• Resources Recycling
• /
• v.30 no.4
• /
• pp.3-10
• /
• 2021

The field of recycling for waste electronic components, which is the typical example of an urban mine, requires the development of useful sorting techniques. In this study, a sorter based on image identification by deep learning was developed to select electronic components into four groups. They were recovered from waste printed circuit boards and should be separated to depend on the difference after treatment. The sorter consists of a workstation with GPU, camera, belt conveyor, air compressor. A small piece (less than 3.5 cm) of electronic components on the belt conveyor (belt speed: 6 cm/s) was taken and learned as teaching data. The accuracy of the image identification was 96% as kinds and 99% as groups. The optimum condition of sorting was determined by evaluating accuracies of image identification and recovery rates by blowdown when changing the operating condition such as belt speed and blowdown time of compressed air. Under the optimum condition, the accuracy of image classification in groups was 98.7%. The sorting rate was more than 70%.

• Journal of the Korea Society of Computer and Information
• /
• v.27 no.9
• /
• pp.13-20
• /
• 2022

It's proposed and analyzed ML(Machine Learning) models to predict vehicle FC(Fuel Consumption) in real-time. The test driving was done for a car to measure vehicle speed, acceleration, road gradient and FC for training dataset. The various ML models were trained with feature data of speed, acceleration and road-gradient for target FC. There are two kind of ML models and one is regression type of linear regression and k-nearest neighbors regression and the other is classification type of k-nearest neighbors classifier, logistic regression, decision tree, random forest and gradient boosting in the study. The prediction accuracy is low in range of 0.5 ~ 0.6 for real-time FC and the classification type is more accurate than the regression ones. The prediction error for total FC has very low value of about 0.2 ~ 2.0% and regression models are more accurate than classification ones. It's for the coefficient of determination (R²) of accuracy score distributing predicted values along mean of targets as the coefficient decreases. Therefore regression models are good for total FC and classification ones are proper for real-time FC prediction.

• Journal of the Korean Geotechnical Society
• /
• v.31 no.9
• /
• pp.17-27
• /
• 2015

We perform a series of experimental tests to evaluate whether the shear strength of clean sands can be reliably predicted from shear wave velocity. Isotropic drained triaxial tests on clean sands reconstituted at different relative densities are performed to measure the shear strength and bender elements are used to measure the shear wave velocity. Laboratory tests reveal that a correlation between shear wave velocity, void ratio, and confining pressure can be made. The correlation can be used to determine the void ratio from measured shear wave velocity, from which the shear strength is predicted. We also show that a unique relationship exists between maximum shear modulus and effective axial stress at failure. The accuracy of the equation can be enhanced by including the normalized confining pressure in the equation. Comparisons between measured and predicted effective friction angle demonstrate that the proposed equation can accurately predict the internal friction angle of granular soils, accounting for the effect of the relative density, from shear wave velocity.

• Korean Journal of Optics and Photonics
• /
• v.19 no.1
• /
• pp.1-8
• /
• 2008

Shack-Hartmann sensors are widely employed as a wavefront measuring device in various applications. Adaptive optics is one of the major applications. Since an adaptive optics system should be operated in real-time, high-speed wavefront sensing is essential. In high-speed operation, integration time of an image detector is very short. In this case, noises such as readout noise and photon noise greatly influence the accuracy of wavefront sensing. Therefore a fast and noise-insensitive centroid finding algorithm is required for the real-time wavefront sensing. In this paper, the multi-resolution correlation method is proposed. By employing multi-resolution images, this method greatly reduces the computation time when compared to the fast Fourier transform (FFT) correlation method. The verification is performed through the computational simulation. In this paper, the center of mass method, correlation method and multi-resolution correlation method are employed to compare the measurement accuracy of the centroid finding algorithms. The accuracy of a Shack-Hartmann wavefront sensor using the proposed algorithm is proved to be comparable to that of the conventional correlation method.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.01a
• /
• pp.91-92
• /
• 2023

본 논문에서는 Yolov5와 Deepsort를 사용한 Tracking by detection을 구현하여 특정 영역을 통과하는 차량의 수를 집계하고, 각 차량의 추정속도를 계산하는 시스템을 구현한다. 실시간 객체 탐지 기능을 수행하는 Yolov5 모델의 학습에는 Kaggle의 개방 데이터인 '도요타 자동차 이미지'를 사용한다. 이미지 크기 640*640, 배치사이즈 16, Early stopping 플래그를 사용하여 학습했을때, Yolov5의 객체 탐지 성능은 정확도 98%, 정밀도 0.961, mAP 0.72을 보여주었다.

