CCDカメラによる表面検査とニューラルネットワークを用いた電子部品の故障推定

概要

論文の詳細を見る
Reliability estimation of electric parts by using CCD camera and neural network is described. Breakdown of carbon film resistors in reliability test is estimated from image features by feedforward neural network. In general, degradation of electric device should be detected for constantaneous operation of every machines. Image processing techniques plays an important role in many engineering applications, developments and more efficient quality control. This paper describes some experimental results on reliability estimation using a neural network and image processing for reliability test. Performance degradation of carbon film resistor is estimated from their colour images. By using the sequence of images of the tested carbon filmed resistors, we attempt to find an appropriate estimated fault time for tested carbon film resistors, in order to establish a convinient method for reliability estimation. We first discuss some essential issues to be considered in reliability problems. It is shown that the image processing and neural network approach give a good results compared with real measured results. Furthermore, the efficiency of the method is indicated by applying to image data for calculation of mean time to failure.
一般社団法人日本機械学会の論文
1999-06-25