional manual recognition and classification operation and results in the benefits of times-saving and cost-down. In this study, the original image from CCD is binarized to acquire the binary image first. Then, a filter was employed to remove the noise from the binary image and the characteristic parameters such as the perimeter, the area, the color and the moment invariants of the passive component were measured. Finally, integrating the Backpropagation Neural Network model, the recognition and classification operation for passive components was accomplished. The available recognition rate of 200 test samples by using the developed system is as high as , 96.5% with which validates this developed system.