• Geophysics and Geophysical Exploration
• /
• v.9 no.4
• /
• pp.304-315
• /
• 2006

We have examined the applicability of f-k analysis to the GPR direct wave measurement for water content to characterize vadose zone condition. When the vadose zone consists of a dry surface layer over wet substratum, we obtained f-k spectra where most of the energy is bounded by the air and dry soil velocities. In this case, dry soil velocity was successfully estimated by using high frequency data. On the other hands, when wet soil overlies dry substratum, the f-k spectra show a contrasting response where most of the energy travels with the velocity bounded by dry and wet soil velocities. In this case, the radar waves are trapped and guided within wet soil layer, exhibiting velocity dispersion. By adopting modal propagation theory, we could formulae a simple inversion code to find two layer's dielectric constants as well as layer thickness. By inverting the velocity dispersion curve obtained from f-k spectra of synthetic modeling data, we could obtain good estimates of dielectric constants of each layer as well as first layer thickness. Moreover, we could obtain more accurate results by including the higher mode data. We expect this method will be useful to get the quantitative property of real subsurface when the field condition is similar.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.39 no.8
• /
• pp.766-773
• /
• 2011

The inclination attitude of the Ground-Based Radar can be measured by the accelerometer due to its static operation environment, but the measurement error is generated from the angular acceleration of the accelerometer, which is created in mechanical oscillation by the dynamic environment, like the wind, gust, rotating antenna, etc. In this paper, the technique of reducing the measurement error of the attitude by the dynamic attitude is proposed and the result of the simulation and the analysis of tracking error by the attitude error are presented.

• Proceedings of the Korean Information Science Society Conference
• /
• 2010.06b
• /
• pp.397-400
• /
• 2010

최근의 과학 실험은 그 규모나 내용에 있어서 점차 대형화되는 동시에 복잡해지고 있다. 이로 인하여 다양한 측정 장비로부터 도출된 실험 결과를 효율적으로 분석, 관리, 종합하는 도구의 필요성이 커지고 있다. 본 논문에서는 원격 측정 장치로부터 서로 다른 포맷의 실험 데이터를 자동 수집한 후 이중 정제한 데이터들만 추출하여 웹에서 시각화하는 실험 데이터 통합 관리 시스템을 제안한다. 먼저 원격 측정 장치의 데이터를 자동으로 수집하기 위해 폴링 서버를 설계하여 장치마다 폴링 에이전트를 도입하였다. 이를 통해 관리자가 각 측정 장치에 직접 접근하지 않고도 데이터를 수집할 수 있다. 폴링으로 확보한 데이터는 파싱을 통해 정제하고, 이들 데이터로 데이터베이스를 구축한다. 정제한 데이터는 시각화하여 사용자가 웹에서 쉽게 파악할 수 있다. 데이터 폴링은 TCP/IP Socket을 통해 수행하므로 보편적으로 사용하는 FTP 방식에 비해 데이터 확보 시 신뢰성을 높일 수 있으며, 폴링 여부 판단 시에는 동기식, 실제 폴링 시에는 비동기식 통신 방법을 사용하여 폴링의 효율을 높였다. 본 시스템을 활용하여 사용자의 임의적인 데이터 접근을 최소화하였고 데이터의 전송, 저장, 관리를 자동화함으로써 편의성을 높였다. 본 시스템을 활용하여 원격 실험 장치로부터 데이터를 확보할 때의 정확성과 폴링 및 파싱 속도를 실험을 통해 측정하였고, 그 결과 폴링 시 100%의 정확도와 정상 포맷의 데이터에 대해서 100%의 파싱 결과를 보임으로써 본 시스템이 원격 장치의 실험 데이터를 통합 관리할 때 적합함을 알 수 있었다. 추후 데이터의 속성에 따라 클러스터링 할 예정이며 클러스터링에 따른 시각화 서비스를 제공할 계획이다.

• KIISE Transactions on Computing Practices
• /
• v.23 no.5
• /
• pp.299-303
• /
• 2017

By piling up hidden layers in artificial neural networks, deep learning is delivering outstanding performances for high-level abstraction problems such as object/speech recognition and natural language processing. Alternatively, deep-learning users often struggle with the tremendous amounts of time and resources that are required to train deep neural networks. To alleviate this computational challenge, many approaches have been proposed in a diversity of areas. In this work, two of the existing Apache Spark-based acceleration frameworks for deep learning (SparkNet and DeepSpark) are compared and analyzed in terms of the training accuracy and the time demands. In the authors' experiments with the CIFAR-10 and CIFAR-100 benchmark datasets, SparkNet showed a more stable convergence behavior than DeepSpark; but in terms of the training accuracy, DeepSpark delivered a higher classification accuracy of approximately 15%. For some of the cases, DeepSpark also outperformed the sequential implementation running on a single machine in terms of both the accuracy and the running time.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.21 no.3
• /
• pp.281-287
• /
• 2001

For the conventional two-dimensional source location of acoustic emission (AE) based on the threshold crossing, wave velocity has to be measured in the actual structure to calculate the arrival-time difference and thus to form the two hyperbolae. Velocity is dependent on the fiber orientation, however, due to the dependence of elastic modulus on fiber orientation in anisotropic materials such as compost#e plates. This tan affect the accuracy of AE source location and make the source location procedure complicated. In this study, we propose a method to reduce the location error in anisotropic plates by using the numerical solution of nonlinear equations, where the velocity term has been removed by employing the fourth sensor. The efficiency and validity of the proposed method has also been experimentally verified.