• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.402-404
• /
• 2003

We present two methods for the automatic selection of the threshold values in unsupervised change detection. Both methods consist of the same two procedures: 1) to determine the parameters of Gaussian mixtures from a difference image or ratio image, 2) to determine threshold values using the Bayesian rule for minimum error. In the first method, the Expectation-Maximization algorithm is applied for estimating the parameters of the Gaussian mixtures. The second method is based on the iterative thresholding that successively employs thresholding and estimation of the model parameters. The effectiveness and applicability of the methods proposed here are illustrated by an experiment on the multi-temporal KOMPAT-1 EOC images.

• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.493-495
• /
• 2003

The establishment of the investigation and the maintenance technique is required for preservation of old structures in Japan. This study attempts to diagnose the deterioration status of the historic structures using the thermal infrared camera. In some structures, the difference of the spatio-temporal change was observed in the surface temperature. For example, the cold joint of concrete was examined using this method effectively. As a result of this study, we have found useful guidelines in developing methodology to conduct diagnosis of historic buildings by using thermal infrared camera.

• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.244-247
• /
• 2003

Multi-temporal data is used to determine the rate of deforestation between the years 1976, 1987 and 2000. Three Landsat TM images, for each period, are pre-processed, mosaicked and normalized difference vegetation index (NDVI) values computed. Based on the values, totally non-forested areas are masked out. The forested areas, both partially and wholly, show a very high degree of correlation between all the bands (reflective), thus necessitating application of principal component analysis. The first two principal components and NDVI values (scaled to 0 ? 255) are used in K-means unsupervised classification to distinguish forest from non-forest areas (that appeared as forest at first). Comparison of the resulting thematic maps gives an annual deforestation rate of roughly 15 0000ha. or 2% between any two epochs.

• Korean Journal of Remote Sensing
• /
• v.19 no.5
• /
• pp.341-349
• /
• 2003

The so-called B-method, a simplified surface energy budget, permits calculation of daily actual evapotranspiration (ET) using remotely sensed data, such as NOAA-AVHRR. Even if the use of satellite data allows estimation of the albedo and surface temperature, this model requires meteorological data measured at ground-level to obtain the other inputs. In addition, a difficulty may be occurred by the difference of temporal scales between the net radiation in daily scale and instantaneous measurement at midday of the surface and air temperatures because the data covered whole day are necessary to obtain accumulated daily net radiation. In order to solve these problems, this study attempted a modification of B-method through an extension of hourly ET value calculated using a complete instantaneous inputs. The estimation of the daily apparent ET from newly proposed system showed a root mean square error of 0.26 mm/day as compared the output obtained from the classical model. It is evident that this may offer more rapid estimation and reduced data volume.

• Journal of Korea Multimedia Society
• /
• v.22 no.6
• /
• pp.639-645
• /
• 2019

This paper proposes a salient chromagram components extraction method based on the temporal discrete cosine transform of a chromagram block to improve cover song retrieval accuracy. The proposed salient chromagram emphasizes tonal contents of music, which are well-preserved between an original song and its cover version, while reducing the effects of timbre difference. We apply the proposed salient chromagram extraction method as a preprocessing step for the Fourier-transform based cover song matching. Experiments on two cover song datasets confirm that the proposed salient chromagram improves the cover song search accuracy.

• Current Applied Physics
• /
• v.18 no.11
• /
• pp.1230-1234
• /
• 2018

We report a new method of measuring the non-radiative recombination rate in bulk Silicon. Synchrotron timeresolved x-ray scattering (TRXS) combines femtometer spatial sensitivity with nanosecond time resolution to record the temporal evolution of a crystal lattice following intense ultrafast laser excitation. Modeling this data requires an Auger recombination time that is considerably slower than previous measurements, which were made at lower laser intensities while probing only a relatively shallow surface depth. We attribute this difference to an enhanced Coulomb interaction that has been predicted to occur in bulk materials with high densities of photoexcited charge carriers.

• Journal of Advanced Information Technology and Convergence
• /
• v.9 no.2
• /
• pp.29-39
• /
• 2019

Tidal Flats are very important natural resource and various efforts have been made to protect it from environmental pollutions. The projects to monitor the environmental changes by periodically observing the creatures in tidal flats are underway. However, they are being done inefficiently by people directly observing. In this paper, we propose an object segmentation method that can be applied to the applications which automatically monitor the living creatures in tidal flats. In the proposed method, a foreground map representing the location of objects is obtained by using a temporal difference method, and then a superpixel method is applied to detect the detailed boundary of an object. The region of a crab is extracted finally by combining the foreground map and the superpixel information. Experimental results show that the proposed method separates crab regions from a tidal flat image easily and accurately.

• Journal of the Korean Geographical Society
• /
• v.47 no.5
• /
• pp.641-653
• /
• 2012

We analyzed two radiocarbon-dated lagoonal sediment cores from Cheonjinho and Ssangho using various biogeochemical methods. As a result, the start times of rice agriculture are estimated to be AD 780 for Cheonjinho and 100 BC for Ssangho. There is a large temporal difference in the beginning of rice agriculture between two study sites even though they are closely located on the coast. This result indicates that pollen records are not sufficient to approximate the start time of rice agriculture accurately. A temporal lag seems to exist between the time when rice agriculture was first introduced and the time when full-scale rice agriculture began in the Yeongdong region, probably because of low agricultural productivity. In both study sites, rice agriculture intensified and slope erosion increased 250 years after full-scale agriculture began. This suggests that intensified rice agriculture resulted in an increased number of inhabitants, settlement expansion to hilly areas, and the consequent deforestation.

• Journal of Digital Contents Society
• /
• v.19 no.4
• /
• pp.789-799
• /
• 2018

This article presents machine learning based approach on Big data to analyzing time series data for anomaly detection in such industrial complex system. Long Short-Term Memory (LSTM) network have been demonstrated to be improved version of RNN and have become a useful aid for many tasks. This LSTM based model learn the higher level temporal features as well as temporal pattern, then such predictor is used to prediction stage to estimate future data. The prediction error is the difference between predicted output made by predictor and actual in-coming values. An error-distribution estimation model is built using a Gaussian distribution to calculate the anomaly in the score of the observation. In this manner, we move from the concept of a single anomaly to the idea of the collective anomaly. This work can assist the monitoring and management of Smart Factory in minimizing failure and improving manufacturing quality.

• The Journal of the Acoustical Society of Korea
• /
• v.14 no.1E
• /
• pp.5-13
• /
• 1995

In order to estimate the variability of the wave propagation loss (PL) du e to hourly variations of the sound speed profiles (SSPs), we conducted oceanographic measurements every hour for 39 hours in October 1993 in the Korea Strait. Currents and meteorological data were measured simultaneously to examine the causes of the temporal variations of temperatures. During the experiment, the temporal variations of temperatures in the surface layer highly depend on the water mass transport from adjacent seas. The PL for low frequency (75-300 Hz) is calculated using the parabolic equation scheme and averaged over the whole water depth. The hourly variation of the SSP may cause a PL difference of up to 10 dB over a 30-50 km range. The variability of PL, represented by standard deviation for the 39 SSPs, is as large as 3 dB over a 50 km range.