• 제목/요약/키워드: Adaptive distance field

검색결과 24건 처리시간 0.021초

초분광 원격탐사 기반 항공관측 및 현장자료를 활용한 선박탐지 (The Ship Detection Using Airborne and In-situ Measurements Based on Hyperspectral Remote Sensing)

  • 박재진;오상우;박경애;;장재철;이문진;김태성;강원수
    • 한국지구과학회지
    • /
    • 제38권7호
    • /
    • pp.535-545
    • /
    • 2017
  • 한반도 주변 해상사고가 증가함에 따라 원격탐사 자료를 활용한 선박탐지 연구의 중요성이 점점 더 강조되고 있다. 이 연구는 고해상도 광학영상에 의존하는 기존 선박탐지 분야에 수백 개 채널의 분광정보를 포함하는 초분광영상을 활용하여 새로운 선박탐지 알고리즘 제시하였다. 두 차례의 현장관측을 통해 측정한 선박 선체의 반사 스펙트럼과 AVIRIS (Airborne Visible/Infrared Imaging Spectrometer) 초분광센서 영상의 선박 및 해수 반사 스펙트럼 간의 분광정합 기법을 적용하였다. 총 다섯 개의 탐지 알고리즘 spectral distance similarity (SDS), spectral correlation similarity(SCS), spectral similarity value (SSV), spectral angle mapper (SAM), spectral information divergence (SID)를 사용하였다. SDS는 선박 일부가 해수로 탐지되는 오차를 나타내었고, SAM은 선박과 해수 사이에 약 1.8배의 차이를 나타내어 명확한 분류 결과를 보여주었다. 이와 더불어 본 연구에서는 각 기법의 최적 임계값을 제시하여 초분광 영상에 포함되어 있는 선박을 분류하였으며 그 결과 SAM, SID가 다른 탐지 알고리즘에 비해 우수한 선박탐지 능력을 보여주었다.

두꺼비 서식지 적합성 지수(HSI) 모델개발을 위한 연구 (A Study on the Development of "Bufo gargarizans" Habitat Suitability Index(HSI))

  • 조근영;구본학
    • 한국환경복원기술학회지
    • /
    • 제25권2호
    • /
    • pp.23-38
    • /
    • 2022
  • This study investigates the characteristics and physical habitat requirements for each Bufo gargarizans life history through a literature survey. After deriving variables for each component of Bufo gargarizans, in order to reduce regional deviations from eight previously studied literature research areas for deriving the criteria for variables, a total of 12 natural habitats of Bufo gargarizanss are selected as spatial ranges by selecting four additional sites such as Umyeonsan Ecological Park in Seoul, Wonheungibangjuk in Cheongju in the central region, Changnyeong Isan Reservoir in the southern region, and Mangwonji in Daegu. This study presents Bufo gargarizans SI, a species endemic to Korea, whose population is rapidly declining due to large-scale housing site development and road development, and develops a Bufo gargarizans HSI model accordingly to improve the function of the damaged Bufo gargarizans habitat and to present an objective basis for site selection of alternative habitat. At the same time, it provides basic data for adaptive management and follow-up monitoring. The three basic habitat requirements of amphibians, the physical habitat requirements of Bufo gargarizans, synthesized with shelter, food, and water, and the characteristics of each life history, are classified into five components by adding space and threats through literature research and expert advice. Variables are proposed by synthesizing and comparing the general characteristics of amphibians, among the previously studied single species of amphibians, the components of HSI of goldfrogs and Bufo gargarizans, and the ecological and physical environmental characteristics of Bufo gargarizans. Afterwards, through consultation with an amphibian expert, a total of 10 variables are finally presented by adjacent forest area(ha), the distance between spawning area and the nearest forest land(m), the soil, the distance from the wetland(m), the forest layered structure, the low grassland space, the permanent wetland area(ha), shoreline slope(%), PH, presence of predators, distance from road(m), presence or absence of obstacles. n order to derive the final criteria for each of the 10 variables, the criteria(alternative) for each variable are presented through geographic information analysis of the site survey area and field surveys of the previously studied literature research area. After a focus group interview(FGI) of 30 people related to the Bufo gargarizans colony in Cheongju, a questionnaire and in-depth interviews with three amphibians experts are conducted to verify and supplement the criteria for each final variable. Based on the finally developed Bufo gargarizans HSI, the Bufo gargarizans habitat model is presented through the SI graph model and the drawing centering on the Bufo gargarizans spawning area

The growth and productivity of native Indonesian rice progenies and its relationship with root development during dry-season

