• 국제학술발표논문집
• /
• The 9th International Conference on Construction Engineering and Project Management
• /
• pp.814-822
• /
• 2022

The deployment of sensors for Structural Health Monitoring requires a complicated network arrangement, ground truthing, and calibration for validating sensor performance periodically. Any conventional sensor on a structural element is also subjected to static and dynamic vertical loadings in conjunction with other environmental factors, such as brightness, noise, temperature, and humidity. A structural model with strain gauges was built and tested to get realistic sensory information. This paper investigates different deep learning architectures and algorithms, including unsupervised, autoencoder, and supervised methods, to benchmark blind drift calibration methods using deep learning. It involves a fully connected neural network (FCNN), a long short-term memory (LSTM), and a gated recurrent unit (GRU) to address the blind drift calibration problem (i.e., performing calibrations of installed sensors when ground truth is not available). The results show that the supervised methods perform much better than unsupervised methods, such as an autoencoder, when ground truths are available. Furthermore, taking advantage of time-series information, the GRU model generates the most precise predictions to remove the drift overall.

• ETRI Journal
• /
• 제36권6호
• /
• pp.894-904
• /
• 2014

Blind mismatch correction of time-interleaved analog-to-digital converters (TI-ADC) is a challenging task. We present a practical blind calibration technique for low-computation, low-complexity, and high-resolution applications. Its key features are: dramatically reduced computation; simple hardware; guaranteed parameter convergence with an arbitrary number of TI-ADC channels and most real-life input signals, with no bandwidth limitation; multiple Nyquist zone operation; and mixed-domain error correction. The proposed technique is experimentally verified by an M = 4 400 MSPS TI-ADC system. In a single-tone test, the proposed practical blind calibration technique suppressed mismatch spurs by 70 dB to 90 dB below the signal tone across the first two Nyquist zones (10 MHz to 390 MHz). A wideband signal test also confirms the proposed technique.

• Journal of Astronomy and Space Sciences
• /
• 제17권2호
• /
• pp.295-308
• /
• 2000

OSMI(Ocean Scanning Multi-Spectral Imager) raw image data(Level 0) were acquired and radiometrically corrected. We have applied two methods, using solar & dark calibration data from OSMI sensor and comparing with the SeaWiFS data, to the radiometric correction of OSMI raw image data. First, we could get the values of the gain and the offset for each pixel and each band from comparing the solar & dark calibration data with the solar input radiance values, calculated from the transmittance, BRDF (Bidirectional Reflectance Distribution Function) and the solar incidence angle($\beta$, $\theta$) of OSMI sensor. Applying this calibration data to OSMI raw image data, we got the two odd results, the lower value of the radiometric corrected image data than the expected value, and the Venetian Blind Effect in the radiometric corrected image data. Second, we could get the reasonable results from comparing OSMI raw image data with the SeaWiFS data, and get a new problem of OSMI sensor.

• 비파괴검사학회지
• /
• 제10권1호
• /
• pp.91-95
• /
• 1990

Welding residual stress was measured by Barkhausen noise method. The calibration experiment was done for the quantitative analysis. The specimen for the calibration experiment must has the same thermo-mechanical history as the actual material to be tested. The Barkhausen noise were analysed by the pulse-height distribution. The results show that the distribution and magnitude of welding residual stress from Barkhausen noise method are in good agreement with those from blind hole method.

• ETRI Journal
• /
• 제38권3호
• /
• pp.551-559
• /
• 2016

In this paper, we propose a hybrid sensor calibration scheme for mobile crowdsensing applications. As the number of newly produced mobile devices containing embedded sensors continues to rise, the potential to use mobile devices as a sensor data source increases. However, because mobile device sensors are generally of a lower performance and cost than dedicated sensors, sensor calibration is crucial. To enable more accurate measurements of natural phenomena through the use of mobile device sensors, we propose a hybrid sensor calibration scheme for such sensors; the scheme makes use of mobile device sensors and existing sensing infrastructure, such as weather stations, to obtain dense data. Simulation results show that the proposed scheme supports low mean square errors. As a practical application of our proposed scheme, we built a temperature map of a city using six mobile phone sensors and six reference sensors. Thanks to the mobility of the sensors and the proposed scheme, our map presents more detailed information than infrastructure-based measurements.

• 한국근적외분광분석학회:학술대회논문집
• /
• 한국근적외분광분석학회 2001년도 NIR-2001
• /
• pp.1091-1091
• /
• 2001

Previous reports have shown that Near Infrared Spectroscopy (NIRS) can be used to assess physical and chemical properties of flax fibre and fabric quality. Currently, spinners assess yarn quality mainly based on strength and regularity measurements. There two key characteristics are influenced by quality of raw fibres used, especially the degree of rotting and strength. The aim of this investigation was to evaluate the use of NIRS for assessing quality of weft grade yarn available on the commercial market. In order to develop the NIR calibrations, a range of samples representing poor, medium and good quality weft yarn samples was included in the calibration and validation sample sets. The samples were analysed for physical and chemical parameters including caustic weight loss, fibre fractions, lipid, ash and minerals. A detailed protocol for assessing yarn quality has been developed to maximize the accuracy of the reflectance spectra. The development of partial least squares regression models and validation of the calibration equations using blind samples will be presented and discussed.

• Journal of Information Processing Systems
• /
• 제13권1호
• /
• pp.83-94
• /
• 2017

Near-field source localization algorithms are very sensitive to sensor gain/phase response errors, and so it is important to calibrate the errors. We took into consideration the uniform linear array and are proposing a blind calibration algorithm that can estimate the directions-of-arrival and range parameters of incident signals and sensor gain/phase responses jointly, without the need for any reference source. They are estimated separately by using an iterative approach, but without the need for good initial guesses. The ambiguities in the estimations of 2-D electric angles and sensor gain/phase responses are also analyzed in this paper. We show that the ambiguities can be remedied by assuming that two sensor phase responses of the array have been previously calibrated. The behavior of the proposed method is illustrated through simulation experiments. The simulation results show that the convergent rate is fast and that the convergent precision is high.

