• 한국물환경학회지
• /
• 제28권3호
• /
• pp.347-358
• /
• 2012

The SWAT (Soil and Water Assessment Tool) should be calibrated and validated with observed data to secure accuracy of model prediction. Recently, the SWAT-CUP (Calibration and Uncertainty Program for SWAT) software, which can calibrate SWAT using various algorithms, were developed to help SWAT users calibrate model efficiently. In this study, three algorithms (GLUE: Generalized Likelihood Uncertainty Estimation, PARASOL: Parameter solution, SUFI-2: Sequential Uncertainty Fitting ver. 2) in the SWAT-CUP were applied for the Soyang-gang dam watershed to evaluate these algorithms. Simulated total streamflow and 0~75% percentile streamflow were compared with observed data, respectively. The NSE (Nash-Sutcliffe Efficiency) and $R^2$ (Coefficient of Determination) values were the same from three algorithms but the P-factor for confidence of calibration ranged from 0.27 to 0.81 . the PARASOL shows the lowest p-factor (0.27), SUFI-2 gives the greatest P-factor (0.81) among these three algorithms. Based on calibration results, the SUFI-2 was found to be suitable for calibration in Soyang-gang dam watershed. Although the NSE and $R^2$ values were satisfactory for total streamflow estimation, the SWAT simulated values for low flow regime were not satisfactory (negative NSE values) in this study. This is because of limitations in semi-distributed SWAT modeling structure, which cannot simulated effects of spatial locations of HRUs (Hydrologic Response Unit) within subwatersheds in SWAT. To solve this problem, a module capable of simulating groundwater/baseflow should be developed and added to the SWAT system. With this enhancement in SWAT/SWAT-CUP, the SWAT estimated streamflow values could be used in determining standard flow rate in TMDLs (Total Maximum Daily Load) application at a watershed.

• Journal of the Korean Society for Industrial and Applied Mathematics
• /
• 제27권2호
• /
• pp.146-159
• /
• 2023

This paper proposes a hybrid algorithm combining K-means clustering and watershed algorithms for flower segmentation and counting. We use the K-means clustering algorithm to obtain the main colors in a complex background according to the cluster centers and then take a color space transformation to extract pixel values for the hue, saturation, and value of flower color. Next, we apply the threshold segmentation technique to segment flowers precisely and obtain the binary image of flowers. Based on this, we take the Euclidean distance transformation to obtain the distance map and apply it to find the local maxima of the connected components. Afterward, the proposed algorithm adaptively determines a minimum distance between each peak and apply it to label connected components using the watershed segmentation with eight-connectivity. On a dataset of 30 images, the test results reveal that the proposed method is more efficient and precise for the counting of overlapped flowers ignoring the degree of overlap, number of overlap, and relatively irregular shape.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2020년도 학술발표회
• /
• pp.96-96
• /
• 2020

In order to establish the risk criteria of inundation due to typhoons or heavy rainfall, research is underway to predict the limit rainfall using basin characteristics, limit rainfall and artificial intelligence algorithms. In order to improve the model performance in estimating the limit rainfall, the learning data are used after the pre-processing. When 50.0% of the entire data was removed as an outlier in the pre-processing process, it was confirmed that the accuracy is over 90%. However, the use rate of learning data is very low, so there is a limitation that various characteristics cannot be considered. Accordingly, in order to predict the limit rainfall reflecting various watershed characteristics by increasing the use rate of learning data, the watersheds with similar characteristics were clustered. The algorithms used for clustering are K-Means, Agglomerative, DBSCAN and Spectral Clustering. The k-Means, DBSCAN and Agglomerative clustering algorithms are clustered at the impervious area ratio, and the Spectral clustering algorithm is clustered in various forms depending on the parameters. If the results of the clustering algorithm are applied to the limit rainfall prediction algorithm, various watershed characteristics will be considered, and at the same time, the performance of predicting the limit rainfall will be improved.

• Journal of Information Processing Systems
• /
• 제16권1호
• /
• pp.113-119
• /
• 2020

This study proposed the structure of the cluster's cell counting algorithm for cell analysis. The image required for cell count is taken under a microscope. At present, the cell counting algorithm is reported to have a problem of low accuracy of results due to uneven shape and size clusters. To solve these problems, the proposed algorithm has a feature of calculating the number of cells in a cluster by applying a radius change analysis to the existing distance conversion and watershed algorithm. Later, cell counting algorithms are expected to yield reliable results if applied to the required field.

