Watershed Segmentation Research Papers - Academia.edu (original) (raw)
Turkey has been divided into 26 basins for studies related to water resources development. Surface areas of the basins range between 6,300 – 127,500 km2, the average being approximately 30,000 km2. Compared with nested watershed... more
Turkey has been divided into 26 basins for studies related to water resources development. Surface areas of the basins range between 6,300 – 127,500 km2, the average being approximately 30,000 km2. Compared with nested watershed classification of U.S., the average size of Turkish national basins corresponds to the third-level (basins). Currently, a forth-level (sub-basins) with an average size of 1,820 km2 reside below the basins in U.S. classification. Even these sub-basins are found to be too large to adequately serve many water resource management needs. Two additional levels (watersheds and sub-watersheds) are currently under development in order to solve this problem. Taking this fact into consideration, currently existing basins of Turkey are supposed to be too large for effective water resources management and there is a need for delineation of sub-basins in a scientific manner. In this study, the boundaries of proto-type sub-basins are determined by using DEM-based automated delineation techniques. For this purpose, 30 arc-seconds SRTM30 DEM is re-projected into an equal area projection, hydrologic sinks are filled and inaccurate elevation values are corrected. In order to prevent invalid delineation of closed basins that cover approximately 10% of the country, water bodies located at the center of these basins are extracted from Global Land Cover Characteristics dataset of USGS and excluded from the process. Among the alternatives, Deterministic-8 (D-8) algorithm with modifications by Garbrecht and Martz is used to calculate flow direction and flow accumulation grids. Using these grids, a set of candidate sub-basins are generated by applying different area thresholds. A threshold of 1,000 km2 resulted in 359 sub-basins with an average size of 1,858 km2, both of which are found to be appropriate and manageable for nationwide usage. Similarity of the average size of the delineated sub-basins to that of U.S. sub-basins also supports this result.
Gandaki province has the good potentiality of hydroelectricity generation with existing twenty-nine hydroelectricity projects. Since the Province is rich in water resources, analysis of watersheds needs to be done for management, planning... more
Gandaki province has the good potentiality of hydroelectricity generation with existing twenty-nine hydroelectricity projects. Since the Province is rich in water resources, analysis of watersheds needs to be done for management, planning and identification of water as well as natural resources. GIS offers integration of spatial and no spatial data to understand and analyze the watershed processes and helps in drawing a plan for integrated watershed development and management. The Digital Elevation Model (DEM) available on the NASA-Earth data has been taken as a primary data for morphometric analysis of watershed in Gandaki Province using QGIS. Delineation of watershed was conducted from a DEM by computing the flow direction and using it in the Watershed tool. Necessary fill sink correction was made before proceeding to delineation. A raster representing the direction of flow was created using Flow Direction tool to determine contributing area. Flow accumulation raster was created f...
We propose in this work a study of an image processing engine able to detect automatically the features of electronic board weldings. The engine has been developed by using ImageJ and OpenCV libraries. Specifically the image processing... more
We propose in this work a study of an image processing engine able to detect automatically the features of electronic board weldings. The engine has been developed by using ImageJ and OpenCV libraries. Specifically the image processing segmentation has been improved by watershed approach. After a complete design of the automation processes, different test have been performed showing the engine efficiency in terms of features extraction, scale setting and thresholding calibration. The engine provides as outputs the storage of the cropped images of each single defects. The proposed engine together with the post-processing 3D imaging represent a good tool for the management of the production quality of electronic boards.
- by G. Seetharaman and +2
- •
- Computer Vision, Kalman Filter, Motion Detection, Object Tracking
The presented work through this paper is outcome of need for correct brain tumor identification, which is done here through a handy approach of segmenting tumor portion from different MRI slices taken of different patients. Tumor, which... more
The presented work through this paper is outcome of need for correct brain tumor identification, which is done here through a handy approach of segmenting tumor portion from different MRI slices taken of different patients. Tumor, which is of swelling appearance, is abruptly growth of cells in brain, causes a painful death. It’s correct diagnosis and within less time is necessary, which has been tried to achieve through this paper. Paper shows serially connected steps of image process, consisting pre-processing using thresholding through morphological opening, watershed segmentation, entropy filtration, and again morphological opening for tumor extraction, performed via MatLab software. Idea behind is using high intensity value of image which is normally tumor indication. This algorithm is very efficient for best case of tumors and equally helpful for other cases, as area, centroid, pixel info and condition can be determined through present work.
In this paper, the use of Multi-Layer Perceptron (MLP) Neural Network model is proposed for recognizing unconstrained offline handwritten Numeral strings. The Numeral strings are segmented and isolated numerals are obtained using a... more
In this paper, the use of Multi-Layer Perceptron (MLP) Neural Network model is proposed for recognizing unconstrained offline handwritten Numeral strings. The Numeral strings are segmented and isolated numerals are obtained using a connected component labeling (CCL) algorithm approach. The structural part of the models has been modeled using a Multilayer Perceptron Neural Network. This paper also presents a new technique to remove slope and slant from handwritten numeral string and to normalize the size of text images and classify with supervised learning methods. Experimental results on a database of 102 numeral string patterns written by 3 different people show that a recognition rate of 99.7% is obtained on independent digits contained in the numeral string of digits includes both the skewed and slant data.