• 전자공학회논문지
• /
• 제54권1호
• /
• pp.96-110
• /
• 2017

기존의 차량 검출 연구들의 대부분은 일반렌즈 또는 광각렌즈를 가지는 후방 카메라를 사용하기 때문에 사각지대가 넓으며, 영상에 노이즈 및 다양한 외부 환경에 취약한 부분이 있다. 본 논문에서는 사각지대를 줄이고, 노이즈 및 가혹한 외부 환경에서도 인식이 가능한 검출 방법을 제안한다. 먼저 광각렌즈보다 더 넓은 화각을 가진 어안렌즈를 이용해 사각지대를 최소화한다. 렌즈의 화각이 커진 만큼 비선형 방사왜곡도 커지게 되므로, 정확한 영상 결과를 얻기 위해서 왜곡 상수 초기화와 최적화를 실시한 후 Calibration을 이용하였다. 그리고 Calibration과 동시에 원본 영상을 분석하여 안개가 자욱한 상황과 갑작스러운 조도 변화로 인해 생기는 명순응, 암순응 현상에 의한 시야 방해 상황에서도 인식이 가능하도록 안개 제거와 밝기 보정을 이용하였다. 안개 제거는 일반적으로 계산 시간이 매우 크다. 따라서 계산 시간을 줄이기 위해 대표적인 안개 제거 알고리즘인 Dark channel prior를 기반으로 안개를 제거하였다. 밝기 보정 시에는 Gamma correction을 이용했고, 보정에 필요한 Gamma value를 결정하기 위해 영상에 대한 밝기 및 명암 평가가 수행하였다. 평가는 영상의 전체가 아닌 일부분을 이용하여 할애되는 계산시간을 줄였다. 밝기 및 명암 값이 계산되면 그 값을 이용해 Gamma value를 결정하고 전체 영상에 보정을 실시하였다. 그리고 밝기 보정과 안개 제거로 나누어 병렬 처리한 후, 영상을 하나로 정합함으로써 전 처리 과정의 연산시간을 줄였다. 이후 보정된 영상으로부터 특징추출법인 HOG를 이용하여 차량을 검출하였다. 그 결과 본 논문에서 제안하는 방법의 영상 보정을 이용한 차량 검출을 하는데 1프레임당 0.064초가 걸렸으며, 기존의 차량 검출 방법에 비해 7.5%의 향상된 검출률을 얻었다.

• 방사선산업학회지
• /
• 제18권2호
• /
• pp.101-107
• /
• 2024

Rapid screening for internal contamination by alpha- and beta-emitting radionuclides is essential in situations involving radiation workers or radiation accidents. This study focused on the use of urine samples and liquid scintillation counting to quickly and accurately assess contamination. Calibration of the alpha and beta detection areas ensured precise measurement results. The major radionuclides recommended for surveillance during accidents were also considered. This study evaluated the effectiveness of the method by examining various parameters, including the limit of detection, linearity, sensitivity, selectivity, accuracy, ruggedness, and blind test sample analysis. The liquid scintillation counting method is an effective tool for screening urinary samples to detect alpha- and beta-emitting radionuclides, particularly during radiation emergencies, despite some limitations in precision.

• 한국수자원학회논문집
• /
• 제41권10호
• /
• pp.969-982
• /
• 2008

수문 모형의 발전이 거듭되면서, 최적 수자원의 관리를 위한 적정한 방법으로 인식되고 있다. 특히 수자원관리에 있어서 토지이용 변화 및 기후 변화에 따른 수문학적 영향 평가에 대한 요구가 증가하고 있다. 이 영향들을 평가하기 위해서는 우선 적용된 수문 모형의 강력한 검증이 요구된다. 그리고 수문모형의 적용 시 많은 지점에서 유량이 미 계측 되었거나, 측정된 자료마저 많은 오차를 포함하고 있는 경우가 있기 때문에 모형의 예측 값을 이용하여 수문분석이 이루어지는 경우가 많다. 이와 같은 경우에는 모형 결과 값에 대한 오차를 줄이기 위해서 강력한 모형 검증방법이 요구된다. 본 연구에서는 다른 물리적 특성을 가진 두 유역을 대상에 측정 유량을 통한 SWAT 모형의 검증 방법을 증명하고자 하였다. 이를 위하여, 금강유역에 위치한 수문학적 특징이 상이한 갑천유역과 용담유역을 선정하여, 먼저 갑천유역에 대하여 정밀한 검 보정을 실시하고, 갑천유역에서 개발된 방법론을 용담유역에 적용하여 모형의 검증을 시도하였다. 용담유역에 대하여 SWAT 모형을 적용한 결과 각 소유 역에서 $R_{eff}$는 0.49$\sim$0.85, $R^{2}$는 0.49$\sim$0.84로 모형은 관찰 값을 양호하게 모의하고 있는 것으로 나타났다. 그리고 모의 결과는 첨두유량 값은 다소 과소 산정하였지만, 전체적인 경향 및 기저유출량을 잘 모의하는 것으로 나타났다. 본 연구의 결과로부터 SWAT모형은 향후 토지이용변화 및 기후변화에 따른 유역특성변화 분석에 사용될 수 있을 것으로 판단된다. 하지만, 본 연구에서 사용된 혼용기법의 신뢰성을 높이기 위해서는 향후 추가적인 유역에 대한 방법론의 타당성 검증 절차를 거쳐야 할 것으로 사료된다.