• The KIPS Transactions:PartB
• /
• v.13B no.4 s.107
• /
• pp.353-360
• /
• 2006

This paper presents a robust algorithm for segmenting a vehicle license plate area from a road image. We consider the features of license plates in three aspects : 1) edges due to the characters in the plate, 2) colors in the plate, and 3) geometric properties of the plate. In the preprocessing step, we compute the thresholds based on each feature to decide whether a pixel is inside a plate or not. A statistical approach is applied to the sample images to compute the thresholds. For a given road image, our algorithm binarizes it by using the thresholds. Then, we select three candidate regions to be a plate by searching the binary image with a moving window. The plate area is selected among the candidates with simple heuristics. This algorithm robustly detects the plate against the transformation or the difference of color intensity of the plate in the input image. Moreover, the preprocessing step requires only a small number of sample images for the statistical processing. The experimental results show that the algorithm has 97.8% of successful segmentation of the plate from 228 input images. Our prototype implementation shows average processing time of 0.676 seconds per image for a set of $1280{\times}960$ images, executed on a 3GHz Pentium4 PC with 512M byte memory.

• Journal of Biomedical Engineering Research
• /
• v.34 no.4
• /
• pp.182-188
• /
• 2013

This study aims to develop a practical algorithm which can detect freezing of gait(FOG) in patients with Parkinson's disease(PD). Eighteen PD patients($68.8{\pm}11.1yrs.$) participated in this study, and three($68.7{\pm}4.0yrs.$) of them showed FOG. We suggested two time-domain algorithms(with 1-axis or 3-axes acceleration signals) and compared them with the frequency-domain algorithm in the literature. We measured the acceleration of left foot with a 3-axis accelerometer inserted at the insole of a shoe. In the time-domain method, the root-mean-square(RMS) acceleration was calculated in a moving window of 4s and FOG was defined as the periods during which RMS accelerations located within FOG range. The parameters in each algorithm were optimized for each subject using the simulated annealing method. The sensitivity and specificity were same, i.e., $89{\pm}8%$ for the time-domain method with 1-axis acceleration and were $91{\pm}7%$ and $90{\pm}8%$ for the time-domain method with 3-axes acceleration, respectively. Both performances were better in the time-domain methods than in the frequency-domain method although the results were statistically insignificant. The amount of calculation in the time-domain method was much smaller than in the frequency-domain method. Therefore it is expected that the suggested time domain algorithm would be advantageous in the systematic implementation of FOG detection.

• Korean Journal of Cognitive Science
• /
• v.16 no.1
• /
• pp.17-27
• /
• 2005

After prolonged viewing of a moving pattern, a stationary pattern can appear to move in the opposite direction, a phenomenon known as motion aftereffect (MAE). Unlike the classical explanation MAE was not confined to an adapted region; instead it can spread to an adjacent region, which was not adapted previously. In order to examine the relative locus of the mechanism responsible for MAE spreading, a rotating harmonic spiral pattern was presented as an adapting stimulus within an annulus window, and then the duration of MAE was measured in both the adapted annulus region and the non-adapted inner region. Two different kinds of test patterns were used: the same and mirror images of the original adapting pattern. An interesting characteristic of a harmonic spiral is that the orientation of a contour at a given location is different from thar of its mirror image by 90 degrees, and consequently the adapting effect of local motion detector is not expected to occur in the mirror image. The results showed that MAE duration in an adapted region was longer in the same image condition than in its mirror image condition, while MAE duration in an non-adapted region was not found to be different between those two different image conditions. These results suggest that MAE spreading might be produced by the adaptation of global motion detectors, not by local motion detectors.

• Proceedings of KOSOMES biannual meeting
• /
• 2005.11a
• /
• pp.165-170
• /
• 2005

