Earticle

The objectives of this study were to extract the major variables of forest structure characteristics such as the shape of the crown and the location of the individual trees by using Unmanned Aerial Vehicle(UAV) and evaluate the accuracy by comparing the extracted information with field data conducted in the Pinus koraiensis plot(Plot 1) in Wonju-si, and Larix leptolepis plot(Plot 2) in Hongcheon-gun, Gangwon-do. UAV images were classified into forest gap areas by using Otsu’s and Entropy methods. Moreover, the crown of an individual tree was delineated by applying Watershed(WS) and Valley Following(VF) algorithms. The estimated crown variables were compared with the field data and the accuracy was analyzed. Additionally, the point showing the highest reflectance value in the crown was considered as the location of each tree by using the local maxima filter. In the Plot 1, the overall accuracy and Khat of crown overlap ratio were 0.80 and 0.59, respectively, when using the Entropy_WS method, which were the highest values. The positional error was 0.74m in X-coordinate and 0.62m in Y-coordinate, which were within 1m of high accuracy. In the Plot 2, the overall accuracy and Khat of crown overlap ratio were 0.78 and 0.51, respectively, when using the Entropy_VF method, which were the highest. The positional error RMSE was 0.59m in X-coordinate and 0.38m in Y-coordinate, which were the highest among all.

영상 세션화는 기본적으로 이미지를 의미있는 구조로 나누는 것이다. 항상 특정 요구에 따라 이미지를 완전히 분할 한다. 하지만 결과가 원하지 않는 오버세션화가 이루어질 수 있다. 따라서 본 논문에서는 오버세션화를 줄이기 위해 가우시안 혼합 모델(GMM) 기반의 변형된 워터쉐드 알고리즘을 제안한다. 거리 변환 영상은 가우시안 혼합 모델에 의해 처리된다. 그리고 기대 최대화(EM) 알고리즘은 가우시안 혼합모델 데이터에 맞추는 데 사용된다. 이들 평균 의 평균을 최적의 임계값으로 선택하고 획득된 임계값에 기초하여 이진 영상으로 변환한다. 이렇게 얻은 이진 영상 은 적절한 구조 요소의 적용하여 침식화한다. 침식된 영상에서 수행된 워터쉐드 분할은 효과적인 분할 결과를 제공 한다. 제안된 방법이 사용될 때 옥수수, 쌀 및 밀과 같은 다른 커널이 효과적으로 분할된다. 실험 목적으로 38 개의 옥수수 커널, 95 개의 쌀 커널 및 32 개의 밀 커널을 사용한다. 분할 정확도는 옥수수, 쌀 및 밀에 대해 각각 94.7 %, 96.8 % 및 90.6 % 이다.

Image segmentation is basically the division of an image into meaningful structures. It always produces complete division of the image as the specific demand. However, it is prone to over-segmentation which makes the result unfavorable. Therefore, this paper proposes modified watershed algorithm based on Gaussian mixture model(GMM) to reduce the over-segmentation. Distance transformed image is processed by GMM. And expectation maximization(EM) algorithm is used to fit GMM to data. Average of these means is chosen as optimal threshold and the image is converted into binary image based on the obtained threshold value. Binary image thus obtained is eroded with the help of proper structuring element. Watershed segmentation thus carried out on the eroded image gives the effective segmentation results. Different kernels such as corn, rice, and wheat are effectively segmented when the proposed methodology is used. 38 corn kernels, 95 rice kernels, and 32 wheat kernels are used for the experimental purpose. The accuracy of segmentation is as high as 94.7%, 96.8%, and 90.6% for corn, rice, and wheat respectively.

This paper proposes an adaptive threshold watershed algorithm, which is the method used for image segmentation and boundary detection, which extends the region on the basis of regional minimum point. First, apply adaptive thresholds to determine regional minimum points. Second, it extends the region by applying adaptive thresholds based on determined regional minimum points. Traditional watershed algorithms create over-segmentation, resulting in the disadvantages of breaking boundaries between regions. These segmentation results mainly from the boundary of the object, creating an inaccurate region. To solve these problems, this paper applies an improved watershed algorithm applied with adaptive threshold in regional minimum point search and region expansion in order to reduce over-segmentation and breaking the boundary of region. This resulted in over-segmentation suppression and the result of having the boundary of precisely divided regions. The experimental results show that the proposed algorithm can apply adaptive thresholds to reduce the number of segmented regions and see that the segmented boundary parts are correct.

