• Journal of the Korean Society of Visualization
• /
• v.3 no.2
• /
• pp.71-78
• /
• 2005

Uncertainty introduced by a cross-correlation algorithm based on FFT has been investigated using PIV standard images. The standard images were generated by the Monte Carlo simulation method. Both bias and random errors from the velocity vector have been analyzed with regard to the particle diameter, displacement, and the number of particles. The uncertainty of velocity is evaluated based upon the IS0/IEC standard. As a result, a total error of $0.26\%$ is included in the PIV cross-correlation algorithm. In addition, the uncertainty budget is presented, where the effect of the above three variables is examined. According to the budget, the variation of the number of particles within the interrogation window mainly contributes to the combined standard uncertainty of the real measured velocity field when excluding the effect of errors by the experiments itself. Finally, the expanded uncertainty is found to be about $12\%$ at the $95\%$ confidence level.

• Journal of the Society of Naval Architects of Korea
• /
• v.51 no.4
• /
• pp.311-320
• /
• 2014

Test uncertainty of a towed underwater Stereoscopic Particle Image Velocimetry (SPIV) system was assessed in a towing tank. To estimate the systematic error and random error of mean velocity and turbulence properties measurement, velocity field of uniform flow was measured. Total uncertainty of the axial component of mean velocity was 1.45% of the uniform flow speed and total uncertainty of turbulence properties was 3.03%. Besides, variation of particle displacement was applied to identify the change of error distribution. In results for variation of particle displacement, the error rapidly increases with particle movement under one pixel. In addition, a nominal wake of a model ship was measured and compared with existing experimental data by five-hole Pitot tubes, Pitot-static tube, and hot wire anemometer. For mean velocity, small local vortex was identified with high spatial resolution of SPIV, but has serious disagreement in local maxima of turbulence properties due to limited sampling rate.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.33 no.2
• /
• pp.549-557
• /
• 2013

An flood inundation map is able to convey spatial distribution of inundation to a decision maker for flood risk management. A roughness coefficient with unclear values and a discharge obtained from the stage-discharge rating equation are key sources of uncertainty in flood inundation mapping by using a hydraulic model. Also, the uncertainty analysis needs an observation for the flood inundation, and satellite images is useful to obtain spatial distribution of flood. Accordingly, the objective of this study is to quantify uncertainty arising roughness and discharge in flood inundation mapping by using a hydraulic model and a satellite image. To perform this, flood inundations were simulated by HEC-RAS and terrain analysis, and ISODATA (Iterative Self-Organizing Data Analysis) was used to classify waterbody from Landsat 5TM imagery. The classified waterbody was used as an observation to calculate F-statistic (likelihood measure) in GLUE (Generalized Likelihood Uncertainty Estimation). The results from GLUE show that flood inundation areas are 74.59 $km^2$ for lower 5 % uncertainty bound and 151.95 $km^2$ for upper 95% uncertainty bound, respectively. The quantification of uncertainty in flood inundation mapping will play a significant role in realizing the efficient flood risk management.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.35 no.5
• /
• pp.341-348
• /
• 2015

Active infrared thermography methods have been known to possess good fault detection capabilities for the detection of defects in materials compared to the conventional passive thermal infrared imaging techniques. However, the reliability of the technique has been under scrutiny. This paper proposes the lock-in thermography technique for the detection and estimation of artificial subsurface defect size and depth with uncertainty measurement.

• Korean Journal of Remote Sensing
• /
• v.38 no.4
• /
• pp.327-341
• /
• 2022

Cloud removal is often required to construct time-series sets of optical images for environmental monitoring. In regression-based cloud removal, the selection of an appropriate regression model and the impact analysis of the input images significantly affect the prediction performance. This study evaluates the potential of Gaussian process (GP) regression for cloud removal and also analyzes the effects of cloud-free optical images and spectral bands on prediction performance. Unlike other machine learning-based regression models, GP regression provides uncertainty information and automatically optimizes hyperparameters. An experiment using Sentinel-2 multi-spectral images was conducted for cloud removal in the two agricultural regions. The prediction performance of GP regression was compared with that of random forest (RF) regression. Various combinations of input images and multi-spectral bands were considered for quantitative evaluations. The experimental results showed that using multi-temporal images with multi-spectral bands as inputs achieved the best prediction accuracy. Highly correlated adjacent multi-spectral bands and temporally correlated multi-temporal images resulted in an improved prediction accuracy. The prediction performance of GP regression was significantly improved in predicting the near-infrared band compared to that of RF regression. Estimating the distribution function of input data in GP regression could reflect the variations in the considered spectral band with a broader range. In particular, GP regression was superior to RF regression for reproducing structural patterns at both sites in terms of structural similarity. In addition, uncertainty information provided by GP regression showed a reasonable similarity to prediction errors for some sub-areas, indicating that uncertainty estimates may be used to measure the prediction result quality. These findings suggest that GP regression could be beneficial for cloud removal and optical image reconstruction. In addition, the impact analysis results of the input images provide guidelines for selecting optimal images for regression-based cloud removal.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.1897-1902
• /
• 2005

