• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.16 no.2
• /
• pp.36-54
• /
• 2017

The purpose of this paper is to develop a multi-sensor fusion-based traffic information acquisition system with robust to environmental changes. it combines the characteristics of each sensor and is more robust to the environmental changes than the video detector. Moreover, it is not affected by the time of day and night, and has less maintenance cost than the inductive-loop traffic detector. This is accomplished by synthesizing object tracking informations based on a radar, vehicle classification informations based on a video detector and reliable object detections of a infrared range finder. To prove the effectiveness of the proposed system, I conducted experiments for 6 hours over 5 days of the daytime and early evening on the pedestrian - accessible road. According to the experimental results, it has 88.7% classification accuracy and 95.5% vehicle detection rate. If the parameters of this system is optimized to adapt to the experimental environment changes, it is expected that it will contribute to the advancement of ITS.

• Spatial Information Research
• /
• v.16 no.4
• /
• pp.455-465
• /
• 2008

Augmented reality (AR) has tremendous potential in visualizing geospatial information, especially on the actual physical scenes. However, to utilize augmented reality in mobile system, many researches have undergone with GPS or ubiquitous marker based approaches. Although there are several papers written with vision based markerless tracking, previous approaches provide fairly good results only in largely under "controlled environments." Localization and tracking of current position become more complex problem when it is used in indoor environments. Many proposed Radio Frequency (RF) based tracking and localization. However, it does cause deployment problems of large RF-based sensors and readers. In this paper, we present a noble markerless AR approach for indoor (possible outdoor, too) navigation system only using monoSLAM (Monocular Simultaneous Localization and Map building) algorithm to full-fill our grand effort to develop mobile seamless indoor/outdoor u-GIS system. The paper briefly explains the basic SLAM algorithm, then the implementation of our system.

• The KIPS Transactions:PartB
• /
• v.18B no.4
• /
• pp.193-200
• /
• 2011

This paper proposes a low-cost visual analyzer algorithm of human body motion for real-time applications such as human-computer interfacing, virtual reality applications in medicine and telemonitoring of patients. To reduce cost of its use, we design the algorithm to use a single camera. To make the proposed system to be used more conveniently, we avoid from using optical markers. To make the proposed algorithm be convenient for real-time applications, we design it to have a closed-form with high accuracy. To design a closed-form algorithm, we propose an idea that formulates motion of a human body joint as a 2D universal joint model instead of a common 3D spherical joint model, without any kins of approximation. To make the closed-form algorithm has high accuracy, we formulates the estimation process to be an optimization problem. Thus-desined algorithm is applied to each joint of the human body one after another. Through experiments we show that human body motion capturing can be performed in an efficient and robust manner by using our algorithm.

• Journal of Drive and Control
• /
• v.20 no.2
• /
• pp.15-23
• /
• 2023

In this study, two-dimensional location of crops for auto weeding was detected using deep learning. To construct a dataset for soybean detection, an image-capturing system was developed using a mono camera and single-board computer and the system was mounted on a weeding robot to collect soybean images. A dataset was constructed by extracting RoI (region of interest) from the raw image and each sample was labeled with soybean and the background for classification learning. The deep learning model consisted of four convolutional layers and was trained with a weakly supervised learning method that can provide object localization only using image-level labeling. Localization of the soybean area can be visualized via CAM and the two-dimensional position of the soybean was estimated by clustering the pixels associated with the soybean area and transforming the pixel coordinates to world coordinates. The actual position, which is determined manually as pixel coordinates in the image was evaluated and performances were 6.6(X-axis), 5.1(Y-axis) and 1.2(X-axis), 2.2(Y-axis) for MSE and RMSE about world coordinates, respectively. From the results, we confirmed that the center position of the soybean area derived through deep learning was sufficient for use in automatic weeding systems.