A basic study for the development of a disk type oil skimmer for a tank lorry, compares via experiments to the oil recovery differences by material between the existing oil skimmer disk and several discs which oil adhesion are possible. The experiment results in this study are summarized as follows: In all the disks, the mass of recovery of bunker-A oil was greater than light oil. In light oil, there was nearly no differences in the mass of oil recovery by disk material, but in the case of bunker A oil, recovery efficiency showed big differences depending on the disk materials. For light oil, the differences in the mass of oil recovery per unit of operation time from the initial time zone were hardly shown. However, the mass of recovery of bunker-A oil linearly increased from the initial operation time in all the disks and the increase shaped a slow moving trend as time went on; therefore, the mass of oil recovery per unit of operation time showed differences depending on time. This result shows that oil viscosity has an effect on the disk recovery efficiency. When comparing the mass of pure oil recovery and the mass of water recovery in the total mass of recovery by bunker-A oil, there was no difference in the mass of oil recovery between the window-aluminum material disk and the disks consisting of other materials, but the mass of water recovery of the former was relatively very small. This shows the most ideal result in view of oil recovery efficiency that considers the mass of water recovery. In conclusion, it was found that aluminum is the most advantageous as the material for tank lorry oil skimmer disk.

• Proceedings of the KSRS Conference
• /
• v.2
• /
• pp.1011-1014
• /
• 2006

Coarse resolution (9 - 50 km pixels) Sea Surface Temperature satellite data are frequently considered adequate for open ocean research. However, coastal regions, including coral reef, estuarine and mesoscale upwelling regions require high-resolution (1-km pixel) SST data. The AVHRR SST data often suffer from navigation errors of several kilometres and still require manual navigation adjustments. The second serious problem is faulty and ineffective cloud-detection algorithms used operationally; many of these are based on radiance thresholds and moving window tests. With these methods, increasing sensitivity leads to masking of valid pixels. These errors lead to significant cold pixel biases and hamper image compositing, anomaly detection, and time-series analysis. Here, after manual navigation of over 40,000 AVHRR images, we implemented a new cloud filter that differs from other published methods. The filter first compares a pixel value with a climatological value built from the historical database, and then tests it against a time-based median value derived for that pixel from all satellite passes collected within ${\pm}3$ days. If the difference is larger than a predefined threshold, the pixel is flagged as cloud. We tested the method and compared to in situ SST from several shallow water buoys in the Florida Keys. Cloud statistics from all satellite sensors (AVHRR, MODIS) shows that a climatology filter with a $4^{\circ}C$ threshold and a median filter threshold of $2^{\circ}C$ are effective and accurate to filter clouds without masking good data. RMS difference between concurrent in situ and satellite SST data for the shallow waters (< 10 m bottom depth) is < $1^{\circ}C$, with only a small bias. The filter has been applied to the entire series of high-resolution SST data since1993 (including MODIS SST data since 2003), and a climatology is constructed to serve as the baseline to detect anomaly events.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• v.29 no.1
• /
• pp.263-271
• /
• 2005

The most important point when we engage on waterway dredging work is supplying safe navigational passage to the vessels underway by narrowing dredge work area and removing submerged dangers. In order to meet this end it is neccessary to use auxiliary equipment for shifting actively and mooring and adopt automation of dredging work by integrating information on real time position, dredging depth, and work information. The danger with a spud control system in this study, by the way, is able to employed on continuous dredging work with the narrowest working area allowing wide and safe passages to vessels underway, by moving the dredger to the working zone with the spud controlled automatically. Furthermore, it has been improved definitely compared with the existing dredging proccess management system such that it shows the track of spud and working depth on the electronic navigation chart of window, together with the final outcome of dredging work. The test dredging work at the entrance of Busan North Port for system evaluation showed that actual working time available was twice of the one by the existing anchor system, and that it reduced 38% of time for preparation work and one man power.

• Journal of the Korean Society of Marine Environment & Safety
• /
• v.11 no.2 s.23
• /
• pp.103-108
• /
• 2005

For the development of a disk type oil skimmer for a tank lorry, a basic study on comparison of oil recovery performance between disk materials of oil skimmer was conducted. The experiment results are summarized as follows: In all the disks, the volume of recovery of bunker-A oil was greater than diesel oil. In light oil, there was nearly no differences in the volume of oil recovery by disk material, but in the case of bunker A oil, recovery efficiency showed big differences depending on the disk materials. For diesel oil, the differences in the volume of oil recovery per unit of operation time from the initial time zone were hardly shown. However, the volume of recovery of bunker-A oil linearly increased from the initial operation time in all the disks and the increase showed a slew moving trend as time. went on; therefore, the volume of oil recovery per unit of operation time showed differences depending on time. This result shows that oil viscosity has an effect on the disk recovery efficiency. When comparing the mass of pure oil recovery and the volume of water recovery in the total mass of recovery by bunker-A oil, there was no difference in the volume of oil recovery between the window-aluminum material disk and the disks consisting of other materials, but the volume of water recovery of the former was relatively very small This shows the most ideal result in view of oil recovery efficiency that considers the volume of water recovery. In conclusion, it was found that aluminum is the most advantageous as the material for tank lorry oil skimmer disk.