In this paper, we propose a watershed algorithm that applies a high-frequency enhancement filter to emphasize the boundary and a local adaptive threshold to search for minimum points. The previous method causes the problem of over-segmentation, and over- segmentation appears around the boundary of the object, creating an inaccurate boundary of the region. The proposed method applies a high-frequency enhancement filter that emphasizes the high-frequency region while preserving the low-frequency region, and performs a minimum point search to consider local characteristics. When merging regions, a fixed threshold is applied. As a result of the experiment, the proposed method reduced the number of segmented regions by about 58% while preserving the boundaries of the regions compared to when high frequency emphasis filters were not used.

The image segmentation technology is of great significance to the target identification. The watershed segmentation algorithm has wide application in image segmentation. The traditional watershed segmentation often causes the problems of over segmentation and noise sensitivity. Therefore, a medical image segmentation algorithm is proposed based on K-means clustering algorithm and improved watershed algorithm. First, K - means clustering algorithm is used for initial segmentation, and then the concept of similarity is put forward to improve the original watershed algorithm. Finally, the adjacent tiles of the initial segmentation is merged. The magnetic resonance image is regarded as the segmentation object. The experimental result shows that the proposed algorithm effectively solves the problem of the over-segmentation of traditional watershed algorithm, and achieves a satisfactory effect for the image segmentation.

Taking the high-resolution remote sensing image of industrial waste dump as the research object, this paper proposes a new image segmentation method based on marker-controlled watershed transform and region merging. The method gets the finial partition result by two-phrase segmentation on the pan-sharpened true color ALOS image. In the first phase, color gradient image of original image should be calculated and then it is modified by morphological impose minima, which should use markers extracted in two different ways. Lastly, preliminary segmentation result is obtained by watershed transform operates on the modified color gradient image. To solve the over-segmentation of industrial solid waste and other ground objects, region merging operation is performed according to the similarity measure criterion based on segmented objects’ color histogram Bhattacharyya coefficient in the second phrase, and then the final result is obtained. This method has been tested on the pan-sharpened ALOS image of 2.5 meters resolution in Shizuishan industrial zone, China. Experiment results demonstrate that this method is feasible and effective to segment the remote sensing industrial solid waste image.

얼굴영역을 분할하기 위해서 Watershed Algorithm 와 Object Grouping 을 이용한 얼굴영역 분할기법을 제안한다. 영상분할에 단점은 단일 알고리즘으로 영역분할이 어렵고, 또한 복잡한 영상에서 정확한 영역을 분할하기가 어렵다는 것이다. 그래서 본 논문에서는 Watershed Segmentation 기법과 Grouping 작업을 통한 병합, 그리고 색상의 선형회귀분석을 이용한 분석법을 적용하여 분할하고자 한다. 얼굴영역 분할방법을 전처리 과정과 영역 병합 그리고 얼굴 부분을 추출하는 3 단계의 과정으로 나누고, 전처리 과정에서는 수리형태학적(Mophological) 연산자를 이용한 영상 분할기법을 이용하여 분할한 후 얼굴 후보 영역을 검출, 영역병합과정에서 기존의 학습데이터와의 유사도를 측정, 얼굴객체추출 조건에 맞지 않는 객체들을 모두 제거함으로써, 정확한 얼굴부분을 분할해 낸다. 실험결과 제안한 방법을 통해 비교적 정확한 얼굴영역을 분할 할 수 있었다.

In this paper, we propose a modified watershed algorithm to obtain exact edge of region. The proposed method adjusts priority at zero-crossing point of gradient in order to make the point of region decision time postponed. We compare the proposed method with a previous method and prove that this method can extract more correct edge of region.

오프라인 필기체 한글 문자인식에서 대부분의 연구들은 영상획득 장비로부터 얻어진 이진영상(Binary image)을 바탕으로 이루어진다. 이 과정 중 영상에 잡음이나 영상패턴의 훼손을 가져오는 경우가 많다. 획이 끊기거나 영상 내 홀(holes)이 발생한 경우 인식에 많은 질적인 문제를 가져온다. 오프라인 필기체 한글 문자인식 과정 중 영상 내 골격을 추출하는 연구는 아직도 많은 난제를 가지고 있다. 또한 골격추출과정은 인식에 많은 영향을 준다. 잡영이 포함된 영상은 잘못된 골격선 추출에 기인한다. 본 논문에 사용된 Binary Watershed Algorithm은 잡영이 포함된 영상개선에 사용하였고, 이 Algorithm은 많은 다양성을 가지고 있어 여러 분야의 응용에 사용되어지고 있다. 본 논문은 이러한 잡영이 포함된 영상의 개선을 통해 기존의 Morphological 세선화 방법과 Zang-Suen 세선화 방법을 통해 골격선 추출을 평가하였다. 여기에는 아직도 자소의 교차 획에 있어서 효과적인 골격선을 추출하는 문제를 가지고 있다.

