• 한국수자원학회논문집
• /
• 제36권2호
• /
• pp.195-209
• /
• 2003

본 연구에서는 낙동강 상류유역의 병렬 다목적댐군인 안동 및 임하다목적 댐의 장기간 유입량을 산정하는데 공간추계 신경망모형이 사용되었다. 공간추계 신경망모형은 역전파 알고리즘으로 LMBP와 BFGS-QNBP를 각각 사용하였다. 공간추계 신경망모형의 구조는 입력층, 은닉층 및 출력층의 3개의 층과 차례대로 8-8-2개의 노드로 구성되어 있다. 입력층 노드는 안동 및 임하다목적 댐의 월평균유입량, 월면적강우량, 월별 증발접시 증발량과 월평균기온으로 구성되어 있으며, 자료시계열은 시간적으로 차이가 있다. 공간추계 신경망모형의 훈련을 위하여 추계학적 모형중 하나인 PARMA(1,1)에 의해서 훈련자료를 모의발생시켰으며, 모의발생된 자료는 공간추계 신경망모형의 훈련에 사용되었다. 훈련을 통하여 공간추계 신경망모형의 매개변수인 최적연결강도와 편차를 산정하였다. 산정된 매개변수는 안동 및 임하다목적 댐의 실측자료를 이용하여 공간추계 신경망모형의 검증에 이용되었으며, 통계분석과 수문곡선의 비교를 통하여 우수한 결과를 나타내었다. 따라서 공간추계 신경망모형은 낙동강 상류유역의 병렬저수지군의 장기간 연계운영기법 개발을 위하여 기초적인 자료를 제공하고, 용수분배 및 관리에 도움을 줄 것이다.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2020년도 학술발표회
• /
• pp.125-125
• /
• 2020

European Space Agency's Sentinel-1 has an improved spatial and temporal resolution, as compared to previous satellite data such as Envisat Advanced SAR (ASAR) or Advanced Scatterometer (ASCAT). Thus, the assumption used for low-resolution retrieval algorithms used by ENVISAT ASAR or ASCAT is not applicable to Sentinel-1, because a higher degree of land surface heterogeneity should be considered for retrieval. The assumption of homogeneity over land surface is not valid any more. In this study, considering that soil roughness is one of the key parameters sensitive to soil moisture retrievals, various approaches are discussed. First, soil roughness is spatially inverted from Sentinel-1 backscattering over Yanco sites in Australia. Based upon this, Artificial Neural Networks data (feedforward multiplayer perception, MLP, Levenberg-Marquadt algorithm) are compared with Fractal approach (brownian fractal, Hurst exponent of 0.5). When using ANNs, training data are achieved from theoretical forward scattering models, Integral Equation Model (IEM). and Sentinel-1 measurements. The network is trained by 20 neurons and one hidden layer, and one input layer. On the other hand, fractal surface roughness is generated by fitting 1D power spectrum model with roughness spectra. Fractal roughness profile is produced by a stochastic process describing probability between two points, and Hurst exponent, as well as rms heights (a standard deviation of surface height). Main interest of this study is to estimate a spatial variability of roughness without the need of local measurements. This non-local approach is significant, because we operationally have to be independent from local stations, due to its few spatial coverage at the global level. More fundamentally, SAR roughness is much different from local measurements, Remote sensing data are influenced by incidence angle, large scale topography, or a mixing regime of sensors, although probe deployed in the field indicate point data. Finally, demerit and merit of these approaches will be discussed.

• 한국공간정보시스템학회:학술대회논문집
• /
• 한국공간정보시스템학회 2002년도 춘계학술대회 논문집
• /
• pp.45-52
• /
• 2002

Artificial neural networks (ANN) have been successfully used for classifying remotely sensed imagery. However, ANN still is not the preferable choice for classification over the conventional classification methodology such as the maximum likelihood classifier commonly used in the industry production environment. This can be attributed to the ANN characteristic built-in stochastic process that creates difficulties in dealing with unequally represented training classes, and its training performance speed. In this paper we examined some practical aspects of training classes when using a back propagation neural network model for remotely sensed imagery. During the classification process of remotely sensed imagery, representative training patterns for each class are collected by polygons or by using a region-growing methodology over the imagery. The number of collected training patterns for each class may vary from several pixels to thousands. This unequally populated training data may cause the significant problems some neural network empirical models such as back-propagation have experienced. We investigate the effects of training over- or under- represented training patterns in classes and propose the pattern repopulation algorithm, and an adaptive alpha adjustment (AAA) algorithm to handle unequally represented classes. We also show the performance improvement when input patterns are presented in random fashion during the back-propagation training.