• 대한원격탐사학회:학술대회논문집
• /
• 대한원격탐사학회 2003년도 Proceedings of ACRS 2003 ISRS
• /
• pp.525-527
• /
• 2003

Recently, the spatial resolution of earth observation satellites is significantly increased to a few meters. Such high spatial resolution images definitely will provide lots of information for detail-thirsty remote sensing users. However, it is more difficult to develop automated image algorithms for automated image feature extraction and pattern recognition. In this study, we propose a two-stage procedure to extract road information from high resolution satellite images. At first stage, a watershed segmentation technique is developed to classify the image into various regions. Then, a knowledge is built for road and used to extract the road regions. In this study, we use panchromatic and multi-spectral images of the IKONOS satellite as test dataset. The experiment result shows that the proposed technique can generate suitable and meaningful road objects from high spatial resolution satellite images. Apparently, misclassified regions such as parking lots are recognized as road needed further refinement in future research.

• 한국전자통신학회논문지
• /
• 제5권3호
• /
• pp.309-314
• /
• 2010

본 논문에서는 인터넷상에서의 지능형 감시 카메라 시스템(Intelligent Security Camera: ISC)을 제안한다. ISC 방법은 워터쉐드 알고리즘에 기반하여 카메라에 입력된 영상을 분할하는 단계와 skin-color model을 사용하여 얼굴의 후보지역을 탐지하는 단계, 그리고 마지막으로 SVM(Support Vector Machine)을 사용하여 얼굴 후보영역에서 얼굴을 검증하는 단계로 구성되어 있다. Skin-color Model을 이용하여 찾아진 얼굴후보 영역으로부터 웨이블렛 변환계수들을 추출한다. 웨이블렛 변환계수들을 SVM의 입력으로 하여 실제 얼굴영역을 검증한다. SVM의 입력으로 실험결과에서 제안된 방법이 감시시스템, 화상회의 시스템과 같은 얼굴을 인식 추적하는 시스템에 적용될 수 있음을 보인다.

• 한국수자원학회논문집
• /
• 제45권12호
• /
• pp.1213-1226
• /
• 2012

유역유출 모의 모형의 자동 보정에 주로 사용되는 진화계열의 알고리즘은 무제약 최적화 알고리즘이다. 이러한 진화계열 알고리즘에 제약조건을 반영하기 위해서는 제약조건을 다룰 수 있는 별도의 방법이 요구된다. 본 연구의 목적은 진화계열 알고리즘의 일종인 집합체 혼합진화 알고리즘에 벌칙함수를 적용하여 제약조건을 고려할 수 있도록 하는 것이다. 또한, 제약조건을 고려할 수 있는 집합체 혼합진화 알고리즘을 SWMM의 자동 보정 모듈에 적용하여 기존 자동 보정 모듈의 기능을 개선하는 것이다. 홍수유출 해석에서는 첨두유량과 관련된 지표가 중요하므로 첨두유량의 오차와 첨두유량 발생시간의 오차를 제어할 수 있는 제약조건을 구성하였다. 제약조건을 포함하여 구성된 자동 보정 모듈은 밀양댐 유역과 구로1 빗물펌프장 배수유역의 홍수유출 모의 모형에 대하여 적용되었다. 자동 보정의 결과는 제약조건의 포함 유무에 따른 결과를 비교하여제시되었다. 그 결과, 제약조건을 고려함에 따라 본래의 목적함수를 크게 위배하지 않으면서, 첨두유량과 첨두유량 발생시간의 오차가 크게 개선되었다. 또한, 검증을 통해서도 제약최적화를 통한 자동보정의 적절성이 검토되었다. 결론적으로 벌칙함수를 이용한 제약조건의 반영을 통해 자동 보정 모듈의 기능을 향상시킬 수 있었다.

