• 한국환경복원기술학회지
• /
• 제11권3호
• /
• pp.94-106
• /
• 2008

As populations inhabiting in natural ecosystem are fragmented by artificial barriers and habitats are destructed by development, extinction possibility of species is getting higher. It is necessary to design and to manage conservation areas and corridors considering animals' movement and migration for sustainable species diversity in present circumstances. 'Least-cost modeling' is one commonly employed approach in which dispersal costs are assigned to distinct habitat types and the last-costly dispersal paths among habitat patches are calculated using a geographical information system (GIS). This study aims to design ecological corridor using least-cost path method and to apply it to a regional ecological network considering movability of medium-large size mammals. This study was carried out over the metropolitan area, which has been deforested by rapid urbanization. Nevertheless there is connected with Gangwon province, Baekdudaegan mountain range and DMZ, considered where many forest species can migrate to this region. This study employs such an approach to develop least-cost path models for medium-large size mammals, have inhabited for this entire region. Considering those species, two forest areas as a source of species supply and forest areas more than 1,000ha are selected as focal forest areas. Movement and migration paths from species supply sources to focal forest areas are calculated by applying landscape permeability theory using land cover map, road density map and land slope map. Results showed least-cost paths from species supply sources to focal forest areas on two species. Wildcat and roe deer are different in some least-cost paths caused by their landscape permeability but paths show generally same specifics. The result of considering regional distribution of expected movement and migration paths to regional ecological network, low altitude mountains of western metropolitan area are evaluated important area for species connectivity. In national or regional levels ecological connectivity is essential to promote species diversity and to preserve integrated ecosystem. This study concludes that developing least-cost models from similar empirical data could significantly improve the utility of these tools.

• 한국측량학회지
• /
• 제36권5호
• /
• pp.423-432
• /
• 2018

Surface water mapping has been widely used in various remote sensing applications. Water indices have been commonly used to distinguish water bodies from land; however, determining the optimal threshold and discriminating water bodies from similar objects such as shadows and snow is difficult. Deep learning algorithms have greatly advanced image segmentation and classification. In particular, FCN (Fully Convolutional Network) is state-of-the-art in per-pixel image segmentation and are used in most benchmarks such as PASCAL VOC2012 and Microsoft COCO (Common Objects in Context). However, these data sets are designed for daily scenarios and a few studies have conducted on applications of FCN using large scale remotely sensed data set. This paper aims to fine-tune the pre-trained FCN network using the CRMS (Coastwide Reference Monitoring System) data set for surface water mapping. The CRMS provides color infrared aerial photos and ground truth maps for the monitoring and restoration of wetlands in Louisiana, USA. To effectively learn the characteristics of surface water, we used pre-trained the DeepWaterMap network, which classifies water, land, snow, ice, clouds, and shadows using Landsat satellite images. Furthermore, the DeepWaterMap network was fine-tuned for the CRMS data set using two classes: water and land. The fine-tuned network finally classifies surface water without any additional learning process. The experimental results show that the proposed method enables high-quality surface mapping from CRMS data set and show the suitability of pre-trained FCN networks using remote sensing data for surface water mapping.

• 한국정보통신학회논문지
• /
• 제11권1호
• /
• pp.206-215
• /
• 2007

Ad-Hoc 네트워크 기술이 미래의 이동 인터넷 기술로서 이동통신망(Mobile Network) 뿐만 아니라 공중무선랜망(WPAN) 그리고 유비쿼터스 망 등에 광범위하게 활용되기 위해서는 개선 및 보완되어야 할 기술적인 문제들이 많다. 특히, 최근 네트워크 보안의 허점을 이용한 보안상 공격이 급증하고 있는 상황에 반하여 Ad-Hoc 라우팅 프로토콜 연구 대부분은 보안 위협 요소를 배제하고 안전한 환경을 가정한 채 수행되고 있다. 또한 Ad-Hoc 네트워크가 무선 매체 특성상 더욱 많은 보안상 위협에 노출되기 쉽고 유선에서 사용되던 보안 메커니즘을 그대로 적용되는 것이 부적합함을 고려할 때 Ad-Hoc 보안에 관한 연구는 더욱 활발히 이루어져야 할 것이다. 따라서 본 논문에서는 Ad-Hoc 네트워크의 특성을 고려하여 중앙 집중적인 인증기관이나 키 분배센터에 의존하지 않고 클러스터를 구성하고 있는 멤버 노드들 간의 협력적이고, 분산된 인증서를 이용한 인증 서비스를 제공할 수 있는 모델을 제안한다. 아울러 시뮬레이션을 통해 제안한 모델의 확장성과, 견고성을 평가해 본다.

