• Proceedings of the Korean Society of Computer Information Conference
• /
• 2018.07a
• /
• pp.325-326
• /
• 2018

본 논문에서는 PCB 공정상의 육안검사를 통한 불량 분류 방식에서 CNN을 이용한 PCB 불량 분류 방식을 제안한다. 이 방식은 육안검사의 문제점인 작업자의 숙련도에 따른 검사 효율을 자동화 검사 시스템에 의해 해결하며, 불량 위치와 종류를 결과 이미지에 표시한다. 또한 이미지 분류 결과를 모니터링할 수 있도록 시리얼 통신을 통하여 Darknet 프레임워크와 LCD를 연동하였다. 적은 량의 데이터 셋으로도 좋은 결과를 냈으며, 다양한 데이터 셋을 이용해 훈련할 시 전반적인 PCB 불량의 분류가 가능할 것으로 예상된다.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2020.07a
• /
• pp.327-328
• /
• 2020

본 논문에서는 저전력 저비용 임베디드 환경에서 PCB 검사 기법을 제안한다. 특히, IC 미삽에 대한 검출 알고리즘을 제안하고 실험한다. 고사양의 컴퓨팅 시스템에서는 CNN과 같은 딥러닝 뉴럴 네트워크를 사용하여 특별한 알고리즘을 고려하지 않아도 대규모의 데이터를 입력함으로써 모델을 완성하고 이를 통해 PCB 검사를 수행할 수 있다. 그러나 데이터의 양이 충분하지 않거나 충분한 전력과 비용을 투입하지 못하는 임베디드 환경에서는 각 부품에 따른 컴퓨터 비전 알고리즘이 필요하다. IC의 경우 타부품에 비하여 형태가 직사각으로 정형화 되있으며 색상도 균일한 특징을 가지고 있기에 미삽에 대한 검출이 가능하다. 베어보드(Bare Board)의 색상과 IC 부품의 색상이 확연히 다를 경우에는 RGB 픽셀을 카운트 하는 히스토그램 카운팅 알고리즘만으로 검출이 가능하다. 베어보드의 색삭과 IC의 색상이 유사할 경우에는 베어보드의 핀 혹은 홀의 형태를 감지하여 검출이 가능하다. 본 논문에서는 베어보드의 색상와 IC의 색상이 같을 경우에 다를 경우를 나누어 미삽 검사를 수행하고 그 정확도를 확인한다.

• International Journal of Control, Automation, and Systems
• /
• v.4 no.1
• /
• pp.96-104
• /
• 2006

We propose a path planning method to improve the productivity of AOI (automated optical inspection) machines in PCB (printed circuit board) assembly lines. The path-planning problem is the optimization problem of finding inspection clusters and the visiting sequence of cameras to minimize the overall working time. A unified method is newly proposed to determine the inspection clusters and visiting sequence simultaneously. We apply a hybrid genetic algorithm to solve the highly complicated optimization problem. Comparative simulation results are presented to verify the usefulness of the proposed method.

• Journal of the Semiconductor & Display Technology
• /
• v.16 no.3
• /
• pp.53-58
• /
• 2017

Recently, advance in technology have increased the importance of visual inspection in semiconductor inspection areas. In PCB visual inspection, accurate line estimation is critical to the accuracy of the entire process, since it is utilized in preprocessing steps such as calibration and alignment. We propose a line estimation method that is differently weighted for the line candidates using a histogram of gradient information, when the position of the initial approximate corner points is known. Using the obtained line equation of the outline, corner points can be calculated accurately. The proposed method is compared with the existing method in terms of the accuracy of the detected corner points. The proposed method accurately detects corner points even when the existing method fails. For high-resolution frames of 3.5mega-pixels, the proposed method is performed in 89.01ms.

• 제어로봇시스템학회:학술대회논문집
• /
• 2001.10a
• /
• pp.109.2-109
• /
• 2001

In the PCB assembly lines, as components become more complex and smaller, the conventional inspection method using traditional ICT and function test show their limitations in application. The automatic optical inspection(AOI) gradually becomes the alternative in the PCB assembly line. In Particular, the PCB inspection machines need more reliable and flexible object recognition algorithms for high inspection accuracy. The conventional AOI machines use the algorithmic approaches such as template matching, Fourier analysis, edge analysis, geometric feature recognition or optical character recognition (OCR), which mostly require much of teaching time and expertise of human operators. To solve this problem, in this paper, a statistical learning based part recognition method is proposed. The performance of the ...

