Defect detection for stamped part of hard disk shell
碩士 === 國立高雄應用科技大學 === 光電與通訊工程研究所 === 105 === Free edge is an important issue in sheet metal stamping as the stamping process affects burr appearance and thus influences product quality. In USB drive frame manufacturing, samples of cut frames are taken immediately after stamping and examined by human...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2017
|
Online Access: | http://ndltd.ncl.edu.tw/handle/4xe9np |
id |
ndltd-TW-105KUAS0801015 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-105KUAS08010152019-05-15T23:32:20Z http://ndltd.ncl.edu.tw/handle/4xe9np Defect detection for stamped part of hard disk shell 硬碟殼沖壓件瑕疵檢測系統 TU, HSUN-CHENG 涂勛城 碩士 國立高雄應用科技大學 光電與通訊工程研究所 105 Free edge is an important issue in sheet metal stamping as the stamping process affects burr appearance and thus influences product quality. In USB drive frame manufacturing, samples of cut frames are taken immediately after stamping and examined by human inspectors traditionally. However, human operators have several disadvantages compared to smart machines including subjectivity, productivity, consistency, and repeatability. Early defect detection is critical for reducing waste of raw material and achieving a high quality product. A great amount of time is therefore spent during the product development process on the optimization of stamping defects, particularly defect lows during tryout. By implementing such technologies in this thesis, the localized energy analysis with cross-projection is first adopted to capture the stamping region of component. Burr along the stamping region with thickness larger than 10 micrometers is classified with the proposed up and down ramping criterion. Then, the bilateral filter is applied to enhance regions with scratch and contamination. Energy analysis and cross-projection are also combined to detect the scratch and contamination regions by following with localized histogram analysis. Using samples taken from the local company based on Kaohsiung city, the experiments of stamping defects have been proved that the proposed framework well performed with 96.39 % recognition rate for the burr defects and 100 % recognition rate for the surface defects, respectively. The proposed framework is possessed of the advantage of autonomous detection with high recognition rate. Leveraging the unique integration of stamping on sheet metal with defect inspection has led to the development of new technologies that transform the performance of stamping press. WANG, JING-WEIN 王敬文 2017 學位論文 ; thesis 66 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國立高雄應用科技大學 === 光電與通訊工程研究所 === 105 === Free edge is an important issue in sheet metal stamping as the stamping process affects burr appearance and thus influences product quality. In USB drive frame manufacturing, samples of cut frames are taken immediately after stamping and examined by human inspectors traditionally. However, human operators have several disadvantages compared to smart machines including subjectivity, productivity, consistency, and repeatability. Early defect detection is critical for reducing waste of raw material and achieving a high quality product. A great amount of time is therefore spent during the product development process on the optimization of stamping defects, particularly defect lows during tryout. By implementing such technologies in this thesis, the localized energy analysis with cross-projection is first adopted to capture the stamping region of component. Burr along the stamping region with thickness larger than 10 micrometers is classified with the proposed up and down ramping criterion. Then, the bilateral filter is applied to enhance regions with scratch and contamination. Energy analysis and cross-projection are also combined to detect the scratch and contamination regions by following with localized histogram analysis.
Using samples taken from the local company based on Kaohsiung city, the experiments of stamping defects have been proved that the proposed framework well performed with 96.39 % recognition rate for the burr defects and 100 % recognition rate for the surface defects, respectively. The proposed framework is possessed of the advantage of autonomous detection with high recognition rate. Leveraging the unique integration of stamping on sheet metal with defect inspection has led to the development of new technologies that transform the performance of stamping press.
|
author2 |
WANG, JING-WEIN |
author_facet |
WANG, JING-WEIN TU, HSUN-CHENG 涂勛城 |
author |
TU, HSUN-CHENG 涂勛城 |
spellingShingle |
TU, HSUN-CHENG 涂勛城 Defect detection for stamped part of hard disk shell |
author_sort |
TU, HSUN-CHENG |
title |
Defect detection for stamped part of hard disk shell |
title_short |
Defect detection for stamped part of hard disk shell |
title_full |
Defect detection for stamped part of hard disk shell |
title_fullStr |
Defect detection for stamped part of hard disk shell |
title_full_unstemmed |
Defect detection for stamped part of hard disk shell |
title_sort |
defect detection for stamped part of hard disk shell |
publishDate |
2017 |
url |
http://ndltd.ncl.edu.tw/handle/4xe9np |
work_keys_str_mv |
AT tuhsuncheng defectdetectionforstampedpartofharddiskshell AT túxūnchéng defectdetectionforstampedpartofharddiskshell AT tuhsuncheng yìngdiékéchōngyājiànxiácījiǎncèxìtǒng AT túxūnchéng yìngdiékéchōngyājiànxiácījiǎncèxìtǒng |
_version_ |
1719149381168398336 |