In this paper, we propose an efficient method not only f3r producing accurate region segmentation, solving the over-segmentation problem of watershed algorithm but also f3r reducing post-processing time by reducing computation loads. Through this proposed method, region segmentation of neighboring objects and discrimination of similar intensities were effectively obtained. Input image of watershed algorithm has used the derivative-based detectors such as Sobel and Canny. But proposed method uses the pixels-similarity-based detector, that is, SUSAN. By adopting this proposed method, we can reduce the noise problem and solve the problem of over-segmentation and not lose the edge information of objects. We also propose Zero-Crossing SUSAN. With Zero-Crossing SUSAN, the edge localization, times and computation loads can be improved over those obtained from existing SUSAN

Text line segmentation is a critical task in handwritten document recognition. In this paper, we propose a novel text-line-segmentation method using baseline estimation and watershed. The baseline-detection algorithm estimates the baseline using Adaptive Head-Tail Connection (AHTC) on the document. Then, the watershed method segments the line region using the baseline-detection result. Finally, the text lines are separated by watershed result and a post-processing algorithm defines the lines more correctly. The scheme successfully segments text lines with 97% accuracy from the handwritten document images in the ICDAR database.

In this paper, we proposed a liver extracting procedure for computer aided liver diagnosis system. Extraction of liver region in an abdominal CT image is difficult due to interferences of other organs. For this reason, liver region is extracted in a region of interest(ROI). ROI is selected by the window which can measure the distribution of Hounsfield Unit(HU) value of liver region in an abdominal CT image. The distribution is measured by an existential probability of HU value of lever region in the window. If the probability of any window is over 50%, the center point of the window would be assigned to ROI. Actually, liver region is not clearly discerned from the adjacent organs like muscle, spleen, and pancreas in an abdominal CT image. Liver region is extracted by the watershed segmentation algorithm which is effective in this situation. Because it is very sensitive to the slight valiance of contrast, it generally produces over segmentation regions. Therefore these regions are required to merge into the significant regions for optimal segmentation. Finally, a liver region can be selected and extracted by prier information based on anatomic information.

해상 안보, 국제 동향 파악 등 다양한 이유로 해상 사진에서 선박을 탐지하고자하는 연구는 지속되어 왔다. 인공지능의 발달로 인해 사진 및 영상 내 객체 탐지를 위한 R-CNN 모델이 등장하였고 객체탐지의 성능이 비약적으로 상승하였다. R-CNN 모델을 이용한 해상 사진에서의 선박 탐지는 인공위성 사진에도 적용되기 시작하였다. 하지만 인공위성 사진은 넓은 지역을 투사하기 때문에 선박 외에도 차량, 지형, 건물 등 다양한 객체들이 선박으로 인식되는 경우가 있다. 본 논문에서는 R-CNN계열 모델을 이용한 인공위성 사진에서의 선박 탐지의 성능을 개선하기 위한 새로운 방법론을 제안한다. 표지자 기반 watershed 알고리즘을 통해 육지와 바다를 분리하고 morphology 연산을 수행하여 RoI를 한 차례 더 특정한 뒤 특정된 RoI에 R-CNN 계열 모델을 사용하여 선박을 탐지하여 오탐을 줄인다. 해당 방법을 이용하여 Faster R-CNN을 사용하였을 경우, Faster R-CNN만을 사용했을 때에 비해 오탐률을 80% 줄일 수 있었다.

Research has been ongoing to detect ships from offshore photographs for a variety of reasons, including maritime security, identifying international trends, and social scientific research. Due to the development of artificial intelligence, R-CNN models for object detection in photographs and images have emerged, and the performance of object detection has risen dramatically. Ship detection in offshore photographs using the R-CNN model has also begun to apply to satellite photography. However, satellite images project large areas, so various objects such as vehicles, landforms, and buildings are sometimes recognized as ships. In this paper, we propose a novel methodology to improve the performance of ship detection in satellite photographs using R-CNN series models. We separate land and sea via marker-based watershed algorithm and perform morphology operations to specify RoI one more time, then detect vessels using R-CNN family models on specific RoI to reduce typology. Using this method, we could reduce the misdetection rate by 80% compared to using only the Fast R-CNN.