• 한국농공학회지
• /
• 제22권4호
• /
• pp.108-114
• /
• 1980

A deterministic conceptual erosion model which simulates detachment, entrainment, transport and deposition of eroded soil particles by rainfall impact and flowing water is presented. Both upland and channel phases of sediment yield are incorporated into the erosion model. The algorithms for the soil erosion and sedimentation processes including land and crop management effects are taken from the literature and then solved using a digital computer. The erosion model is used in conjunction with the modified Kentucky Watershed Model which simulates the hydrologic characteristics from watershed data. The two models are linked together by using the appropriate computer code. Calibrations for both the watershed and erosion model parameters are made by comparing the simulated results with actual field measurements in the Four Mile Creek watershed near Traer, Iowa using 1976 and 1977 water year data. Two water years, 1970 and 1978 are used as test years for model verification. There is good agreement between the mean daily simulated and recorded streamflow and between the simulated and recorded suspended sediment load except few partial differences. The following conclusions were drawn from the results after testing the watershed and erosion model. 1. The watershed and erosion model is a deterministic lumped parameter model, and is capable of simulating the daily mean streamflow and suspended sediment load within a 20 percent error, when the correct watershed and erosion parameters are supplied. 2. It is found that soil erosion is sensitive to errors in simulation of occurrence and intensity of precipitation and of overland flow. Therefore, representative precipitation data and a watershed model which provides an accurate simulation of soil moisture and resulting overland flow are essential for the accurate simulation of soil erosion and subsequent sediment transport prediction. 3. Erroneous prediction of snowmelt in terms of time and magnitute in conjunction with The frozen ground could be the reason for the poor simulation of streamflow as well as sediment yield in the snowmelt period. More elaborate and accurate snowmelt submodels will greatly improve accuracy. 4. Poor simulation results can be attributed to deficiencies in erosion model and to errors in the observed data such as the recorded daily streamflow and the sediment concentration. 5. Crop management and tillage operations are two major factors that have a great effect on soil erosion simulation. The erosion model attempts to evaluate the impact of crop management and tillage effects on sediment production. These effects on sediment yield appear to be somewhat equivalent to the effect of overland flow. 6. Application and testing of the watershed and erosion model on watersheds in a variety of regions with different soils and meteorological characteristics may be recommended to verify its general applicability and to detact the deficiencies of the model. Futhermore, by further modification and expansion with additional data, the watershed and erosion model developed through this study can be used as a planning tool for watershed management and for solving agricultural non-point pollution problems.

• 한국수자원학회논문집
• /
• 제52권12호
• /
• pp.963-973
• /
• 2019

본 연구는 도시침수 위험기준이 산정되지 않은 지역의 예·경보 기준을 예측하기 위해 유역특성 자료와 피해이력 기반으로 산정된 한계강우량을 활용하여 도시침수 위험기준을 추정하는 모델을 검토하였다. 위험기준 추정모델은 머신러닝 알고리즘의 하나인 Support Vector Machine을 이용하여 설계하였으며, 학습자료는 지역별 한계강우량과 유역특성으로 구성하였다. 학습자료는 정규화 한 후 SVM 알고리즘에 적용하였으며, SVM에 적용시 Leave-One-Out과 K-fold 교차검증 알고리즘을 이용하여 절대평균오차와 표준편차를 계산한 후 모델의 성능을 평가하였다. Leave-One-Out의 경우 표준편차가 작은 모델이 최적모델로 선정되었으며, K-fold의 경우 fold의 개수가 적은 모델이 선정되었다. 선정된 모델의 지속시간별 평균 정확도는 80% 이상으로 나타나 침수 위험기준 추정을 위해 SVM을 활용가능 할 것으로 판단된다.

• 한국멀티미디어학회논문지
• /
• 제10권4호
• /
• pp.441-449
• /
• 2007

영상 분할을 위한 클러스터링에서는 방대한 계산량과 전형적인 분할 오류가 중요한 문제점으로 지적되어 왔다. 본 연구에서는 이러한 문제들을 최소화하기 위한 새로운 기법을 제안한다. 마커-제어 유역변환(marker- controlled watershed transform)에서 마커는 영역 확장의 시작점이므로, 분할된 각 영역을 대표하는 성질을 가진다. 따라서 마커 화소로 제한하는 클러스터링으로 계산 복잡도를 줄일 수 있다. 제안한 기법에서는 가보 필터(gabor filter)의 질감 에너지에서 마커를 선택하고, FCM(fuzzy c-means) 클러스터링으로 마커의 군집을 형성하며, 유역변환에서 생성된 영역들을 마커의 군집정보를 이용하여 병합한다. Brodatz 영상 조합에 대한 성능 실험에서 클러스터링 특유의 얼룩(blob) 분할 오류를 현저하게 개선하였으며, 영상 분할 소요 시간 비교에서 기존의 FCM 클러스터링 알고리즘보다 소요 시간이 적었다. 또한, 전체적으로 일정한 분할 소요시간을 보여주었다.