• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.19 no.11
• /
• pp.31-37
• /
• 2018

This paper proposes an effective single-image haze-removal algorithm with low complexity by using a dual dark channel prior (DCP) and an adaptive brightness correction technique. The dark channel of a small patch preserves the edge information of the image, but is sensitive to noise and local brightness variations. On the other hand, the dark channel of a large patch is advantageous in estimation of the exact haze value, but halo effects from block effects deteriorate haze-removal performance. In order to solve this problem, the proposed algorithm builds a dual DCP as a combination of dark channels from patches with different sizes, and this meets low-memory and low-complexity requirements, while the conventional method uses a matting technique, which requires a large amount of memory and heavy computations. Moreover, an adaptive brightness correction technique that is applied to the recovered image preserves the objects in the image more clearly. Experimental results for various hazy images demonstrate that the proposed algorithm removes haze effectively, while requiring much fewer computations and less memory than conventional methods.

• Journal of Digital Convergence
• /
• v.16 no.11
• /
• pp.335-340
• /
• 2018

Commercialized speech recognition systems that have an accuracy recognition rates are used a learning model from a type of speaker dependent isolated data. However, it has a problem that shows a decrease in the speech recognition performance according to the quantity of data in noise environments. In this paper, we proposed the vector quantization based speech recognition performance improvement using maximum log likelihood in Gaussian distribution. The proposed method is the best learning model configuration method for increasing the accuracy of speech recognition for similar speech using the vector quantization and Maximum Log Likelihood with speech characteristic extraction method. It is used a method of extracting a speech feature based on the hidden markov model. It can improve the accuracy of inaccurate speech model for speech models been produced at the existing system with the use of the proposed system may constitute a robust model for speech recognition. The proposed method shows the improved recognition accuracy in a speech recognition system.

• Atmosphere
• /
• v.29 no.3
• /
• pp.269-282
• /
• 2019

While it is important to obtain the accurate information on snowfall data due to the increase in damage caused by the heavy snowfall in the winter season, it is not easy to observe the snowfall quantitatively. Recently, snow measurements using a weighing precipitation gauge have been carried out, but there is a problem that high snowfall intensity results in low accuracy. Also, the observed snowfall data are sensitive depending on wind speed, temperature, and humidity. In this study, a new process of quality control for snow water equivalent (SWE) data of the weighing precipitation gauge were proposed to cover the low accuracy of snow data and maximize the data utilization. Snowfall data (SWE) observed by Pluvio, Parsivel, snow-depth meter using laser or ultrasonic, and rainfall gauge in Cloud Physics Observation Site (CPOS) were compared and analyzed. Applying the QC algorithm including the use of number of hydrometeor particles as reference, the increased SWE per the unit time was determined and the data noise was removed and marked by flag. The SWE data converted by the number concentration of hydrometeor particles are tested as a method to restore the QC-removed data, and show good agreement with those of the weighing precipitation gauge, though requiring more case studies. The three events data for heavy snowfall disaster in Pyeongchang area was analyzed. The SWE data with improved quality was showed a good correlation with the eye-measured data ($R^2$ > 0.73).

• Structural Engineering and Mechanics
• /
• v.71 no.4
• /
• pp.391-405
• /
• 2019

Compared to the ambient vibration test mainly identifying the structural modal parameters, such as frequency, damping and mode shapes, the impact testing, which benefits from measuring both impacting forces and structural responses, has the merit to identify not only the structural modal parameters but also more detailed structural parameters, in particular flexibility. However, in traditional impact tests, an impacting hammer or artificial excitation device is employed, which restricts the efficiency of tests on various bridge structures. To resolve this problem, we propose a new method whereby a moving vehicle is taken as a continuous exciter and develop a corresponding flexibility identification theory, in which the continuous wheel forces induced by the moving vehicle is considered as structural input and the acceleration response of the bridge as the output, thus a structural flexibility matrix can be identified and then structural deflections of the bridge under arbitrary static loads can be predicted. The proposed method is more convenient, time-saving and cost-effective compared with traditional impact tests. However, because the proposed test produces a spatially continuous force while classical impact forces are spatially discrete, a new flexibility identification theory is required, and a novel structural identification method involving with equivalent load distribution, the enhanced Frequency Response Function (eFRFs) construction and modal scaling factor identification is proposed to make use of the continuous excitation force to identify the basic modal parameters as well as the structural flexibility. Laboratory and numerical examples are given, which validate the effectiveness of the proposed method. Furthermore, parametric analysis including road roughness, vehicle speed, vehicle weight, vehicle's stiffness and damping are conducted and the results obtained demonstrate that the developed method has strong robustness except that the relative error increases with the increase of measurement noise.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.23 no.9
• /
• pp.1060-1069
• /
• 2019

Pedestrian-based camera self-calibration methods are suitable for video surveillance systems since they do not require complex calibration devices or procedures. However, using arbitrary pedestrians as calibration targets may result in poor calibration accuracy due to the unknown height of each pedestrian. To solve this problem in the real surveillance environments, this paper proposes a novel Bayesian approach. By assuming known statistics on the height of pedestrians, we construct a probabilistic model that takes into account uncertainties in both the foot/head locations and the pedestrian heights, using foot-head homology. Since solving the model directly is infeasible, we use variational Bayesian inference, an approximate inference algorithm. Accordingly, this makes it possible to estimate the height of pedestrians and to obtain accurate camera parameters simultaneously. Experimental results show that the proposed algorithm is robust to noise and provides accurate confidence in the calibration.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.20 no.5
• /
• pp.87-92
• /
• 2019

