• Journal of the Korea Society of Computer and Information
• /
• v.27 no.10
• /
• pp.19-28
• /
• 2022

In this paper, we propose the Detector Scheduler which determines the best tracking-by-detection (TBD) mechanism to perform real-time high-accurate multi-object tracking (MOT). The Detector Scheduler determines whether to run a detector by measuring the dissimilarity of features between different frames. Furthermore, we propose a self-supervision method to learn the Detector Scheduler with tracking results since it is difficult to generate ground truth (GT) for learning the Detector Scheduler. Our proposed self-supervision method generates pseudo labels on whether to run a detector when the dissimilarity of the object cardinality or appearance between frames increases. To this end, we propose the Detector Scheduling Loss to learn the Detector Scheduler. As a result, our proposed method achieves real-time high-accurate multi-object tracking by boosting the overall tracking speed while keeping the tracking accuracy at most.

• Journal of electromagnetic engineering and science
• /
• v.18 no.3
• /
• pp.169-174
• /
• 2018

In this paper, we propose an estimation method using multiple in-phase and quadrature (IQ) signals of different frequencies to evaluate the sample-time errors in the sampling oscilloscope. The estimator is implemented by ODRPACK, and a novel iteration scheme is applied to achieve fast convergence without any prior information. Monte-Carlo simulation is conducted to confirm the proposed method. It clearly shows that the multiple IQ approach achieves more accurate results compared to the conventional method. Finally, the criteria for the frequency selection and the signal capture time are investigated.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 1999.11a
• /
• pp.490-493
• /
• 1999

Due to the miniaturization of electrical components and assembles on signal and telecommunication circuits, transient overvoltages caused by switching operation or lightning surge has become more interesting concern to the field of electrical engineering. This paper describes the development of transient voltage blocking devices (TBDs) that can protect sensitive signal and telecommunication devices from overvoltages. Two kinds of TBDs are designed and tested by using a combination surge generator which can produce the standard impulse current of 8/20${\mu}$s 2.1kA according to the IEC 1000-4-5 standard. From the simulation and experimental reults, it is confirmed that the proposed TBD has an enough protection performance in low insertion loss and tight clamping voltage.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.461-463
• /
• 2023

본 연구는 실시간 MOT(Multiple-Object-Tracking)의 성능을 향상시키기 위해 다양한 기법을 적용한 MOT 모델을 설계한다. 연구에서 사용하는 Backbone 모델은 TBD(Tracking-by-Detection) 기반의 Tracking 모델을 사용한다. Heatmap Detection을 통해 객체를 검출하고 Transformer 기반의 Feature를 연결하여 Tracking 한다. 제안하는 방법은 Anchor 기반의 Detection의 장시간 문제와 추적 객체 정보 전달손실을 감소하여 실시간 객체 추적에 도움이 될 것으로 사료된다.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.14 no.3
• /
• pp.509-521
• /
• 1994

This study is an effort to develop computer simulation model that produce precipitation patterns from stochastic model. A stochastic model is formulated for the process of daily precipitation with considering the sequences of wet and dry days and the precipitation amounts on wet days. This study consists of 2 papers and the process of precipitation occurrence is modelled by an alternate renewal process (ARP) in paper (I). In the ARP model for the precipitation occurrence, four discrete distributions, used to fit the wet and dry spells, were as follows; truncated binomial distribution (TBD), truncated Poisson distribution (TPD), truncated negative binomial distribution (TNBD), logarithmic series distribution (LSD). In companion paper (II) the process of occurrence is developed by Markov chain. The amounts of precipitation, given that precipitation has occurred, are described by a Gamma. Pearson Type-III, Extremal Type-III, and 3 parameter Weibull distribution. Daily precipitation series model consists of two models, A-Wand A-G model, by combining the process of precipitation occurrence and a continuous probability distribution on the precipitation of wet days. To evaluate the performance of the simulation model, output from the model was compared with historical data of 7 stations in the Nakdong and Seomjin river basin. The results of paper (1) show that it is possible to design a model for the synthetic generation of IX)int precipitation patterns.

• Journal of Forest and Environmental Science
• /
• v.32 no.4
• /
• pp.329-343
• /
• 2016

The present study was conducted on ridge tops of moist temperate Oak forests in Garhwal Himalaya to assess the plant diversity, regeneration, biomass production and carbon assimilation in different Oak forests. For this purpose, three Oak forest types viz., (a) Quercus leucotrichophora or Banj Oak (FT1; between 1,428-2,578 m asl), (b) Quercus floribunda or Moru Oak (FT2; between 2,430-2,697 m asl) and (c) Quercus semecarpifolia or Kharsu Oak (FT3; between 2,418-3,540 m asl) were selected on different ridge tops in Bhagirathi catchment area of Garhwal Himalaya. A total of 91 plant species including 23 trees (8 gymnosperms and 15 angiosperms), 21 shrubs and 47 herbs species belonging to 46 families were recorded from all the ridge top Oak forests. The highest mean tree density ($607{\pm}33.60trees\;ha^{-1}$) was observed in Q. floribunda forest with lower mean total basal cover (TBC) value ($48.02{\pm}3.67m^2ha^{-1}$), whereas highest TBC value ($80.16{\pm}3.30m^2ha^{-1}$) was recorded for Q. semecarpifolia forest, with lowest mean stem density ($594{\pm}23.43stems\;ha^{-1}$). The total biomass density (TBD) across three Oak forests ranged between $497.32{\pm}83.70$ (FT1) and $663.16{\pm}93.85t\;ha^{-1}$ (FT3), while the total carbon density (TCD) values ranged between $228.75{\pm}22.27$ (FT1) and $304.31{\pm}18.12t\;ha^{-1}$ (FT3). Most of the tree species were found with good regeneration (GR) status (average 45%) in all the forest types whereas, few species were found not regenerating (NR) (average 17%) however, few new recruitments were also recorded. ANOVA (Post-Hoc Tukey's test at 5% significance level) indicated significant forest-wise differences in TBC, TBD and TCD (in tree layer); family and evenness (in shrub layer only) values, while insignificant differences were noticed in density values of tree, seedling and herb layer.

