• 한국멀티미디어학회논문지
• /
• 제12권3호
• /
• pp.397-408
• /
• 2009

영상 학습은 컴퓨터를 이용한 자동 시각 검사에서 매우 중요하고 어려운 문제이다. 최근 생산되는 인쇄회로 기판(PCB : Printed Circuit Board)은 부품의 크기가 작아지고 회로 패턴이 점점 복잡해져서 신제품의 개발 주기가 짧고 다양한 제품들을 검사해야 하는 분야에서 어렵고 복잡한 학습 과정은 큰 문제가 되고 있다. 본 논문은 CAD(Gerber: 거버)파일을 이용하여 PCB 자동 시각 검사의 기준이 되는 참조 영상을 생성하였다. Gerber 파일로 생성된 참조 영상은 결함이 없는 PCB 패턴을 보장한다. 시스템의 구현과 실험을 통하여 Gerber 파일을 이용하여 PCB 자동 시각 검사 시스템의 학습 과정을 손쉽게 할 수 있는 방안을 제시하였다.

• 산업기술연구
• /
• 제21권B호
• /
• pp.67-74
• /
• 2001

According as the patterns of PCB (Printed Circuit Board) become denser and complicated, quality and accuracy of PCB influence the performance of final product. It's attempted to obtain trust of 100% about all of parts. Because human inspection in mass-production manufacturing facilities are both time-consuming and very expensive, the automation of visual inspection has been attempted for many years. Thus, automatic visual inspection of PCB is required. In this paper, we used an algorithm which compares the reference PCB patterns and the input PCB patterns are separated an object and a scene by filtering and edge detection. And than compare two image using pattern matching algorithm. We suggest an defect inspection algorithm in PCB pattern, to be satisfied low cost, high speed, high performance and flexibility on the basis of $640{\times}480$ binary pattern.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2002년도 하계종합학술대회 논문집(4)
• /
• pp.81-84
• /
• 2002

This paper presents the real-time PCB(Printed circuit board) inspection system that can detect thin open/short error using the line scan camera. After a overall introduction of our system, the outline of our inspection methods are described. The goal of our inspection system is the real time and detailed inspection using the line scan camera. To perform inspection processing in real-time, we utilize double buffering structure. In order to solve the problem of unexpectable pixels of PCB, we propose melting process which eliminates unexpectable pixels of PCB. The design and development of our prototype of PCB ins- pection system is discussed and test results are presented to show the effectiveness of the developed inspection algorithm.

• 한국정밀공학회지
• /
• 제23권3호
• /
• pp.102-109
• /
• 2006

Several methods for PCB pattern inspection have been tried to detect fine detects in pad contours, but their low detection accuracy results from pattern variations originating from etching, printing and handling processes. The adaptive inspection algorithm has been newly proposed to extract minute defects based on movable segments. With gerber master images of PCB, vertex extractions of a pad boundary are made and then a lot of segments are constructed in master data. The pad boundary is composed of segment units. The proposed method moves these segments to optimal directions of a pad boundary and so adaptively matches segments to pad contours of inspected images, irrespectively of various pattern variations. It makes a fast, accurate and reliable inspection of PCB patterns. Its performances are also evaluated with several images.

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 2003년도 하계종합학술대회 논문집 Ⅳ
• /
• pp.2335-2338
• /
• 2003

This paper presents a real-time PCB (Printed Circuit Board) vision inspection system. This system can detect the OPEN and SHORT of the PCB which of the line width is 150$\mu\textrm{m}$. Our PCB inspection system is based on the referential method. Since the size of the captured PCB image is very large, the image is divided into 512${\times}$512 images to apply the accurate alignment efficiently. To correct the misalignment between the reference image and the inspection image, pattern matching is performed. In order to implement the proposed algorithm in real-time, we use the SIMD instruction and the double buffering structures. Our experiential results show the effectiveness of the developed inspection algorithm.

