• Journal of Korea Multimedia Society
• /
• v.22 no.12
• /
• pp.1457-1465
• /
• 2019

3D convolution is to stack multiple consecutive frames to form a cube, and then apply the 3D convolution kernel in the cube. In this structure, each feature map of the convolutional layer is connected to multiple adjacent sequential frames in the previous layer, thus capturing the motion information. However, due to the changes of pedestrian posture, motion and position, the convolution at the same place is inappropriate, and when the 3D convolution kernel is convoluted in the time domain, only time domain features of three consecutive frames can be extracted, which is not a good enough to get action information. This paper proposes an action recognition method based on feature fusion of 3D convolutional neural network. Based on the VGG16 network model, sending a pre-acquired optical flow image for learning, then get the time domain features, and then the feature of the time domain is extracted from the features extracted by the 3D convolutional neural network. Finally, the behavior classification is done by the SVM classifier.

• Journal of the Korean Society of Industry Convergence
• /
• v.23 no.2_2
• /
• pp.173-180
• /
• 2020

In Korea, due to the rapid increase in population due to industrialization, rivers were developed and managed with a focus on the completion and dimension of rivers. Due to the rapid increase in river use, there are so many river facilities indiscriminately that the administrative and management tasks are complicated and diversified in computerizing facilities and hydrologic data. Many methods have been proposed to solve this problem, but many problems exist. Among them, water has the same spectral characteristics, so it is difficult to subdivide into rivers, reservoirs, and dams. Therefore, this study subdivided the water system using supervised classification and GIS in order to efficiently manage the water resources by classifying the water system with the same spectral characteristics. In order to analyze the accuracy of the results, the accuracy of the objects classified using land cover map provided by environmental spatial information service was evaluated, and the result was an average of 91.75%, with 97.50% of rivers, 87.76% of reservoirs, and 90.00% of others.

• Journal of The Korean Society of Agricultural Engineers
• /
• v.50 no.3
• /
• pp.61-69
• /
• 2008

Land-use in northern Cheongju region is changing rapidly because of the increased interactions of human activities with the environment as population increases. Land-use change detection is considered essential for monitoring the growth of an urban complex. The analysis was undertaken mainly on the basis of the multi-temporal Landsat images (1991, 1992 and 2000) and DEM data in a post-classification analysis with GIS to map land-use distribution and to analyse factors influencing the land-use changes for Cheongju city. The area of each land-use category was also calculated for monitoring land-use changes. Land-use statistics revealed that substantial land-use changes have taken place and that the built-up areas have expanded by about $17.57km^2$ (11.47%) over the study period (1991 - 2000). This study illustrated an increasing trend of urban and barren lands areas with a decreasing trend of agricultural and forest areas. Land-use changes from one category to others have been clearly represented by the NDVI composite images, which were found suitable for delineating the development of urban areas and land use changes in northern Cheongju region. Rapid economic developments together with the increasing population were noted to be the major factors influencing rapid land use changes. Urban expansion has replaced urban and barren lands.

• Proceedings of the KIEE Conference
• /
• 2005.11b
• /
• pp.162-164
• /
• 2005

This paper presents an hour-ahead System Marginal Price (SMP) forecasting framework based on a neural network. Recently, the deregulation in power industries has impacted on the power system operational problems. The bidding strategy of market participants in energy market is highly dependent on the short-term price levels. Therefore, short-term SMP forecasting is a very important issue to market participants to maximize their profits. and to market operator who may wish to operate the electricity market in a stable sense. The proposed hybrid neural network is composed of tow parts. First part of this scheme is pattern classification to input data using Kohonen Self-Organizing Map (SOM) and the second part is SMP forecasting using back-propagation neural network that has three layers. This paper compares the forecasting results using classified input data and unclassified input data. The proposed technique is trained, validated and tested with historical date of Korea Power Exchange (KPX) in 2002.

• Spatial Information Research
• /
• v.8 no.1
• /
• pp.101-116
• /
• 2000

Seoul Metropolitan Government(SMG) as well as the other local governments should have a role in managing groundwater according to the Groundwater Law. As a guidance of the management and control of groundwater use, this study emphasizes the necessity of the construction of 1 ; 5,000 hydrogeological maps and proposes possble usages of the maps. As expected to be a part of the groundwater GIS database constructed by SMG, the maps can be used as base maps to facilitate the whole processes related to the groundwater development and preservation. For the purpose of the construction of the 1 ; 5,000 hydrogelogical GIS database, we develop the classification system of the hydro-geological entities and their attributes.

