• 제목/요약/키워드: Scale defect

검색결과 237건 처리시간 0.028초

표면 결함 검출을 위한 데이터 확장 및 성능분석 (Performance Analysis of Data Augmentation for Surface Defects Detection)

  • 김준봉;서기성
    • 전기학회논문지
    • /
    • 제67권5호
    • /
    • pp.669-674
    • /
    • 2018
  • Data augmentation is an efficient way to reduce overfitting on models and to improve a performance supplementing extra data for training. It is more important in deep learning based industrial machine vision. Because deep learning requires huge scale of learning data to learn a model, but acquisition of data can be limited in most of industrial applications. A very generic method for augmenting image data is to perform geometric transformations, such as cropping, rotating, translating and adjusting brightness of the image. The effectiveness of data augmentation in image classification has been reported, but it is rare in defect inspections. We explore and compare various basic augmenting operations for the metal surface defects. The experiments were executed for various types of defects and different CNN networks and analysed for performance improvements by the data augmentations.

NGC 영상시스템 기반의 패턴 결함검출기 설계 (Design of Embedded Iamge System based Pattern Defect Detector)

  • 이동원;엄예지;강민구;조문신;이문용
    • 한국콘텐츠학회:학술대회논문집
    • /
    • 한국콘텐츠학회 2007년도 추계 종합학술대회 논문집
    • /
    • pp.869-873
    • /
    • 2007
  • 본 논문은 고속으로 생산되는 제품의 영상을 캡쳐한 후 영상처리 기법 중 에지 추출 알고리즘을 응용하여 조명에 투과된 제품의 에지를 추출 및 필터링 하는 방법으로 결함 검출 시스템을 설계한다. 소형의 임베디드 기기에 패턴 매칭 영상처리 기법을 이용하여 결함을 검출하고 패턴의 비 매칭 정도를 기준점에 따라 정상 또는 불량 판정을 할 수 있는 어플리케이션을 개발하고 탑재하였고, 어플리케이션의 불량 판정 알고리즘으로는 NGC (Normalized GrayScale Corelation) 기법을 사용하였고 검출 판정 결과 적절한 판정값을 입력하는 것으로 기준 패턴과 형상이 다른 대상의 불량을 판정한다.

  • PDF

칼라와 공간 정보를 이용한 평균 이동에 기반한 물체 추적 (Mean Shift Based Object Tracking with Color and Spatial Information)

  • 안광호;정명진
    • 대한전기학회:학술대회논문집
    • /
    • 대한전기학회 2006년도 제37회 하계학술대회 논문집 D
    • /
    • pp.1973-1974
    • /
    • 2006
  • The mean shift algorithm has achieved considerable success in object tracking due to its simplicity and robustness. It finds local maxima of a similarity measure between the color histograms of the target and candidate image. However, the mean shift tracking algorithm using only color histograms has a serious defect. It doesn't use the spatial information of the target. Thus, it is difficult to model the target more exactly. And it is likely to lose the target during the occlusions of other objects which have similar color distributions. To deal with these difficulties we use both color information and spatial information of the target. Our proposed algorithm is robust to occlusions and scale changes in front of dynamic, unstructured background. In addition, our proposed method is computationally efficient. Therefore, it can be executed in real-time.

  • PDF

Void Formation Induced by the Divergence of the Diffusive Ionic Fluxes in Metal Oxides Under Chemical Potential Gradients

  • Maruyama, Toshio;Ueda, Mitsutoshi
    • 한국세라믹학회지
    • /
    • 제47권1호
    • /
    • pp.8-18
    • /
    • 2010
  • When metal oxides are exposed to chemical potential gradients, ions are driven to diffusive mass transport. During this transport process, the divergence of ionic fluxes offers the formation/annihilation of oxides. Therefore, the divergence of ionic flux may play an important role in the void formation in oxides. Kinetic equations were derived for describing chemical potential distribution, ionic fluxes and their divergence in oxides. The divergence was found to be the measure of void formation. Defect chemistry in scales is directly related to the sign of divergence and gives an indication of the void formation behavior. The quantitative estimation on the void formation was successfully applied to a growing magnetite scale in high temperature oxidation of iron at 823 K.

AE 신호 형상 인식법에 의한 회전체의 신호 검출 및 분류 연구 (Detection and Classification of Defect Signals from Rotator by AE Signal Pattern Recognition)

  • 김구영;이강용;김희수;이현
    • 한국철도학회논문집
    • /
    • 제4권3호
    • /
    • pp.79-86
    • /
    • 2001
  • The signal pattern recognition method by acoustic emission signal is applied to detect and classify the defects of a journal bearing in a power plant. AE signals of main defects such as overheating, wear and corrosion are obtained from a small scale model. To detect and classify the defects, AE signal pattern recognition program is developed. As the classification methods, the wavelet transformation analysis, the frequency domain analysis and time domain analysis are used. Among three analyses, the wavelet transformation analysis is most effective to detect and classify the defects of the journal bearing..