• 한국광학회:학술대회논문집
• /
• 한국광학회 2006년도 하계학술발표회 논문집
• /
• pp.469-471
• /
• 2006

This paper describes a efficient insepction method of PCB PADs for AFVI. The methods for PCB inspection have been tried to detect the defects in PCB PADs, but their low detection rate results from pattern variations that are originating from etching, printing and handling processes. The adaptive inspection method has been newly proposed to extract minute defects based on dynamic segments and filters. The vertexes are extracted from CAM master images of PCB and then a lot of segments are constructed in master data. The proposed method moves these segments to optimal directions of a PAD contour and so adaptively matches segments to PAD contours of inspected images, irrespectively of various pattern variations. It makes a fast, accurate and reliable inspection of PCB patterns. Experimental results show that proposed methods are found to be effective for flexible defects detection.

• 반도체디스플레이기술학회지
• /
• 제19권2호
• /
• pp.11-15
• /
• 2020

In a typical initial setup of a PCB component inspection system, operators should manually input various information such as category, position, and inspection area for each component to be inspected, thus causing much inconvenience and longer setup time. Although there are many deep learning based object detectors, RetinaNet is regarded as one of best object detectors currently available. In this paper, a method using an extended RetinaNet is proposed that automatically detects its component category and position for each component mounted on PCBs from a high-resolution color input image. We extended the basic RetinaNet feature pyramid network by adding a feature pyramid layer having higher spatial resolution to the basic feature pyramid. It was demonstrated by experiments that the extended RetinaNet can detect successfully very small components that could be missed by the basic RetinaNet. Using the proposed method could enable automatic generation of inspection areas, thus considerably reducing the setup time of PCB component inspection systems.

• 제어로봇시스템학회논문지
• /
• 제15권8호
• /
• pp.772-777
• /
• 2009

The geometric distortion of image is one of the most important parameters that take effect on the accuracy of optical inspection systems. We propose a new correction method of the image distortion to increase the accuracy of PCB inspection systems. The model-free method is applied to correct the randomly distorted image that cannot be represented by mathematical model. To reduce the correction time of inspection system, we newly propose a grid reduction algorithm that minimize the number of grids by the quad-tree approach. We apply the proposed method to a PCB inspection system, and verify its usefulness through experiments using actual inspection images.

• 한국정밀공학회지
• /
• 제15권5호
• /
• pp.9-18
• /
• 1998

The components with J-lead. which are more integrated and smaller than ones with Gull-wing. are rapidly being used in electronic board such as the PCB, for they have the advantage of occupying a small space compared to the other components. However, the development of inspection system for these new components is not so rapid as component development. Component-inspection with J-lead using vision system is difficult because they are hidden from camera optical axis. X-ray inspection, which has the advantage of inspecting the inside of solder state, is used to J-lead inspection. However. it is high cost and dangerous by leaking out X-ray compared to vision system. Therefore, in this paper, we design vision system suited to J-lead inspection and then propose algorithm which have flexibility in mount and rand error.

• 한국멀티미디어학회논문지
• /
• 제24권8호
• /
• pp.988-999
• /
• 2021

AOI (Automatic Optical Inspection) of PCB (Printed Circuit Board) is a very important step to guarantee the product performance. The process of registering components called teaching mode is first perform, and AOI is then carried out in a testing mode that checks defects, such as recognizing and comparing the component mounted on the PCB to the stored components. Since most of registration of the components on the PCB is done manually, it takes a lot of time and there are many problems caused by mistakes or misjudgement. In this paper, A components classifier is proposed using YOLO (You Only Look Once) v2's object detection model that can automatically register components in teaching modes to reduce dramatically time and mistakes. The network of YOLO is modified to classify small objects, and the number of anchor boxes was increased from 9 to 15 to classify various types and sizes. Experimental results show that the proposed method has a good performance with 99.86% accuracy.