• Progress in Medical Physics
• /
• v.20 no.4
• /
• pp.216-224
• /
• 2009

$^{131}I$ is a radiological isotope being used widely for treatment of cancer as emitting gamma-ray and it is also applied to estimate the function of thyroid for its accumulation in thyroid. However, $^{131}I$ is more difficult to quantitate comapred to $^{99m}Tc$, because $^{131}I$ has multiple energy gamma-ray emissions compared to $^{99m}Tc$ which is a mono energetic gamma-ray source. Especially, scattered ray and septal penetration resulted by high energy gamma ray have a bad influence upon nuclear medicine image. The purpose of this study was to estimate scatter components depending on the different source locations within a phantom using Monte Carlo simulation (GATE). The simulation results were validated by comparing with the results of real experiments. Dual-head gamma camera (ECAM, Chicago, Illinois Siemens) with high energy, general-purpose, and parallel hole collimators (hole radius: 0.17 cm, septal thickness: 0.2 cm, length: 5.08 cm) was used in this experiment. The NaI crystal is $44.5{\times}59.1\;cm$ in height and width and 0.95 cm in thickness. The diameter and height of PMMA phantom were 16 cm and 15 cm, respectively. The images were acquired at 5 different locations of $^{131}I$ point source within the phantom and the images of $^{99m}Tc$ were also acquired for comparison purpose with low energy source. The simulation results indicated that the scattering was influenced by the location of source within a phantom. The scattering effects showed the same tendency in both simulation and actual experiment, and the results showed that the simulation was very adequate for further studies. The results supported that the simulation techniques may be used to generalize the scattering effects as a function of a point source location within a phantom.

• The Korean Journal of Nuclear Medicine
• /
• v.38 no.3
• /
• pp.253-258
• /
• 2004

Purpose: Chitosan has been studied as a non-viral gene delivery vector, drug delivery carrier, metal chelator, food additive, and radiopharmaceutical, among other things. Recently, galactose-graft chitosan was studied as a non-viral gene and drug delivery vector to target hepatocytes. The aim of this study was to investigate the usefulness of nuclear imaging for in vivo evaluation of targeting the hepatocyte by galactose grafting. Methods and Materials: Galactosyl methylated chitosan (GMC) was produced by methylation to lactobionic acid coupled chitosan. Cytotoxicity of $^{99m}Tc$-GMC was determined by MTT assay. Rabbits were injected via their auricular vein with $^{99m}Tc$-GMC and $^{99m}Tc$-methylated chitosan (MC), the latter of which does not contain a galactose group, and images were acquired with a gamma camera equipped with a parallel hole collimator. The composition of the galactose group in galactosylated chitosan (GC), as well as the tri-, di-, or mono-methylation of GMC, was confirmed by NMR spectroscopy. Results: The results of MTT assay indicated that $^{99m}Tc$-GMC was non-toxic. $^{99m}Tc$-GMC specifically accumulated in the liver within 10 minutes of injection and maintained high hepatic uptake. In contrast, $^{99m}Tc$-MC showed faint liver uptake. $^{99m}Tc$-GMC scintigraphy of rabbits showed that the galactose ligand principally targeted the liver while the chitosan functionalities led to excretion through the urinary system. Conclusion: Bioconjugation with a specific ligand endows some degree of targetability to an administered molecule or drug, as in the case of galactose for hepatocyte in vivo, and evaluating said targetabililty is a clear example of the great benefit proffered by nuclear imaging.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.49 no.4
• /
• pp.90-101
• /
• 2012

The landmark selection is crucial to successful perform in SLAM(Simultaneous Localization and Mapping) with a mono camera. Especially, in unknown environment, automatic landmark selection is needed since there is no advance information about landmark. In this paper, proposed visual attention system which modeled human's vision system will be used in order to select landmark automatically. The edge feature is one of the most important element for attention in previous visual attention system. However, when the edge feature is used in complicated indoor area, the response of complicated area disappears, and between flat surfaces are getting higher. Also, computation cost increases occurs due to the growth of the dimensionality since it uses the responses for 4 directions. This paper suggests to use a corner feature in order to solve or prevent the problems mentioned above. Using a corner feature can also increase the accuracy of data association by concentrating on area which is more complicated and informative in indoor environments. Finally, this paper will prove that visual attention system based on corner feature can be more effective in SLAM compared to previous method by experiment.