Application Of Back-Propagation Neural Network To LED Die Defect Inspection
碩士 === 遠東科技大學 === 機械工程研究所 === 100 === This thesis uses the back-propagation neural network to recognize two types defects of LED die automatically. First of all, a LED die recognition process will be developed. The Otsu's method and template-matching techniques are employed in this recognition...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Published: |
2012
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/37603930264899335770 |
| Summary: | 碩士 === 遠東科技大學 === 機械工程研究所 === 100 === This thesis uses the back-propagation neural network to recognize two types defects of LED die automatically. First of all, a LED die recognition process will be developed. The Otsu's method and template-matching techniques are employed in this recognition process to find a single LED die from the LED wafer image. Through the multi-sample algorithm and the simple logic operators, the characteristics of LED die defects will be obtained quickly. These characteristics of LED die defects will be the training samples of the neural network. After the training process, the LED die defects detection system will be examined by a lot of the single LED die images. The defects detection system will identify whether the defect of the sample images.
The results show that the proposed feature extraction method can increase the detection accuracy effectively. The detection accuracy of these two types of LED die are 99.23% and 84.81%. The time of detection for every single die is about 70 milliseconds on average.
|
|---|