본 논문은 최소점 검색과 레이블 할당 시에 임계치를 적용한 개선된 워터쉐드(watershed) 알고리즘을 제안한다. 기존의 워터쉐드 알고리즘은 과분할의 문제점을 발생시키며 그로 인한 후처리가 필요하다 이러한 과분할 영역은 특히 객체의 경계주위에서 발생하여 부정확한 영역의 경계를 만든다. 이러한 문제점들을 해결하기 위해서 본 논문에서는 과분할을 막기 위해 잡음 제거와 경사(gradient) 영상을 구해 영상을 단순화시킨 후, 최소점 검색과 레이블 할당 시에 임계치를 적용한 개선된 워터쉐드 알고리즘을 적용해 과분할을 억제하고 분할된 영역의 경계를 정확히 나타나도록 하였다.

This paper proposes an improved image segmentation algorithm by the watershed algorithm based on the local adaptive threshold on local minima search and the fixing threshold on label allocation. The previous watershed algorithm generates the problem of over-segmentation. The over-segmentation makes the boundary in the inaccuracy region by occurring around the object. In order to solve those problems we quantize the input color image by the vector quantization, remove noise and find the gradient image. We sorted local minima applying the local adaptive threshold on local minima search of the input color image. The simulation results show that the proposed algorithm controls over-segmentation and makes the fine boundary around segmented region applying the fixing threshold based on sorted local minima on label allocation.

Mixed raster content 모델 기반의 영상 부호화는 전경과 배경 레이어에 빈 영역인 don't care region (DCR)이 발생하게 되며, 이 영역에 대한 신호 채움 방식에 따라 전체적인 부호화 성능이 큰 영향을 받게 된다. DCR을 채우기 위한 대부분의 기존 기법들은 홀 주변 기존 영역의 특성을 효율적으로 이용하지 않아 신호 채움 후에도 기존 신호 영역에 존재하던 고주파 성분이 반영될 뿐 아니라, DCR 경계에서의 신호 불연속으로 인해 고주파 성분이 추가적으로 발생한다는 문제점을 갖고 있다. 이 문제를 해결하기 위해 본 논문은 우선순위 기반의 적응적 영역 확장법을 이용한 새로운 DCR 채움 알고리즘을 제안한다. 제안 알고리즘은 분수령 알고리즘을 이용하여 DCR의 각 홀 픽셀에 대한 신호 특성을 판단한 후, 이를 토대로 영역 채움의 우선순위를 결정한다. 이 우선 순위를 기반으로 영역 확장을 수행함으로써 고주파 성분을 포함하고 있는 영역의 확장이 최소화되어 전체적인 부호화 효율이 향상될 수 있다. 실험 결과를 통해 제안 알고리즘이 비교 대상 알고리즘에 비해 효율적인 신호 채움으로 우수한 부호화 성능을 가짐을 보인다.

Image coding based on mixed raster content model generates don't care regions (DCR) in foreground and background layers, and its overall coding performance is greatly affected by region filling methods for DCRs. Most conventional methods for DCR filling fail in utilizing the local signal properties in hole regions and thus the high frequency components in non-DCR regions are reflected into DCR after signal filling. In addition, further high frequency components are induced to the filled signal because of signal discontinuities in the boundary of DCR. To solve this problem, a new DCR filling algorithm using the priority-based adaptive region growing is proposed in this paper. The proposed method uses the watershed algorithm and the flooding priority of each pixel for region filling is determined from the degree of smoothness in the neighborhood area. By growing the filled region into DCR based on the computed priority, the expansion of high-textured area can be minimized which can improve the overall coding performance. Experimental results show that the proposed method outperforms conventional algorithms.

Grain size of material is important factor in evaluating mechanical properties. Methods for grain size determination are described in ASTM grain size standards. However, conventional method require pretreatment of the surface to clarify grain boundaries. In this study, the grain size from the surface image obtained from scanning electron microscope was measured using the watershed algorithm, which is a region-based method among image segmentation techniques. The shapes of the crystals are similar to each other, but the size and growth height are different. In addition, crystal grains are adjacent to each other, so it is very similar to the shape image of the topography. Therefore, grain boundaries can be efficiently detected using the Watershed algorithm.