• 정보처리학회논문지:소프트웨어 및 데이터공학
• /
• 제11권11호
• /
• pp.465-472
• /
• 2022

본 연구에서는 기존의 동영상 합성 네트워크에 스타일 합성 네트워크를 접목시켜 동영상에 대한 스타일 합성의 한계점을 극복하고자 한다. 본 논문의 네트워크에서는 동영상 합성을 위해 스타일갠 학습을 통한 스타일 합성과 동영상 합성 네트워크를 통해 스타일 합성된 비디오를 생성하기 위해 네트워크를 학습시킨다. 인물의 시선이나 표정 등이 안정적으로 전이되기 어려운 점을 개선하기 위해 3차원 얼굴 복원기술을 적용하여 3차원 얼굴 정보를 이용하여 머리의 포즈와 시선, 표정 등의 중요한 특징을 제어한다. 더불어, 헤드투헤드++ 네트워크의 역동성, 입 모양, 이미지, 시선 처리에 대한 판별기를 각각 학습시켜 개연성과 일관성이 더욱 유지되는 안정적인 스타일 합성 비디오를 생성할 수 있다. 페이스 포렌식 데이터셋과 메트로폴리탄 얼굴 데이터셋을 이용하여 대상 얼굴의 일관된 움직임을 유지하면서 대상 비디오로 변환하여, 자기 얼굴에 대한 3차원 얼굴 정보를 이용한 비디오 합성을 통해 자연스러운 데이터를 생성하여 성능을 증가시킴을 확인했다.

• 한국환경복원기술학회지
• /
• 제20권2호
• /
• pp.13-35
• /
• 2017

This study was carried out to analyze vegetation structure and change of the disturbed area in forest ecosystem(FE), riparian ecosystem(RE) and coastal ecosystem(CE) related to Regional Ecological Network at Chungnam province through constancy and dominance analysis, species diversity index, similarity index and canonical correlation analysis. Data were collected from April to October in 2015. As a result of constancy and dominance analysis of forest disturbed area, non-effective species(NES) was 30 species(17.0%), exported species(ES) was 98 species(55.7%) and imported species(IS) was 48 species(27.3%) among the total 176 species, respectively. In riparian disturbed area with total 139 species, there were 16 NES(11.5%), 98 ES(70.5%) and 25 IS(18.0%) respectively. In coastal disturbed area with 140 species, there were 20 NES(14.3%), 88 ES(62.9%) and 32 IS(22.9%) respectively. In all types of disturbed areas, the ratio of ES was higher than IS. As a result of species diversity, species richness and shannon's diversity index of disturbed area decreased in all kinds of crown strata such as tree, subtree, shrub and herb layer compared to the control area. As a result of similarity index, that of each type between control site and disturbed site was 0.374 in FE, 0.329 in CE and 0.259 in RE in the order. As a result of the CCA analysis, the number of present species, vine plants ratio and exported species ratio in disturbed area of FE and RE were decreased, and the naturalized plant ratio, imported species ratio and herb ratio were increased. But environmental factors of CE were not shown any clear tendency. In conclusion, many species occupied in control site disappeared into the disturbed area, and the naturalized plants and herb species were abundantly imported in the area. Therefore, it was considered that this study could be applied to the development of long-term and short-term ecological restoration techniques in view of vegetation changes.