• Journal of the Korea Institute of Military Science and Technology
• /
• v.12 no.4
• /
• pp.516-523
• /
• 2009

Recently, the applications of unmanned robots are increasing in various fields including surveillance and reconnaissance, planet exploration and disaster relief. To perform their missions with success, the robots should be able to evaluate terrain's characteristics quantitatively and identify traversable regions to progress toward a goal using mounted sensors. Recently, the authors have proposed techniques that extract terrain information and analyze traversability for off-road navigation of an unmanned robot. In this paper, we examine the use of 3D world models(terrain maps) to classify obstacle and safe terrain for increasing the reliability of the proposed techniques. A world model is divided into several patches and each patch is classified as belonging either to an obstacle or a non-obstacle using three types of metrics. The effectiveness of the proposed method is verified on real terrain maps.

• IEIE Transactions on Smart Processing and Computing
• /
• v.6 no.4
• /
• pp.229-236
• /
• 2017

This paper presents an efficient object-based color image-retrieval algorithm that is suitable for the classification and retrieval of images from small to mid-scale datasets, such as images in PCs, tablets, phones, and cameras. The proposed method first finds salient regions by using regional feature vectors, and also finds several dominant colors in each region. Then, each salient region is partitioned into small sub-blocks, which are assigned 1 or 0 with respect to the number of pixels corresponding to a dominant color in the sub-block. This gives a binary map for the dominant color, and this process is repeated for the predefined number of dominant colors. Finally, we have several binary maps, each of which corresponds to a dominant color in a salient region. Hence, the binary maps represent the spatial distribution of the dominant colors in the salient region, and the union (OR operation) of the maps can describe the approximate shapes of salient objects. Also proposed in this paper is a matching method that uses these binary maps and which needs very few computations, because most operations are binary. Experiments on widely used color image databases show that the proposed method performs better than state-of-the-art and previous color-based methods.

• Journal of Ecology and Environment
• /
• v.35 no.3
• /
• pp.167-175
• /
• 2012

Coastal sand dune systems are particularly fragile and threaten the environment. However, these systems provide fundamental ecosystem services to the nearby urban areas, acting, for example, as protective buffers against erosion. In this paper, we attempt to classify the biotope types of coastal sand dune ecosystems and select an index for the assessment of the conservation value. The types of biotopes are categorized based on the vegetation map; floras are examined in order to research the effects of hinterlands on coastal sand dunes. In addition, a naturalization rate and an urbanization index for each biotope type in hinterlands are analyzed. In the ecosystem of coastal sand dunes, the urbanization index and naturalization rate shows a higher value in sand dunes with areas of road, residential, and idle land in farm villages, rice fields, and fields. On the contrary, a lower value in the urbanization index and naturalization rate is present when typical biotope types, such as sand dune vegetation and natural Pinus thunbergii forests, are widely distributed. Based on these results, urbanization index and naturalization rate should be used as critical indices for the assessment of the ecosystem of costal sand dunes.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.24 no.11
• /
• pp.2875-2882
• /
• 2000

Typically, the ultrasonic sensors can be used in navigation systems for modeling of the enviornment, obstacle avoidance, and map building. In this paper, we tried to approach an object classification method using the range data of the ultrasonic sensors. A characterization of the sonar scan is described that allows the differentiation of planes, corners, edges, cylindrical and rectangular pillars by processing the scanned data from three sonars. To use the data from the ultrasonic sensors as input to the neural networks, we have introduced a clustering, threshold, and bit operation algorithm for the obtained raw data, After repeated training of the neural network, the performance of the proposed method was obtained through experiments. Also, the recognition ranges of the proposed method were investigated. As a result of experiments, we found that the proposed method successfully recognized the objects within the accuracy of 78%.

• Korean Institute of Interior Design Journal
• /
• v.13 no.6
• /
• pp.132-139
• /
• 2004

This study is to a question in argument that existing theses about a trait spatial configuration of exhibition space were analyzed without appropriateness verification of analysis characteristics. Firstly, through theoretical studies of established thesis, validity twenty analysis characteristics was chosen by making an investigation into existing analysis characteristics. Secondly, through a subject of our investigation, forty-two exhibition space of nineteen museums and art museum at home and abroad, a distribution map of exhibition space was analyzed by multivariate analysis. As a result of this study : 1) Nine analysis characteristics which extracted through multivariate analysis was the principal analysis characteristics. 2) A scale was important characteristic for the classification of museum therefore a degree of space perception was ought to compare every one of similar scale museum. 3) When comparing a trait of spatial configuration at exhibition space, these characteristics came into effect on middle sized museums. 4) It was visually confirmed a trait of spatial configuration of each group between museum and art museum