뉴럴 네트워크 기반 연성 인쇄 회로 기판 검사 시스템
An Inspection System for Flexible Printed Circuit Board based on Neural Network
- 주제(키워드) Filtering , Flexible Printed Circuit Board , Inspection , Neural Network
- 발행기관 아주대학교
- 지도교수 박상철
- 발행년도 2015
- 학위수여년월 2015. 8
- 학위명 석사
- 학과 및 전공 일반대학원 산업공학과
- 실제URI http://www.dcollection.net/handler/ajou/000000020774
- 본문언어 영어
- 저작권 아주대학교 논문은 저작권에 의해 보호받습니다.
초록/요약
This thesis judges the excessive images among images which are objects to inspection in visual inspection, and suggests the inspection system based on neural network to minimize the images which was regarded as faulty image from image of fair quality. F-PCB (Flexible Printed Circuit Board) operates to inspection quality firstly by the inspection device, the image regarded as faulty should be the object to inspection in macrograph. The objects of visual inspection need to take second inspection by professional people, but they shall be wasted and shall decrease their unity of qualities because the number of images are too much. For these matters, there need to have the inspection system for figuring out whether something would be fair quality or faulty. To structure the proposed inspection system, using the neural network can solve these matters. Neural network have been known as the level of studying or the level of testing. In the stage of studying, it can be far different depending on the data; consistency of data is important. Visual inspection images involve noisy due to the place or position of inspection device. This paper suggest ‘Sweep Searching Filtering,’ ‘Island Filtering’ to find a key from these matters because images involving noisy let down the consistent data. Moreover, it also suggests four different types of methods to extract a parameter, characteristic of images for studying. The suggested inspection system in this studying will judge the excessive images, have an objective to minimize the faulty image from image of fair quality, and demonstrate the efficiency of inspection system through the studied neural network.
more목차
Chapter 1 Introduction 1
Section 1 The Background of a Research 2
Section 2 Inspection of F-PCB (Flexible Printed Circuit Board) 4
Section 3 The related Research 7
Section 4 Purpose of Research and Organization 8
Chapter 2 Neural Network 9
Section 1 The Introduction of Neural Network 10
Section 2 Model of Neural Network 13
Section 3 Learning Method of Neural Network 14
Chapter 3 Technique of Image Filtering 18
Section 1 Introduction of Technique: Previous Image Processing 19
Section 2 Sweep Searching Filtering 22
Section 3 Island Filtering 27
Chapter 4 Extraction of Parameter: Image Characteristic 31
Section 1 Object Images Characteristic 32
Section 2 Method to Extract Parameter : Object Images Characteristic 33
Section 3 Method of Application Parameter Characteristic 43
Chapter 5 Implementation Inspection System for F-PCB and its Result 45
Section 1 The Implementation and Result of Inspection System based on Neural Network 46
Chapter 6 Conclusion and Hereafter Research 48
Bibliography 51
Abstract 53