• 한국도로학회논문집
• /
• 제20권1호
• /
• pp.35-45
• /
• 2018

PURPOSES : This study examines the performance changes of road networks according to the strength of a disaster, and proposes a method for estimating the quantitative resilience according to the road-network performance changes and damage scale. This study also selected high-influence road sections, according to disasters targeting the road network, and aimed to analyze their hazard resilience from the network aspect through a scenario analysis of the damage recovery after a disaster occurred. METHODS : The analysis was conducted targeting Sejong City in South Korea. The disaster situation was set up using the TransCAD and VISSIM traffic-simulation software. First, the study analyzed how road-network damage changed the user's travel pattern and travel time, and how it affected the complete network. Secondly, the functional aspects of the road networks were analyzed using quantitative resilience. Finally, based on the road-network performance change and resilience, priority-management road sections were selected. RESULTS : According to the analysis results, when a road section has relatively low connectivity and low traffic, its effect on the complete network is insignificant. Moreover, certain road sections with relatively high importance can suffer a performance loss from major damage, for e.g., sections where bridges, tunnels, or underground roads are located, roads where no bypasses exist or they exist far from the concerned road, including entrances and exits to suburban areas. Relatively important roads have the potential to significantly degrade the network performance when a disaster occurs. Because of the high risk of delays or isolation, they may lead to secondary damage. Thus, it is necessary to manage the roads to maintain their performance. CONCLUSIONS : As a baseline study to establish measures for traffic prevention, this study considered the performance of a road network, selected high-influence road sections within the road network, and analyzed the quantitative resilience of the road network according to scenarios. The road users' passage-pattern changes were analyzed through simulation analysis using the User Equilibrium model. Based on the analysis results, the resilience in each scenario was examined and compared. Sections where a road's performance loss had a significant influence on the network were targeted. The study results were judged to become basic research data for establishing response plans to restore the original functions and performance of the destroyed and damage road networks, and for selecting maintenance priorities.

• 대한전자공학회논문지TC
• /
• 제40권10호
• /
• pp.67-78
• /
• 2003

인터넷 트래픽의 폭발적인 증가로 인한 높은 대역폭의 요구와 광 네트워크 기술이 발전되면서 DWDM 네트워크가 국가적 혹은 범세계적인 차세대 광 인터넷(NGOI) 백본망의 대안으로 인식되고 있다. 이러한 DWDM 네트워크 기반의 NGOI에서는 RWA(Routing and Wavelength Assignment) 문제와 생존성이 중요한 이슈가 되고 있다. 특히 높은 데이터 전송율을 가지는 DWDM 네트워크에서 일어나는 짧은 서비스 파괴는 막대한 트래픽 손실을 야기하므로, AOTN에서의 fault/attack 검출, 지역화, 그리고 회복시킴은 가장 중요한 이슈 중 하나가 된다. 본 논문에서는 다양한 광 백본망 소자들의 fault/attack 취약성 분석을 통한 fault/attack 관리 모델을 제안하고, IP/GMPLS over DWDM 내의 제어프로토콜인 Extended-LMP (Link Management Protocol)와 RSVP-TE+(Resource Reservation Protocol-Traffic Engineering)를 이용하여 fault/attack 회복 절차를 제시한다.

• 한국컴퓨터정보학회논문지
• /
• 제26권1호
• /
• pp.19-26
• /
• 2021

역사적 지역에서 발굴되는 문화재는 시대적 배경을 바탕으로 고유의 특징을 가지고 있으며, 역사와 전파 지역의 흐름에 따라 그 문양과 특징이 조금씩 변화하는 것을 볼 수 있다. 어떤 지역에서 발굴된 문화재는 그 당시의 문화를 대표하며 온전한 모습을 유지하는 것도 있지만, 대부분이 파손/손실되거나 일부분으로 나누어져 그 구성을 연구하고 파손된 부분을 복구하기 위해 많은 전문가가 동원된다. 이 연구의 목적은 이러한 복원연구에 인공지능 신경망을 통해 과거의 문양과 패턴들을 학습시키고, 출토된 문화재에서 손실된 부분을 복원시키는 데 있으며, 문화재를 복원하기 위해 기본적인 생성적 적대 신경망인 GAN(Generative Adversarial Network)[1]을 사용한다. 연구에서는 GAN을 기반으로 출토된 문화재 일부를 기반으로 하여 손상/손실된 나머지 부분을 복구한 연구 과정으로, 학습에 기반이 되는 온전한 문화재의 이미지로 훈련을 하고, 일부를 마음대로 손상해 복구할 수 있도록 했다. 연구는 문화재 복구에 있어, 시대적 특징을 어느 정도 복구하는지, 색상과 재질을 복구하는지에 중점을 두고 있다. 마지막으로는 실제 출토된 비슷한 문화재를 기반으로 훈련된 신경망을 사용하여 문양을 복구함으로써 인공신경망의 적용 범위를 연구한다.