The objective of this paper is to estimate the concentration distribution in flow field inside the pipeline based on electrical impedance tomography. Special emphasis is given to the development of dynamic imaging technique for two-phase field undergoing a rapid transient change. Nonlinearity-compensation extended Kalman filter is employed to cope with unexpected measurement uncertainty. The nonlinearity-compensation extended Kalman filter compensates for the influence of measurement uncertainty and solves the instability of extended Kalman filter. Extensive computer simulations are carried out to show that nonlinearity-compensation extended Kalman filter has enhanced estimation performance especially in the unexpected measurement environment.

• The KIPS Transactions:PartB
• /
• v.8B no.5
• /
• pp.565-572
• /
• 2001

Analyzing functions with small values of the product of position and frequency uncertainties have many advantages in image processing and data compression. Until now, this values has been computed based on the uncertainty principle, but the computed frequency uncertainty is not practical the human visual filters which have on-zero peak response frequencies. A new metric for the frequency uncertainty is used to calculate a deviation about the frequency which has maximum response. The joint effective widths for various functions are derived. As the result of analysis, the joint uncertainty for many functions converges to 0.5 as the joint parameter increases. Furthermore. Gabor cosine function shows an excellent performance among the mentioned functions.

• Journal of Sensor Science and Technology
• /
• v.12 no.6
• /
• pp.258-264
• /
• 2003

The present paper studies how to calibrate hydrometer using image process. The method aligns particular scales of hydrometer selected for calibrating the hydrometer with the horizontal plane of the reference liquid automatically without man's operation. Major parts composing the system are CCD camera, frame grabber, stepping motor and image process program. The image process program is composed of a part that locates the meniscus and aligns it with a scale and a part that controls the step motor. To verify the performance of the developed method, this study compares the meniscus and scale observed directly with the naked eye with the result of calibration by the manual calibration method. The differences between the corrections were less than $0.004\;kg/m^3$ with uncertainty of $0.06\;kg/m^3$. These showed that the calibration results of the developed hydrometer calibration using image process nearly equal to manual method.

• Publications of The Korean Astronomical Society
• /
• v.7 no.1
• /
• pp.79-87
• /
• 1992

New and improved data on the gravitational lens systems discovered so far are compared with the theoretical predictions of Gott, Park, and Lee (1989, GPL). Systems lensed by a single galaxy, compatible with assumptions of GPL, support flat or near-flat geometry for the universe. But the statistical uncertainty is too large to draw any definite conclusion. We need more lens systems. Also, the probability of multiple image lensing and mean separation of the images averaged over the source distribution are calculated for various cosmological models. Multiple-image lens systems and radio ring systems are compared with the predictions. Although the data reject exotic cosmological models, it cannot discriminate among conventional Friedmann models yet.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.16 no.2
• /
• pp.104-110
• /
• 2016

Dense local image descriptors like SIFT are fruitful for capturing salient information about image, shown to be successful in various image-related tasks when formed in bag-of-words representation (i.e., histograms). In this paper we consider to utilize these dense local descriptors in the object tracking problem. A notable aspect of our tracker is that instead of adopting a point estimate for the target model, we account for uncertainty in data noise and model incompleteness by maintaining a distribution over plausible candidate models within the Bayesian framework. The target model is also updated adaptively by the principled Bayesian posterior inference, which admits a closed form within our Dirichlet prior modeling. With empirical evaluations on some video datasets, the proposed method is shown to yield more accurate tracking than baseline histogram-based trackers with the same types of features, often being superior to the appearance-based (visual) trackers.