• KIEAE Journal
• /
• v.17 no.3
• /
• pp.83-90
• /
• 2017

Purpose: With the increase in the energy consumption in the building sector, many studies have recently been conducted to address this issue. Due to its efficiency, diverse studies on a light-shelf, which is a natural lighting system, are in progress. However, there has been no research on the external illumination that determines the performance of a light-shelf. Therefore, the present study aimed to prepare the external illumination standard for securing the lighting performance of a light-shelf through the light-shelf performance evaluation based on a 1:1 scale testbed, and to suggest a height-adjustable light-shelf based on an open and closed concept for the case in which the lighting performance of the light-shelf significantly deteriorates, by collecting the external illumination. Method: In this study, a 1:1 scale testbed was established for performance evaluation, and the external illumination standard for securing the performance of the light-shelf was prepared by comparing the lighting performance of the light-shelf depending on the open and closed condition of the light-shelf and the external illumination. Result: 1) In this study, a light-shelf that can be opened or closed depending on the external illumination was suggested. As a result, the prospect right can be secured by creating the condition where there is no light-shelf installation by moving the light-shelf to the top of the window when the lighting performance is not secured. 2) In the summer solstice, the external illuminations appropriate for lighting energy reduction were more than 75,000 lx and 60,000 lx for the light-shelf width of 0.3 m and 0.6 m, respectively. 3) In the intermediate period, the external illumination appropriate for lighting energy reduction was 60,000 lx. In the winter solstice, making the condition where there is no light-shelf installation by closing the light-shelf would be appropriate. 4) Based on the aforementioned results, the external illumination standard for the opening and closing of the height-adjustable light-shelf based on an open and closed concept suggested in this study was 60,000 lx, and the light-shelf with a width of 0.6 m would be advantageous for lighting energy reduction.

• Journal of Navigation and Port Research
• /
• v.29 no.5 s.101
• /
• pp.395-402
• /
• 2005

The most important point when we engage on waterway dredging work is supplying safe navigational passage to the vessels underway by narrowing dredge work area and removing submerged dangers. In order to meet this end it is necessary to use auxiliary equipment for shifting actively and mooring and adopt automation of dredging work by integrating information on real time position, dredging depth, and work information. The dredger with a spud control system in this study, by the way, is able to employed on continuous dredging work with the narrowest working area allowing wide and safe passages to vessels underway, by moving the dredger to the working zone with the spud controlled automatically. Furthermore, it has been improved definitely compared with the existing dredging process management system such that it shows the track of spud and working depth on the electronic navigation chart of window, together with the final outcome of dredging work. The test dredging work at the entrance of Busan North Port for system evaluation showed that actual working time available was twice of the one by the existing anchor system, and that it reduced $38\%$ of time for preparation work and one man power.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.11
• /
• pp.310-318
• /
• 2018

Recently, the importance of big data analysis is increasing as a large amount of data is generated by various devices connected to the Internet with the advent of Internet of Things (IoT). Especially, it is necessary to analyze various large-scale IoT streaming sensor data generated in real time and provide various services through new meaningful prediction. This paper proposes a real-time indoor PM10 concentration prediction LSTM model based on streaming data generated from IoT sensor using AWS. We also construct a real-time indoor PM10 concentration prediction service based on the proposed model. Data used in the paper is streaming data collected from the PM10 IoT sensor for 24 hours. This time series data is converted into sequence data consisting of 30 consecutive values from time series data for use as input data of LSTM. The LSTM model is learned through a sliding window process of moving to the immediately adjacent dataset. In order to improve the performance of the model, incremental learning method is applied to the streaming data collected every 24 hours. The linear regression and recurrent neural networks (RNN) models are compared to evaluate the performance of LSTM model. Experimental results show that the proposed LSTM prediction model has 700% improvement over linear regression and 140% improvement over RNN model for its performance level.