• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2018.06a
• /
• pp.235-237
• /
• 2018

많은 분야에서 기기설비들의 고장, 결함은 안전과 관련되어 있기 때문에 연구가 활발히 진행되고 있다. 주로 데이터를 취득하여 제품의 유지보수 및 품질을 향상시키는 연구로 고장을 나타내는 특성 인자를 추출하여 고장진단을 하는 것이다. 하지만, 과거의 룰 기반 결함 탐지 기법은 예외의 경우를 탐지하기 어렵다는 문제를 가져왔다. 최근 들어 인공지능이 특성 인자를 쉽게 추출할 수 있다는 장점으로 인해 인공지능과 결합된 고장진단 시스템이 많이 제안되고 있다. 본 논문에서는 인공지능의 추세와 인공지능과 결합된 고장진단 시스템을 소개한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2009.04a
• /
• pp.436-439
• /
• 2009

본 연구에서는 소프트웨어 제품의 품질 수준을 파악할 수 있는 지표를 도출하여 지표산식을 정의하고 지표의 결과를 산출하기 위해 필요한 수집항목을 선정하며 수집과 분석을 통해 실질적으로 디지털 콘텐츠 소프트웨어의 어떤 결함 유형들이 주로 발생하고 있으며 그에 따른 결함수정 비용의 산출을 통해 소프트웨어의 품질확보를 위해 필요한 비용이 어느 정도가 될 것인지를 객관적 자료를 근거로 나타내고자 한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2011.11a
• /
• pp.223-226
• /
• 2011

반도체 공정에서 부품의 결함을 찾는 것은 완제품의 품질 개선을 위해 중요하다. 현재까지 많은 비전 알고리즘들이 부품의 결함을 찾기 위해 적용되고 있다. 그러나 이런 알고리즘 대부분은 2D 방식의 검사 방식에 머물고 있다. 그러나 이런 2D방식의 검사 방법은 반도체 칩의 Lead나 Pad 그리고 Solder Joint와 같이 3D 정보에 의해 불량 유무를 판결해야 하는 곳에 적용하기 어렵다. 이에 본 논문에서는 PMP(Phase Measuring Profilometry)방법에 의해 반도체 칩의 Lead부분을 검사하기 위한 시스템 구성과 방법을 제안한다.

• Archives of design research
• /
• v.14 no.4
• /
• pp.137-146
• /
• 2001

The purpose of the study is to explore and understand the product development through the theory of evolution. For the purpose of the study, an evolution mechanism was set up in which the products were applied with. The results showed three distinguishable types of product evolution 'the evolution of functions','the evolution of forms'and 'the evolution of symbols what the products represent'. In order to answer the research questions, , , , the research was carried out in three ways: firstly, some ideas of product evolution were looked closely through existing researches; secondly, the study explored the idea of evolution mechanism being a natural process like an organic system; finally by applying various products with the evolution mechanism, different patterns of product evolution were classified. The results of the study can be summarised as follows: ${\bullet}$ The evolution mechanism in this study can be defined as an 'organic system'that consists of the causes of evolution'(based on the needs and the willingness to have pleasure), 'the point of evolving'(based on a degree of imperfection) and 'the motive of evolution'(based on the readiness to make profit). ${\bullet}$ The evolution mechanism seems to suggest three different patterns in product evolution,'the evolution of functions', 'the evolution of forms' and 'the evolution of symbols'

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2010.05a
• /
• pp.569-571
• /
• 2010

In this paper, we develop a vision inspection system for inspecting flaws on plastic products such as insufficient moldings, spots, scratches. The inspection algorithm for this system consist of image binarization for curved structure of plastic products, image noise removal using morphology operation, labeling methods for candidate regions and image filtering and calibration method for flaw inspection. In order to improve its performance, we also develop fast image processing algorithm based on GUI. To verify the effectiveness of this system, we conducted evaluation for the system accuracy and the inspection algorithm processing time.