  • Zakaria, Sabaruddin;Fitrya, Farid;Kurniawan, Trisda;Hereri, Agam Ihsan;Maulana, Teuku
    • 한국작물학회:학술대회논문집
    • /
    • 한국작물학회 2017년도 9th Asian Crop Science Association conference
    • /
    • pp.338-338
    • /
    • 2017
  • One of the problems in increasing rice production is getting lines or varieties that have high adaptability so that able to produce maximum production in a variety of environmental conditions. One strategy that can be done to get adaptive varieties is through the improvement of native varieties. This research was conducted in rain fed paddy field, Aceh province, Indonesia from June to September 2014. The texture of the soil was clay-loam with the soil pH ranged from 6.5-6.7. Five potential progenies of rice from crossing between native Indonesian rice with introduced rice varieties consist of C3, C4, S3, S5, S6 were used in this study. Besides that, one national rice variety Ciherang also used as a comparison. The plants were growth in the plot with the size of $2m{\times}1.4m$ with plant distance was $20cm{\times}20cm$. The fertilizers used in this study were Urea, NPK, and KCl. Randomized block design with 6 rice progenies/variety and 3 replications were used in this study. There were 18 experimental units and each experimental unit had 10 samples for the sources of data. The variables that were observed in this study including plant height at harvesting time, number of productive tiller, the percentage of empty grains and filled grain per panicle, weight of filled grains per hill, weight of filled grain per plot and yield potential per hectare. Analyzed were also conducted for the depth of root penetration, dry-root weight, dry-shoot weight, shoot-root weight ratio and its correlation with the weight of filled grain per hill. The research results show that there was significant difference on plant height at harvesting time, number of productive tillers, the percentage of empty grains and filled grain per panicle, weight of filled grains per hill, weight of filled grain per plot and yield potential per ha (p>0.01) among the treatments. In addition, depth root penetration, dry-root weight, dry-shoot weight, shoot-root weight ratio also had significant difference (p>0.01) among the treatment. The highest plant at harvesting time was found in S6, reaching 129.8 cm and the shortest plants was found in C3 reaching 107.5 cm. The largest number of productive tillers and the highest percentage of filled grains per panicle were found in Ciherang reaching 10.5 tillers and 80.7% respectively. Ciherang also had the heaviest weight of filled grains per hill and per plot reaching 21.1 g and 2.18 kg respectively. Whereas, S6 had the lowest number of tillers and the lowest percentage of filled grain per panicle.. The highest yield potential per ha was found in Ciherang reaching 7.79 tons. Among the progenies, S5 had the highest yield potential reaching 5 tons/Ha. The result also showed significant relationship between shoot-root weight ratio with weight of filled grains per hill. The highest value of shoot-root weight ratio (1.57) in Ciherang is thought had closed relationship with its yield potential.

  • PDF

스마트폰 다종 데이터를 활용한 딥러닝 기반의 사용자 동행 상태 인식 (A Deep Learning Based Approach to Recognizing Accompanying Status of Smartphone Users Using Multimodal Data)

  • 김길호;최상우;채문정;박희웅;이재홍;박종헌
    • 지능정보연구
    • /
    • 제25권1호
    • /
    • pp.163-177
    • /
    • 2019
  • 스마트폰이 널리 보급되고 현대인들의 생활 속에 깊이 자리 잡으면서, 스마트폰에서 수집된 다종 데이터를 바탕으로 사용자 개인의 행동을 인식하고자 하는 연구가 활발히 진행되고 있다. 그러나 타인과의 상호작용 행동 인식에 대한 연구는 아직까지 상대적으로 미진하였다. 기존 상호작용 행동 인식 연구에서는 오디오, 블루투스, 와이파이 등의 데이터를 사용하였으나, 이들은 사용자 사생활 침해 가능성이 높으며 단시간 내에 충분한 양의 데이터를 수집하기 어렵다는 한계가 있다. 반면 가속도, 자기장, 자이로스코프 등의 물리 센서의 경우 사생활 침해 가능성이 낮으며 단시간 내에 충분한 양의 데이터를 수집할 수 있다. 본 연구에서는 이러한 점에 주목하여, 스마트폰 상의 다종 물리 센서 데이터만을 활용, 딥러닝 모델에 기반을 둔 사용자의 동행 상태 인식 방법론을 제안한다. 사용자의 동행 여부 및 대화 여부를 분류하는 동행 상태 분류 모델은 컨볼루션 신경망과 장단기 기억 순환 신경망이 혼합된 구조를 지닌다. 먼저 스마트폰의 다종 물리 센서에서 수집한 데이터에 존재하는 타임 스태프의 차이를 상쇄하고, 정규화를 수행하여 시간에 따른 시퀀스 데이터 형태로 변환함으로써 동행 상태분류 모델의 입력 데이터를 생성한다. 이는 컨볼루션 신경망에 입력되며, 데이터의 시간적 국부 의존성이 반영된 요인 지도를 출력한다. 장단기 기억 순환 신경망은 요인 지도를 입력받아 시간에 따른 순차적 연관 관계를 학습하며, 동행 상태 분류를 위한 요인을 추출하고 소프트맥스 분류기에서 이에 기반한 최종적인 분류를 수행한다. 자체 제작한 스마트폰 애플리케이션을 배포하여 실험 데이터를 수집하였으며, 이를 활용하여 제안한 방법론을 평가하였다. 최적의 파라미터를 설정하여 동행 상태 분류 모델을 학습하고 평가한 결과, 동행 여부와 대화 여부를 각각 98.74%, 98.83%의 높은 정확도로 분류하였다.