본 논문에서는 MPEG-4 와 같은 객체 및 내용 기반 영상 부호화에 활용될 수 있는 영상분할 알고리듬을 제안한다. 기존의 수학적 형태학(mathematical morphology)을 이용한 영상분할은 대개 과분할(over-segmentation)된 결과를 출력하는 경향이 있다. 이러한 과분할 문제 때문에 미소영역(small region)이나 비슷한 특성을 갖는 인접영역들을 서로 병합시켜야 하는 단점을 갖고 있다. 본 논문에서는 기존 영상분할의 문제점을 해결하고자 화소단위가 아닌 블록단위의 마커추출을 이용한 영상분할을 제안한다. 즉, 블록단위로 영상을 분할함으로써 질감부분에 해당하는 영역들이 하나의 큰 영역으로 분할될 수 있도록 하고, 영상내 객체(Object)의 정확한 윤곽선(contour)을 찾기 위해 화소단위로 워터쉐드(watershed) 알고리듬을 적용한다. 결과적으로 본논문에서 제안하는 알고리듬을 기존의 수학적 형태학을 이용한 방법과 비교하여 질감영역에서 향상된 영상분할과 계산시간의 부담을 줄일 수 있다는 것을 실험을 통해 확인하였다.

In this paper, we propose a new image segmentation algorithm which can be use din object-based image coding applications such as MPGA-4. Since the conventional objet segmentation methods based on mathematical morphology tend to yield oversegmented results, they normally need a postprocess which merges small regions to obtain a larger one. To solve this oversegmentation problem, in this paper, we prosed a block-based segmentation algorithm that can identify large texture regions in the image. Also, by applying the watershed algorithm to the image blocks between the homogeneous regions, we can obtain the exact pixel-based contour. Experimental results show that the proposed algorithm yields larger segments, particularly in the textural area, and reduces the computational complexities.

영상 분할(image segmentation)은 주어진 영역에서 의미 있는 영역만을 분리해 내는 과정으로 컴퓨터 비젼 분야에 있어 매우 중요한 단계로 취급되어 진다. 본 연구에서는 기존의 영상 분할 알고리즘들의 문제점을 해결하고자 단계적인 marker 영상을 적용한 watershed 알고리즘을 제안하였다. 제안한 방법은 단계적인 level의 조절에 따라 분할의 정도를 선택할 수 있으며 watershed 알고리즘의 특성인 정확한 물체 검출의 장점들 살릴 수 있다.

본 논문에서는 치아에 대한 컬러영상에서 개별적인 치아영역을 분할하기 위한 새로운 방법을 제안하였다. 제안하는 알고리듬은 치아의 구조적 특징을 이용한 구강구조모델과 워터쉐드 알고리듬의 새로운 경계선 설정방법 등이 사용되었다. 먼저, 컬러영상으로부터 치아영역이 강조된 회색레벨 영상을 획득하고 치아영역 분할시 문제가 될 수 있는 불필요한 부분을 영상에서 제거하였다. 다음으로 제안한 구강구조모델을 이용한 치아영상의 영상향상을 실행하였고, 향상된 영상을 워터쉐드 알고리즘을 이용하여 개별적 치아영역을 분할하였다. 워터쉐드 알고리즘에 필요한 경계선과 시드는 최소 문턱치를 이용한 이진영상의 경계선과 국부 최대값을 적용하였다. 제안한 방법의 성능을 평가하기 위하여 기존의 방법과 제안한 방법에 대하여 비교 실험을 수행하였다. 실험 결과, 제안한 방법이 기존의 방법에 비하여 대구치영역의 검출율이 향상됨을 확인하였으며 치아를 포함한 구강 내 영역의 중복검출 등의 문제를 방지하여 치아영역 검출 성능이 향상되었음을 확인하였다.

In this paper, we proposed a new algorithm for individual tooth region segmentation on tooth color images. The proposed algorithm used oral cavity model based on structural feature of tooth and new boundary of watershed algorithm. First, the gray scale image is obtained with emphasized tooth regions from the color images and unnecessary regions are removed on tooth images. Next, the image enhancement of tooth images is implemented using the proposed oral cavity model, and the individual tooth regions are segmented by watershed algorithm on the enhanced images. Boundary and seeds necessary to watershed algorithm are applied boundary of binary image using minimum thresholding and region maximum value. In order to evaluate performance of proposed algorithm, we conduct experiment to compare conventional algorithm with proposed algorithm. As a result of experiment, we confirmed that the proposed algorithm is more improved detection ratio than conventional algorithm at molar regions and the tooth region detection performance is improved by preventing overlap detection on oral cavity.