• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2022.06a
• /
• pp.323-324
• /
• 2022

복잡한 항만해역에서 자율운항선박 입출항 지원을 위해서는 항내에서 통항하는 선박에 대한 지속적인 인식이 필요하며, 이를 위해 선박자동식별시스템(AIS)에서 송출된 정보를 기반으로 선박의 운동정보(위치/침로/속도), 식별번호 및 크기 정보를 확인한다. 하지만, AIS 탑재 의무가 있는 선박에 대한 정보만 취득이 가능하기 때문에, 보조적으로 육상레이더를 활용하여 AIS 정보로부터 식별이 안되는 선박을 인식할 수 있는 기술이 필요하다. 본 연구에서는 자율운항선박 입출항 지원을 위해 레이더 이미지를 활용하여 선박의 운동정보와 형상정보를 동시에 추정할 수 있는 알고리즘을 설계하였다.

• Journal of the Korea Society for Simulation
• /
• v.29 no.4
• /
• pp.21-31
• /
• 2020

NCTR(Non-Cooperative Target Recognition) refers to the function of radar to identify target on its own without support from other systems such as ELINT(ELectronic INTelligence). ISAR(Inverse Synthetic Aperture Radar) image is one of the representative methods of NCTR, but it is difficult to automatically classify the target without an identification database due to the significant changes in the image depending on the target's maneuver and location. In this study, we discuss how to build an identification database using simulation and deep-learning technique even when actual images are insufficient. To simulate ISAR images changing with various radar operating environment, A model that generates and learns images through the process named 'Perfect scattering image,' 'Lost scattering image' and 'JEM noise added image' is proposed. And the learning outcomes of this model show that not only simulation images of similar shapes but also actual ISAR images that were first entered can be classified.

• Annual Conference of KIPS
• /
• 2020.05a
• /
• pp.5-8
• /
• 2020

매우 작은 크기의 센서들이 산악 등의 험지에 흩뿌려지는 스마트 더스트 환경은 장치들의 컴퓨팅 성능과 리소스가 매우 제한되기 때문에 각 센서들의 위치를 식별하기 매우 힘들다. 또한 초대량의 센서들이 뿌려지는 특성으로 인해 수집, 전송되는 데이터의 크기가 상상하기 힘들 정도로 커질 수 있다. 본 논문에서는 중간 매개 역할을 수행하는 디바이스의 위치와 삼변측량을 이용해 센서들의 위치를 계산하고 계산된 위치를 기반으로 동종의 센서에서 수집된 데이터를 축소, 통합하는 위치 기반 데이터 축소 시스템을 제안한다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.13 no.5
• /
• pp.887-898
• /
• 2009

Conventional routing protocols proposed for wireless sensor networks (WSNs) cannot fully accommodate the characteristics of WSNs. In particular, although it is possible to largely obtain benefits in the solution of energy consumption and global identification problems through applying position information, there are few protocols that actively apply such position information. In the case of geographical and energy aware routing (GEAR) that is a typical algorithm, which uses position information, it does not fully represent the characteristics of WSNs because it is limited to forward query messages and assumed as fixed network environments. The routing protocols proposed in this paper defines the direction of data, which is routed based on the position information of individual and target nodes, in which each node configures its next hop based on this direction and routes signals. Because it performs data-centric routing using position information, it does not require certain global identifications in order to verify individual nodes and is able to avoid unnecessary energy consumption due to the forwarding of packets by defining its direction.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2007.06a
• /
• pp.721-723
• /
• 2007

Recently, the size and location of the acquired container image while the container is loading and unloading in Harbors is not fixed. And it is difficult to get a good image for recognition because of the variation of external environment as those the size of container and where the yard-tractor stop is. In this paper, we estimate where the container ISO-code set is using Top-hat transform from realtime images and get an image to recognize container ISO-code using PAN/TILT/ZOOM camera. We extract the container ISO-code using Top-hat transform and Histogram projection. After binarization, we extract each character from complex background using labeling. We use BP(Backpropagation Network) to recognize extracted characters.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.25 no.4
• /
• pp.459-466
• /
• 2014

This paper deals with the method to classify the aircraft target by analyzing its JEM signal. We propose the method to classify the engine model by analyzing JEM spectrum using the harmonic frequency mask generated from the blade information of jet engine. The proposed method does not need the complicated logic algorithm to find the chopping frequency in each rotor stage and the pre-simulated engine spectrum DB used in the previous methods. In addition, we propose the method to estimate the precise spool rate and it reduces the error in estimating the number of blades or in calculating the harmonic frequency of frequency mask.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2015.05a
• /
• pp.586-589
• /
• 2015

A computer presentation system by using hand-motion identification rules is proposed. To identify hand motions of a presenter, a face region is extracted first using haar classifier. A motion status(patterns) and position of hands is discriminated using the center of gravities of user's face and hand after segmenting the hand area on the YCbCr color model. User's hand is applied to the motion detection rules and then presentation control command is then executed. The proposed system utilizes the motion identification rules without the use of additional equipment and it is then capable of controlling the presentation and does not depend on the complexity of the background. The proposed algorithm confirmed the stable control operation via the presentation of the experiment in the dark illumination range of indoor atmosphere (lx) 15-20-30.

• Journal of the Korean Wood Science and Technology
• /
• v.35 no.6
• /
• pp.73-82
• /
• 2007

A large underground wooden structure of Shilla period which was presumably built in the 5th century A.D. was excavated at Gomo Sanseong Fortress, Mungyeong, Gyeongsangbuk-do. Of 12 wood member samples obtained on the site, 8 hardwoods under family Fagaceae and 4 softwoods under family Pinaceae were separated through light microscopy. Among hardwoods, 5 were found to be Cerris section, 2 Prinus section, and 1 Castanea crenata. On the other hand, 4 softwoods were all identified as Finns densiflora.

• Journal of IKEEE
• /
• v.22 no.4
• /
• pp.1044-1049
• /
• 2018

CNN (Convolutional Neural Networks) is a neural network that models animal visual information processing. And it shows good performance in various fields. In this paper, we use CNN to classify target and non-target data by analyzing the spectrogram of active sonar signal. The data were divided into 8 classes according to the ratios containing the targets and used for learning CNN. The spectrogram of the signal is divided into frames and used as inputs. As a result, it was possible to classify the target and non-target using the characteristic that the classification results of the seven classes corresponding to the target signal sequentially appear only at the position of the target signal.

• Proceedings of the Korea Contents Association Conference
• /
• 2010.05a
• /
• pp.336-338
• /
• 2010

장문인식은 손바닥 중앙부에 나타난 손금과 주름의 패턴을 이용하여 개인을 식별하는 것으로, 효과적인 장문인식을 위해서는 이러한 패턴이 나타나는 관심영역(ROI: region of interest)에 대한 안정적인 추출이 필요하다. 본 논문에서는 윤곽선의 형태 정보를 토대로 적응적으로 굴곡점의 위치를 찾아내고 이로부터 ROI를 추출하는 방법을 제안한다. 제안된 방법의 성능을 확인하기 위하여 유도 막대가 없는 자연스런 장문획득 장치에 의해 수집된 장문영상을 대상으로 실험을 수행하였다. 실험결과 제안된 방법은 손의 위치 변화나 회전에 무관하게 장문영상으로부터 안정적으로 ROI를 추출함을 확인할 수 있었다.