• 대한공업교육학회지
• /
• v.33 no.2
• /
• pp.312-324
• /
• 2008

To improve the newly designed HKS(hinge kit system) in common refrigerators, it was investigated the new robust methodologies. There were the study of failure modes, mechanisms in the marketplace, and the design parameters of HKS with various improvements using accelerated life testing. Based on the claimed marketplace product returns and 1st ALT reproduction, the fracturing and cracking occur in the housing of the HKS. The missing design parameters of the failed HKS in the design phase of the refrigerator was the housing hinge kit structure. The corrective action plans are the modifications of the housing hinge kit structure from the open supporting to all supporting structure. Based on 2nd ALTs, the fracturing and cracking occur in the torsion shaft. The missing design parameter was the roundness of torsion shaft. After a sequence of ALT testing, the levels of the missing design parameters were setup. The yearly failure rate and B1 life of the redesigned HKS, based on the results of ALT, were over 0.01 percent and 10 years, respectively. The parameter design through the inspection of the failed product, load analysis, and three rounds of ALT, was very effective in the new robust design methodologies of the mechanical system and this method can be applied to other design system.

• The Journal of Bigdata
• /
• v.7 no.2
• /
• pp.31-41
• /
• 2022

Steel is one of the most fundamental components to mechanical industry. However, the quality of products are greatly impacted by the surface defects in the steel. Thus, researchers pay attention to the need for surface defects detector and the deep learning methods are the current trend of object detector. There are still limitations and rooms for improvements, for example, related works focus on developing the models but don't take into account real-time application with practical implication on industrial settings. In this paper, a real-time application of steel surface defects detection based on YOLOv4 is proposed. Firstly, as the aim of this work to deploying model on real-time application, we studied related works on this field, particularly focusing on one-stage detector and YOLO algorithm, which is one of the most famous algorithm for real-time object detectors. Secondly, using pre-trained Yolov4-Darknet platform models and transfer learning, we trained and test on the hot rolled steel defects open-source dataset NEU-DET. In our study, we applied our application with 4 types of typical defects of a steel surface, namely patches, pitted surface, inclusion and scratches. Thirdly, we evaluated YOLOv4 trained model real-time performance to deploying our system with accuracy of 87.1 % mAP@0.5 and over 60 fps with GPU processing.

• Journal of Digital Convergence
• /
• v.11 no.2
• /
• pp.115-125
• /
• 2013

Unexpected consumer dissatisfaction emerges through rapid growth and expansion of on-line shopping channel. This research focused on the fashion online retail channels' negative aspect caused by service failure which possibly disappointed consumers. We also tried to seek for appropriate service recovery types based on frequently offered recovery types on-line. Data from college students were analyzed. Results indicate that fitting problem, insufficient information, product defect, inventory problem and slow delivery were the main service failure types in apparel e-shopping. Regression analysis identified that among these types, insufficient information, product defect, and slow delivery had significant influence on channel satisfaction after post recovery effort. Results also confirmed significant relationships between channel satisfaction and channel switching. Consumers perceived benefit level causes overall channel satisfaction level to rise while perceived risk leads to lower level of channel satisfaction. Choosing desirable service recovery activities in each service failure situations is necessary in order to raise consumer's channel satisfaction in online apparel shopping.

• Journal of the KSME
• /
• v.32 no.8
• /
• pp.713-724
• /
• 1992

연속주조공정에서 주형과 2차냉각 영역에서의 주편의 응고과정에 관한 열전달 문제를 다루는데 해석적 및 수치적방법의 기본 원리와 그 결과를 간단히 소개하였다. 연속주조공정중 상황변화가 일어나는 주형과 2차냉각대에서 온도제어를 효율적으로 수행하기 위하여 주조온도, 강종, 주조 속도, 냉각수량 등 운전 인자에 따른 주편온도분포 및 응고층두께를 이론적으로 구하는 방법만을 취급하였지만, 실제로는 주편온도 및 응고층 두께를 측정하는 실험적 방법이 상호보완적으로 이루어져야 향후제품결함이 적고, 높은 생산성을 꾀할 수 있는 신 연주기의 개발이 가능할 것이 다.

• 한국가스학회:학술대회논문집
• /
• 2001.10a
• /
• pp.31-34
• /
• 2001

점차 가시화되고 있는 이루크츠크 가시도입 사업과 같은 장거리 가스배관에 사용될 배관은 현재 국내 가스배관과는 다른 종류가 될 가능성이 매우 높다. 이에 API X70 이상의 고강도 강재 개발부터 배관건설현장 용접 결함평가기준까지 다양한 분야의 기술개발이 필요하다. 이러한 기술개발은 어느 한 기관이 아닌 여러 관련업계가 함께 연구 및 제품개발에 참여하는 대형 과제로서 추진하는 것이 효과적이다.