• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.12
• /
• pp.1835-1845
• /
• 2021

To accurately identify black ice and warn the drivers of information in advance so they can control speed and take preventive measures. In this paper, we propose a lightweight black ice detection network based on infrared road images. A black ice recognition network model based on CNN transfer learning has been developed. Additionally, to further improve the accuracy of black ice recognition, an enhanced lightweight network based on MobileNetV2 has been developed. To reduce the amount of calculation, linear bottlenecks and inverse residuals was used, and four bottleneck groups were used. At the same time, to improve the recognition rate of the model, each bottleneck group was connected to a 3×3 convolutional layer to enhance regional feature extraction and increase the number of feature maps. Finally, a black ice recognition experiment was performed on the constructed infrared road black ice dataset. The network model proposed in this paper had an accurate recognition rate of 99.07% for black ice.

• Korean Journal of Remote Sensing
• /
• v.38 no.1
• /
• pp.111-126
• /
• 2022

Land use and land cover (LULC) mapping is an important factor in geospatial analysis. Although highly precise ground-based LULC monitoring is possible, it is time consuming and costly. Conversely, because the synthetic aperture radar (SAR) sensor is an all-weather sensor with high resolution, it could replace field-based LULC monitoring systems with low cost and less time requirement. Thus, LULC is one of the major areas in SAR applications. We developed a LULC model using only KOMPSAT-5 single co-polarized data and digital elevation model (DEM) data. Twelve HH-polarized images and 18 VV-polarized images were collected, and two HH-polarized images and four VV-polarized images were selected for the model testing. To train the LULC model, we applied the conditional generative adversarial network (cGAN) method. We used U-Net combined with the residual unit (ResUNet) model to generate the cGAN method. When analyzing the training history at 1732 epochs, the ResUNet model showed a maximum overall accuracy (OA) of 93.89 and a Kappa coefficient of 0.91. The model exhibited high performance in the test datasets with an OA greater than 90. The model accurately distinguished water body areas and showed lower accuracy in wetlands than in the other LULC types. The effect of the DEM on the accuracy of LULC was analyzed. When assessing the accuracy with respect to the incidence angle, owing to the radar shadow caused by the side-looking system of the SAR sensor, the OA tended to decrease as the incidence angle increased. This study is the first to use only KOMPSAT-5 single co-polarized data and deep learning methods to demonstrate the possibility of high-performance LULC monitoring. This study contributes to Earth surface monitoring and the development of deep learning approaches using the KOMPSAT-5 data.

• Journal of Digital Policy
• /
• v.3 no.1
• /
• pp.27-32
• /
• 2024

In this paper, proposes a super resolution method that enhances the quality of results by refining texture features, contrasting each, and utilizing the results as weights. For the improvement of quality, a precise and clear restoration result in details such as boundary areas is crucial in super resolution, along with minimizing unnecessary artifacts like noise. The proposed method constructs a residual block structure with multiple paths and skip-connections for feature estimation in conventional Convolutional Neural Network (CNN)-based super resolution methods to enhance quality. Additional learning is performed for sharpened and blurred image results for further texture analysis. By contrasting each super resolution result and allocating weights through this process, the proposed method achieves improved quality in detailed and smoothed areas of the image. The experimental results of the proposed method, evaluated using the PSNR and SSIM values as quality metrics, show higher results compared to existing algorithms, confirming the enhancement in quality.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.14 no.2
• /
• pp.907-923
• /
• 2020

For the severe haze situation in the Beijing-Tianjin-Hebei region, conventional fine particulate matter (PM2.5) concentration prediction methods based on pollutant data face problems such as incomplete data, which may lead to poor prediction performance. Therefore, this paper proposes a method of predicting the PM2.5 concentration based on image analysis technology that combines image data, which can reflect the original weather conditions, with currently popular machine learning methods. First, based on local parameter estimation, autoregressive (AR) model analysis and local estimation of the increase in image blur, we extract features from the weather images using an approach inspired by free energy and a no-reference robust metric model. Next, we compare the coefficient energy and contrast difference of each pixel in the AR model and then use the percentages to calculate the image sharpness to derive the overall mass fraction. Furthermore, the results are compared. The relationship between residual value and PM2.5 concentration is fitted by generalized Gauss distribution (GGD) model. Finally, nonlinear mapping is performed via the wavelet neural network (WNN) method to obtain the PM2.5 concentration. Experimental results obtained on real data show that the proposed method offers an improved prediction accuracy and lower root mean square error (RMSE).

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.10 no.4
• /
• pp.736-739
• /
• 2009

In this paper, an efficient lossless compression algorithm using motion adaptation is proposed. It is divided into two parts: a motion adaptation based nonlinear predictor part and a residual data coding part. The proposed nonlinear predictor can reduce prediction error by learning from its past prediction errors using motion adaption. The predictor decides the proper selection of the intra and inter prediction values according to the past prediction error. The reduced error is coded by existing context adaptive coding method. Experimental results show that the proposed algorithm has the higher compression ratio than context modeling methods, such as FELICS, CALIC, and JPEG-LS.