This study is the analysis of causes of printed circuit board (PCB) in collective protection equipment failure and quality improvement. The equipment is a component of the weapon system currently in operation and serves to defend against enemy chemical and biological attack as well as heating and cooling functions. However, during operation in the military, fans of condensate assembly failed to operate. The cause of the failure is the burning of PCB. It was found that the parts were heated according to the continuous cooling operation under the high temperature environmental conditions. Accordingly, the electronic components exposed to high temperature were deteriorated and destroyed. To solve this problem, PCB apply to heatsink. The performance test of improved PCB has been completed. Futhermore system compatibility, positive pressure maintenance and noise test were performed. This improvement confirmed that no faults have occurred in PCB so far. Therefore, the quality of the equipment has improved.

• The Journal of the Acoustical Society of Korea
• /
• v.38 no.2
• /
• pp.168-176
• /
• 2019

When an array receives a signal with a frequency higher than the design frequency, there is an ambiguity in beamforming due to spatial aliasing. In order to overcome this problem, Abadi proposed frequency-difference beamforming. However, there is a constraint that the minimum frequency bandwidth is required according to the value of the difference frequency. In this paper, we propose a method to find the direction of the target signal with spatial aliasing based on the frequency-wavenumber spectrum combined with Radon transform. The proposed method can estimate the direction of the target without ambiguities when the signal has nonnegligible bandwidth. We tested the algorithm by simulating a broadband signal and verified the results with the frequency-difference beamforming method using SAVEX15 (Shallow Water Acoustic Variability EXperiment 2015)'s shrimp noise data.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.12 no.12
• /
• pp.5860-5876
• /
• 2018

A single snapshot data can only provide limited amount of information so that the rank of covariance matrix is not full, which is not adopted to complete the parameter estimation directly using the traditional super-resolution method. Aiming at solving the problem, a joint time delay and angle estimation using matrix pencil method based on information reconstruction vector for orthogonal frequency division multiplexing (OFDM) signal is proposed. Firstly, according to the channel frequency response vector of each array element, the algorithm reconstructs the vector data with delay and angle parameter information from both frequency and space dimensions. Then the enhanced data matrix for the extended array element is constructed, and the parameter vector of time delay and angle is estimated by the two-dimensional matrix pencil (2D MP) algorithm. Finally, the joint estimation of two-dimensional parameters is accomplished by the parameter pairing. The algorithm does not need a pseudo-spectral peak search, and the location of the target can be determined only by a single receiver, which can reduce the overhead of the positioning system. The theoretical analysis and simulation results show that the estimation accuracy of the proposed method in a single snapshot and low signal-to-noise ratio environment is much higher than that of Root Multiple Signal Classification algorithm (Root-MUSIC), and this method also achieves the higher estimation performance and efficiency with lower complexity cost compared to the one-dimensional matrix pencil algorithm.

• Journal of the Korea Convergence Society
• /
• v.10 no.5
• /
• pp.17-26
• /
• 2019

We aimed to build a patient-based allergy prevention system using the smartphone and focused on the region of interest (ROI) extraction method for Optical Character Recognition (OCR) in the general environment. However, the current ROI extraction method has shown good performance in the experimental environment, but the performance in the real environment was not good due to the noisy background. Therefore, in this paper, we propose the compared methods of reducing noisy background to solve the ROI extraction problem. There five methods used as a SMF, DIN, Denoising Autoencoder(DAE), DAE with Convolution Neural Network(DAECNN) and median filter(MF) with DAECNN (MF+DAECNN). We have shown that our proposed DAECNN and MF+DAECNN methods are 69%, respectively, which is relatively higher than the conventional DAE method 55%. The verification of performance improvement uses MSE, PSNR and SSIM. The system has implemented OpenCV, C++ and Python, including its performance, is tested on real images.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.10
• /
• pp.3729-3749
• /
• 2021

At present, deep convolution network-based salient object detection (SOD) has achieved impressive performance. However, it is still a challenging problem to make full use of the multi-scale information of the extracted features and which appropriate feature fusion method is adopted to process feature mapping. In this paper, we propose a new adjacency auxiliary network (AANet) based on multi-scale feature fusion for SOD. Firstly, we design the parallel connection feature enhancement module (PFEM) for each layer of feature extraction, which improves the feature density by connecting different dilated convolution branches in parallel, and add channel attention flow to fully extract the context information of features. Then the adjacent layer features with close degree of abstraction but different characteristic properties are fused through the adjacent auxiliary module (AAM) to eliminate the ambiguity and noise of the features. Besides, in order to refine the features effectively to get more accurate object boundaries, we design adjacency decoder (AAM_D) based on adjacency auxiliary module (AAM), which concatenates the features of adjacent layers, extracts their spatial attention, and then combines them with the output of AAM. The outputs of AAM_D features with semantic information and spatial detail obtained from each feature are used as salient prediction maps for multi-level feature joint supervising. Experiment results on six benchmark SOD datasets demonstrate that the proposed method outperforms similar previous methods.