• Journal of the Microelectronics and Packaging Society
• /
• v.22 no.1
• /
• pp.15-19
• /
• 2015

We demonstrated an inspection system for detecting discoloration of PCB Cu ball pad with an OSP surface finish. Though the OSP surface finish has many advantages such as eco-friendly and low cost, however, it often shows a discoloration phenomenon due to a heating process. In this study, the discoloration was analyzed with device-independent CIELAB color space. First of all, the PCB samples were inspected with standard lamps and CCD camera. The measured data was processed with Labview program for detecting discoloration of Cu ball pad. From the original PCB sample image, the localized Cu ball pad image was selected to reduce the image size by the binarization and edge detection processes and it was also converted to device-independent CIELAB color space using $3{\times}3$ conversion matrix. Both acquisition time and false acceptance rate were significantly reduced with this proposed inspection system. In addition, $L^*$ and $b^*$ values of CIELAB color space were suitable for inspection of discoloration of Cu ball pad.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2024.01a
• /
• pp.289-290
• /
• 2024

PCB(Printed circuit board)생산시에 중요한 역할을 담당하는 비전검사 시스템의 성능은 지속적으로 발전해왔다. 기존 머신 비전 검사 시스템은 이미지가 불규칙하고 비정형일 경우 해석이 어렵고 전문가의 경험에 의존한다. 그리고 비전검사 시스템 개발 당시의 기준과 다른 불량이 발생한다면 검출이 불가능 하거나 정확도가 낮게 나온다. 본 논문에서는 이를 개선하고자 딥러닝 영상인식을 이용한 PCB 기판 비전 검사 시스템을 구현하였다. 딥러닝 영상인식 알고리즘은 YOLOv4를 이용하고, 워핑(warping)과 시킨 PCB 이미지를 학습하여 비전검사 시스템을 구성하였다. 딥러닝 영상인식 기술의 처리 속도를 보완하고자 QR코드로 PCB 기판 종류를 인식하고, 해당 PCB 부품의 미삽은 정답 이미지 바운딩 박스 좌표와 비교하여 불량품을 발견하면 표시해준다. 기판의 부품 인식을 위해 기판 데이터는 직접 촬영하여 수집하였다. 이를 활용하여 PCB 생산 공정에서 비전검사 시스템의 성능이 향상되었고,, 다양한 PCB를 생산에 신속하게 대응할 수 있다.

• Journal of Korea Multimedia Society
• /
• v.24 no.7
• /
• pp.868-879
• /
• 2021

PCB (Printed Circuit Board) inspection using a deep learning model requires a large amount of data and storage. When the amount of stored data increases, problems such as learning time and insufficient storage space occur. In this study, the existing object detection model is changed to a continual learning model to enable the recognition and classification of PCB components that are constantly increasing. By changing the structure of the object detection model to a knowledge distillation model, we propose a method that allows knowledge distillation of information on existing classified parts while simultaneously learning information on new components. In classification scenario, the transfer learning model result is 75.9%, and the continual learning model proposed in this study shows 90.7%.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.46 no.2
• /
• pp.22-27
• /
• 2009

In this paper, an automatic functional inspection system EVT (Emulated Vision Tester) for LCD drive module PCB has been proposed and implemented. Typical automatic inspection system such as probing methods and vision-based systems are widely known and used, however, there exist undetectable defects due to critical timing factors which they may miss to catch from LCD equipments. Especially typical vision-based systems have inconsistency on acquisition of images so that distinction between gray scales can be difficult which results in low level of performance and reliability on the inspection results. The proposed EVT system is pure hardware solution. It directly compares pattern signals from a pattern generator to output signals from LCD drive module. It also inspects variety of analog signals such as voltage, resistance, wave forms and so forth. The EVT system not only shows high performance in terms of reliability and processing speed but reduces costs on inspection and maintenance. Also, full automation of entire production line can be realized when EVT is applied in in-line inspection processes.

• The Journal of the Korea Contents Association
• /
• v.6 no.6
• /
• pp.1-8
• /
• 2006

This paper presents an intelligent pattern matching method that can be used to acquire the reliable calibration data for automatic PCB pattern inspection. The inaccurate calibration data is often acquired by geometric pattern variations and selecting an inappropriate model manual. It makes low the confidence of inspection and also the inspection processing time has been delayed. In this paper, the geometric features of PCB patterns are utilized to calculate the accurate calibration data. An appropriate model is selected automatically based on the geometric features, and then the calibration data to be invariant to the geometric variations(translation, rotation, scaling) is calculated. The method can save the inspection time unnecessary by eliminating the need for manual model selection. As the result, it makes a fast, accurate and reliable inspection of PCB patterns.