• Steel and Composite Structures
• /
• v.24 no.5
• /
• pp.635-641
• /
• 2017

Safety service improvement and development of efficient maintenance strategies for corroded steel structures are undeniably essential. Therefore, understanding the influence of damage caused by corrosion on the remaining load-carrying capacities such as tensile strength is required. In this study, artificial neural network (ANN) approach is proposed in order to produce a simple, accurate, and inexpensive method developed by using tensile test results, material properties and finite element method (FEM) results to train the ANN model. Initially in reproducing corroded model process, FEM was used to obtain tensile strength of artificial corroded plates, for which surface is developed by a spatial autocorrelation model. By using the corroded surface data and material properties as input data, with tensile strength as the output data, the ANN model could be trained. The accuracy of the ANN result was then verified by using leave-one-out cross-validation (LOOCV). As a result, it was confirmed that the accuracy of the ANN approach and the final output equation was developed for predicting tensile strength without tensile test results and FEM in further work. Though previous studies have been conducted, the accuracy results are still lower than the proposed ANN approach. Hence, the proposed ANN model now enables us to have a simple, rapid, and inexpensive method to predict residual tensile strength more accurately due to corrosion in steel structures.

• ETRI Journal
• /
• v.42 no.2
• /
• pp.230-238
• /
• 2020

Human-object interaction (HOI) detection is a popular computer vision task that detects interactions between humans and objects. This task can be useful in many applications that require a deeper understanding of semantic scenes. Current HOI detection networks typically consist of a feature extractor followed by detection layers comprising small filters (eg, 1 × 1 or 3 × 3). Although small filters can capture local spatial features with a few parameters, they fail to capture larger context information relevant for recognizing interactions between humans and distant objects owing to their small receptive regions. Hence, we herein propose a three-stream HOI detection network that employs a context convolution module (CCM) in each stream branch. The CCM can capture larger contexts from input feature maps by adopting combinations of large separable convolution layers and residual-based convolution layers without increasing the number of parameters by using fewer large separable filters. We evaluate our HOI detection method using two benchmark datasets, V-COCO and HICO-DET, and demonstrate its state-of-the-art performance.

• Journal of Korean Institute of Industrial Engineers
• /
• v.42 no.4
• /
• pp.263-269
• /
• 2016

The purpose of this study is to challenge a computational regression-type problem, that is handling large-size data, in which conventional metamodeling techniques often fail in a practical sense. To solve such problems, regression-type boosting, one of ensemble model techniques, together with bootstrapping-based re-sampling is a reasonable choice. This study suggests weight updates by the amount of the residual itself and a new error decision criterion which constructs an ensemble model of models selectively chosen by rejection limits. Through these ideas, we propose AdaBoost.RMU.R as a metamodeling technique suitable for handling large-size data. To assess the performance of the proposed method in comparison to some existing methods, we used 6 mathematical problems. For each problem, we computed the average and the standard deviation of residuals between real response values and predicted response values. Results revealed that the average and the standard deviation of AdaBoost.RMU.R were improved than those of other algorithms.

• Journal of information and communication convergence engineering
• /
• v.20 no.4
• /
• pp.259-264
• /
• 2022

In wireless sensor networks, sensor nodes are often deployed in large numbers in places that are difficult for humans to access. However, the energy of the sensor node is limited. Therefore, one of the most important considerations when designing routing protocols in wireless sensor networks is minimizing the energy consumption of each sensor node. When the energy of a wireless sensor node is exhausted, the node can no longer be used. Various protocols are being designed to minimize energy consumption and maintain long-term network life. Therefore, we proposed KOCED, an optimal cluster K-means algorithm that considers the distances between cluster centers, nodes, and residual energies. I would like to perform a performance evaluation on the KOCED protocol. This is a study for energy efficiency and validation. The purpose of this study is to present performance evaluation factors by comparing the K-means algorithm and the K-medoids algorithm, one of the recently introduced machine learning techniques, with the KOCED protocol.

• Proceedings of the Korean Institute of Building Construction Conference
• /
• 2022.11a
• /
• pp.109-110
• /
• 2022

Currently, interest in the quality of concrete is increasing. Among the important factors for evaluating the quality of concrete, interest in unit-water content is also increasing. Currently, the air-meter method, the microwave oven drying method, the capacitance method, and the microwave penetration method are used to measure the unit-water content of concrete.. Among the above methods, except for the microwave method, the measurement method is complicated, portability is reduced, and economic efficiency is reduced. This research aims to measure a unit-water content by using a Frequency Domain Reflectometry(FDR) sensor that is economical, simple to measure, and portable among microwave methods. In addition, it is an experimental study to determine the accuracy of unit-water content using a single input residual model during deep learning to solve the limitations of the FDR sensor.