  • PDF

유한요소해석에 기초한 감육곡관 손상압력 평가 모델 개발 (Development of Failure Pressure Evaluation Model for Local Wall-Thinned Elbows Based on Finite Element Analysis)

  • 김진원;박종선
    • 대한기계학회논문집A
    • /
    • 제32권12호
    • /
    • pp.1063-1071
    • /
    • 2008
  • This paper provides a failure pressure evaluation model for local wall-thinned elbows. In this study, parametric finite element analyses are performed on the elbows containing local wall-thinning defect at their intrados and extrados, and the failure pressures are obtained from the analysis results by applying a local failure criterion that was validated by real-scale pipe tests. An evaluation model including the effects of thinning depth, length, circumferential angle, thinning location, and elbow geometries on the failure pressure is derived based on the evaluated failure pressures. The proposed model agrees well with the results of finite element analyses and reasonably estimates the dependence of failure pressure on the wall-thinning dimensions and elbow geometries. Also, the comparison with experimental data demonstrates that the proposed evaluation model can accurately predict the failure pressure of local wall-thinned elbows.

Atomic-scale investigation of Epitaxial Graphene Grown on 6H-SiC(0001) Using Scanning Tunneling Microscopy and Spectroscopy

  • 이한길;최정헌;김세훈
    • 한국진공학회:학술대회논문집
    • /
    • 한국진공학회 2012년도 제42회 동계 정기 학술대회 초록집
    • /
    • pp.125-125
    • /
    • 2012
  • Graphene was epitaxially grown on a 6H-SiC(0001) substrate by thermal decomposition of SiC under ultrahigh vacuum conditions. Using scanning tunneling microscopy (STM), we monitored the evolution of the graphene growth as a function of the temperature. We found that the evaporation of Si occurred dominantly from the corner of the step rather than on the terrace. A carbon-rich $(6{\sqrt{3}}{\times}6{\sqrt{3}})R30^{\circ}$ layer, monolayer graphene, and bilayer graphene were identified by measuring the roughness, step height, and atomic structures. Defect structures such as nanotubes and scattering defects on the monolayer graphene are also discussed. Furthermore, we confirmed that the Dirac points (ED) of the monolayer and bilayer graphene were clearly resolved by scanning tunneling spectroscopy (STS).

  • PDF

AN ASSESSMENT OF THE RADIATION DOSE RATE DUE TO AN OCCURRENCE OF THE DEFECT ON THE SPENT NUCLEAR FUEL ROD

  • Lee, Sang-Hun;Moon, Joo-Hyun
    • Journal of Radiation Protection and Research
    • /
    • 제34권3호
    • /
    • pp.144-150
    • /
    • 2009
  • This study examines how much the radiation dose rate around it varies if a crack occurs on the spent nuclear fuel rod. The spent nuclear fuel rod to be examined is that of Kori unit 3&4. The source terms are evaluated using the ORIGEN-ARP that is part of the version 5.1 of the SCALE package. The radiation dose rate is assessed using the TORT. To check if the structure of a fuel rod is appropriately modeled in the TORT calculation, the calculation results by the TORT are compared with those by the ANISN for the same case. From the code simulation, it is known that if a crack occurs on the spent nuclear fuel rod, the neutron dose rate varies depending on what material is the crack filled with, but the gamma dose rate varies irrespective of type of the material that the crack is filled with.

고속월쉬변환을 이용한 비선형 시스템의 3계층 최적제어 (Three-Level Optimal Control of Nonlinear Systems Using Fast Walsh Transform)

  • 김태훈;신승권;조영호;이한석;이재춘;안두수
    • 대한전기학회논문지:시스템및제어부문D
    • /
    • 제50권11호
    • /
    • pp.505-513
    • /
    • 2001
  • This paper presents the new three-level optimal control scheme for the large scale nonlinear systems, which is based on fast walsh transform. It is well known that optimization for nonlinear systems leads to the resolution of a nonlinear two point boundary value problem which always requires a numerical iterative technique for their solution. However, Three-level costate coordination can avoid two point boundary condition in subsystem. But this method also has the defect that must solve high order differential equation in intermediate level. The proposed method makes use of fast walsh transform, therefore, is simple in computation because of solving algebra equation instead of differential equation.

  • PDF

전력계통의 전력조류제어를 위한 진화연산의 비교 (Comparison of Evolutionary Computation for Power Flow Control in Power Systems)

  • 이상근
    • 전기학회논문지P
    • /
    • 제54권2호
    • /
    • pp.61-66
    • /
    • 2005
  • This paper presents an unified method which solves real and reactive power dispatch problems for the economic operation of power systems using evolutionary computation such as genetic algorithms(GA), evolutionary programming(EP), and evolution strategy(ES). Many conventional methods to this problem have been proposed in the past, but most of these approaches have the common defect of being caught to a local minimum solution. The proposed methods, applied to the IEEE 30-bus system, were run for 10 other exogenous parameters and composed of P-optimization module and Q-optimization module. Each simulation result, by which evolutionary computations are compared and analyzed, shows the possibility of applications of evolutionary computation to large scale power systems.