• 한국컴퓨터그래픽스학회논문지
• /
• 제30권3호
• /
• pp.81-86
• /
• 2024

영상 신호 프로세서(Image Signal Processor, ISP)는 카메라 센서로부터 획득된 RAW 영상을 사람의 눈에 보기 좋은 sRGB 영상으로 변환한다. RAW 영상은 sRGB 영상에 비해 영상 처리에 도움이 되는 정보를 가지고 있지만 상대적으로 큰 용량으로 인해 주로 sRGB 영상만 저장되고 사용된다. 또한, 실제 카메라의 ISP 과정이 공개되어 있지 않아 그 역과정을 모사하는 것은 매우 어렵다. 이에 sRGB와 RAW 영상의 상호 변환을 위한 카메라 ISP 모델링 연구가 활발히 진행되고 있으며, 최근 기존의 단순한 ISP 신경망 구조를 고도화하고 실제 카메라 ISP의 동작과 유사하게 카메라 파라미터(노출 시간, 감도, 조리개 크기, 초점 거리)를 직접 반영하는 ParamISP[1] 모델이 제안되었다. 하지만 ParamISP[1]를 포함한 기존의 연구는 카메라 ISP를 모델링함에 있어 렌즈로 인해 발생하는 렌즈 쉐이딩(Lens Shading), 광학 수차(Optical Aberration), 렌즈 왜곡(Lens Distortion) 등을 고려하지 않아 복원 성능에 한계가 있다. 본 연구는 ISP 신경망이 렌즈로 인해 발생하는 열화를 보다 잘 다룰 수 있도록 위치 정보 인코딩(Positional Encoding)을 도입한다. 제안하는 위치 정보 인코딩 기법은 영상을 분할하여 패치(Patch) 단위로 학습하는 카메라 ISP 신경망에 적합하며 기존 모델에 비해 영상의 공간적 맥락을 반영할 수 있어 더욱 정교한 영상 복원을 가능하게 한다.

• 한국통신학회논문지
• /
• 제29권1B호
• /
• pp.64-72
• /
• 2004

Multiprotocol Label Switching (MPLS) has been developed as a key technology to enhance the reliability, manageability and overall quality of service of core If networks with connection-oriented tunnel LSP and traffic engineering such as constraint-based routing, explicit routing, and restoration. In this paper, we propose a control bandwidth borrowing scheme that maximizes the utilization of tunnel LSPs or physical links by an extension to the RSVP-TE label distribution protocol. MPLS-based core switching network and VPN services rely on the establishment of connection-oriented tunneled LSPs that are configured or predefined by network management systems. The mechanism of network management system varies from (i) a relatively static LSP establishment accounting, to (ii) a dynamic QoS routing mechanisms. With the use of hierarchical LSPs, the extra bandwidth that is unused by the trunk (outer) LSPs should be fully allocated to their constituent end-to-end user traffic (inner) LSPs in order to maximize their utilization. In order to find out the unused extra bandwidth in tunnel LSP or physical link and redistribute these resources to constituent LSPs, we expend the functionality of RSVP-TE and the found unused extra bandwidth is redistributed with a weight-based recursive redistribution scheme. By the extended RSVP-TE and proposed recursive redistributed scheme, we could achieve the instantaneous maximized utilization of tunnel LSP or physical link suffering from the potential under-utilization problem and guarantee the end-to-end QoS requirements. With the proposed scheme, network manager can manage more effectively the extra available bandwidth of hierarchical LSPs and maximize the instantaneous utilization of the tunneled LSP resources.