• The Bulletin of The Korean Astronomical Society
• /
• v.40 no.2
• /
• pp.27.2-27.2
• /
• 2015

Strategic plan of Subaru science and operation will be introduced. Currently, Subaru has wide variety of instruments, conducts only classical observations, with less than 5 nights allocation for each proposal. Near future, Subaru will emphasize on surveys, introduce queue mode observations, reduce the number of instruments, and concentrate on large size programs. Large surveys are called Subaru Strategic Programs (SSPs). HSC-SSP is on-going (300 nights for 5 years), PFS-SSP will start at around 2020 (360 nights for 5 years), and IRD-SSP from 2016 (TBD). HSC science includes 1) cosmology with gravitational lensing, 2) lensing studies of galaxies and clusters, 3) photometric redshifts, 4) the Solar system, 5) the Milky Way and the Local Group, 6) AGN/quasars, 7) transients, 8) galaxies at low/high redshifts, and 9) clusters of galaxies. PFS science includes 1) cosmology, 2) galaxy & AGN, and 3) galactic archaeology. Subaru is planning the third pillar instrument, so called ULTIMATE-Subaru, which is the GLAO optical-NIR wide field camera & multi-IFU spectrograph for finding galaxies at ultra high redshift (z>10). Finally the strategy from Subaru to TMT will be presented. Subaru will conduct four major SSPs (HSC, PFS, IRD, ULTIMATE-Subaru) in coming decade to provide targets to TMT. HSC performs wide field surveys to reveal the distribution of dark matter in the Universe. IRD surveys Earth-like young planets to discover ~20 Earth-like habitable planets. PFS studies the expanding Universe to provide a few million emission line galaxies to TMT.

• Journal of Korea Multimedia Society
• /
• v.20 no.6
• /
• pp.893-910
• /
• 2017

In this paper, we propose a simple online multiple object (human) tracking method, LKDeep (Lucas-Kanade feature and Detection based Simple Online Multiple Object Tracker), which can run in fast online enough on CPU core only with acceptable tracking performance for embedded surveillance purpose. The proposed LKDeep is a pragmatic hybrid approach which tracks multiple objects (humans) mainly based on LK features but is compensated by detection on periodic times or on necessity times. Compared to other state-of-the-art multiple object tracking methods based on 'Tracking-By-Detection (TBD)' approach, the proposed LKDeep is faster since it does not have to detect object on every frame and it utilizes simple association rule, but it shows a good object tracking performance. Through experiments in comparison with other multiple object tracking (MOT) methods using the public DPM detector among online state-of-the-art MOT methods reported in MOT challenge [1], it is shown that the proposed simple online MOT method, LKDeep runs faster but with good tracking performance for surveillance purpose. It is further observed through single object tracking (SOT) visual tracker benchmark experiment [2] that LKDeep with an optimized deep learning detector can run in online fast with comparable tracking performance to other state-of-the-art SOT methods.

• Journal of Astronomy and Space Sciences
• /
• v.35 no.1
• /
• pp.31-37
• /
• 2018

Pc1 pulsations are geomagnetic fluctuations in the frequency range of 0.2 to 5 Hz. There have been several observations of Pc1 pulsations in low earth orbit by MAGSAT, DE-2, Viking, Freja, CHAMP, and SWARM satellites. However, there has been a clear limitation in resolving the spatial and temporal variations of the pulsation by using a single-point observation by a single satellite. To overcome such limitations of previous observations, a new space mission was recently initiated, using the concept of multi-satellites, named the Small scale magNetospheric and Ionospheric Plasma Experiments (SNIPE). The SNIPE mission consists of four nanosatellites (~10 kg), which will be launched into a polar orbit at an altitude of 600 km (TBD) in 2020. Four satellites will be deployed in orbit, and the distances between each satellite will be controlled from 10 to 1,000 km by a high-end formation-flying algorithm. One of the possible science targets of the SNIPE mission is observing electromagnetic ion cyclotron (EMIC) waves. In this paper, we report on examples of observations, showing the limitations of previous EMIC observations in low earth orbit, and suggest possibilities to overcome those limitations through a new mission.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.23 no.3
• /
• pp.204-212
• /
• 2020

For general target tracking works by receiving a set of measurements from sensor. However, if the SNR(Signal to Noise Ratio) is low due to small RCS(Radar Cross Section), caused by remote small targets, the target's information can be lost during signal processing. TBD(Track Before Detect) is an algorithm that performs target tracking without threshold for detection. That is, all sensor data is sent to the tracking system, which prevents the loss of the target's information by thresholding the signal intensity. On the other hand, using all sensor data inevitably leads to computational problems that can severely limit the application. In this paper, we propose an iterative Joint Integrated Probabilistic Data Association as a practical target tracking technique suitable for a low SNR multi-target environment with real time operation capability, and verify its performance through simulation studies.