Tuberculosis (TB) is very dangerous and rapidly spread disease in the world. In the investigating cases for suspected tuberculosis (TB), chest radiography is not only the key techniques of diagnosis based on the medical imaging but also... more
Tuberculosis (TB) is very dangerous and rapidly spread disease in the world. In the investigating cases for suspected tuberculosis (TB), chest radiography is not only the key techniques of diagnosis based on the medical imaging but also the diagnostic radiology. So, Computer aided diagnosis (CAD) has been popular and many researchers are interested in this research areas and different approaches have been proposed for the TB detection and lung decease classification. In this paper, the medical background history of TB decease in chest X-rays and a survey of the various approaches in TB detection and classification are presented. The literature in the related methods is surveyed papers in this research area until now 2014.
Object-Based Image Analysis (OBIA) has gained swift popularity in remote sensing area mainly due to the increasing availability of very high resolution imagery. Image segmentation is a major step within OBIA process. Image segmentation... more
Object-Based Image Analysis (OBIA) has gained swift popularity in remote sensing area mainly due to the increasing availability of very high resolution imagery. Image segmentation is a major step within OBIA process. Image segmentation quality remarkably influences the subsequent image classification accuracy. It is necessary to implement advanced and robust methods to increase image segmentation quality that is generally measured by several accuracy metrics including Area Fit Index (AFI and Quality Rate (Qr). In this study, two widely-used segmentation algorithms, namely region-based multiresolution segmentation and edge-based watershed transform were applied to a very high resolution imagery acquired by VorldView-2 sensor to evaluate and compare their performance in terms of segmentation quality metrics. Totally five segmentation goodness metrics, namely under-segmentation, over-segmentation, root means square, AFI and Qr were applied through the manually digitized reference objects available on the imagery. ENVI and eCognition Developer software packages were used to perform watershed transform and multiresolution segmentation algorithms, respectively. Nearest neighbor classification method was applied and related accuracy assessment was conducted in two software platforms. Results showed that multi-resolution segmentation was superior (about 18% higher in terms of AFI) compared to watershed transform in the delineation of segments of reference objects. Also, higher classification accuracies (about 5%) were achieved by the use of multi-resolution segmentation.
- by Xiaohe Chen and +1
- •
- Algorithms, Biomedical Engineering, Image Processing, Medical Imaging
In this work a marker-controlled and regularized watershed segmentation is proposed. Only a few previous studies address the task of regularizing the obtained watershed lines from the traditional marker-controlled watershed segmentation.... more
In this work a marker-controlled and regularized watershed segmentation is proposed. Only a few previous studies address the task of regularizing the obtained watershed lines from the traditional marker-controlled watershed segmentation. In the present formulation, the topographical distance function is applied in a level set formulation to perform the segmentation, and the regularization is easily accomplished by regularizing the level set functions. Based on the well-known Four-Color theorem, a mathematical model is developed for the proposed ideas. With this model, it is possible to segment any 2D image with arbitrary number of phases with as few as one or two level set functions. The algorithm has been tested on real 2D fluorescence microscopy images displaying rat cancer cells, and the algorithm has also been compared to a standard watershed segmentation as it is implemented in MATLAB. For a fixed set of markers and a fixed set of challenging images, the comparison of these two methods shows that the present level set formulation performs better than a standard watershed segmentation.
Patterns isolation is one of the bases of image processing and artificial intelligence algorithms as it is the key for image segmentation and posterior classification of objects. A brief review is performed in this document, regarding the... more
Patterns isolation is one of the bases of image processing and artificial intelligence algorithms as it is the key for image segmentation and posterior classification of objects. A brief review is performed in this document, regarding the Hough and watershed transforms algorithms, and in regards to its robustness to some conditions and implementations.
The objective of this paper is to provide a texture based segmentation algorithm for better delineation of the epithelial layer from histological images in discriminating normal and oral sub-mucous fibrosis (OSF). As per literature and... more
The objective of this paper is to provide a texture based segmentation algorithm for better delineation of the epithelial layer from histological images in discriminating normal and oral sub-mucous fibrosis (OSF). As per literature and oral clinicians, it is established that the OSF initially originates and propagates in the epithelial layer. So, more accurate segmentation of this layer is extremely important for a clinician to make a diagnostic decision. In doing this, Gabor based texture gradient is computed in gray scale images, followed by preprocessing of the microscopic images of oral histological sections. On the other hand, the color gradients of these images are obtained in the transformed Lab color space. Finally, the watershed segmentation is extended to segment the layer based on the combination of texture and color gradients. The segmented images are compared with the ground truth images provided by the oral experts. The segmentation results depict the superiority of the texture based segmentation in comparison to the Otsu's based segmentation in terms of misclassification error. Results are shown and discussed.► In this study we provide a texture based epithelial layer segmentation algorithm. ► Accurate segmentation is important for clinician to make a diagnostic decision. ► The watershed is used to segment the layer using texture and color gradients. ► The results depict the superiority of the segmentation compared with Otsu's method.
A marker-controlled and regularized watershed segmentation is proposed for cell segmentation. Only a few previous studies address the task of regularizing the obtained watershed lines from the traditional marker-controlled watershed... more
A marker-controlled and regularized watershed segmentation is proposed for cell segmentation. Only a few previous studies address the task of regularizing the obtained watershed lines from the traditional marker-controlled watershed segmentation. In the present formulation, the topographical distance function is applied in a level set formulation to perform the segmentation, and the regularization is easily accomplished by regularizing the level set functions. Based on the well-known Four-Color theorem, a mathematical model is developed for the proposed ideas. With this model, it is possible to segment any 2D image with arbitrary number of phases with as few as one or two level set functions. The algorithm has been tested on real 2D fluorescence microscopy images displaying rat cancer cells, and the algorithm has also been compared to a standard watershed segmentation as it is implemented in MATLAB. For a fixed set of markers and a set of challenging images, the comparison of these two methods shows that the present level set formulation performs better than a standard watershed segmentation.
The quality of underwater images is directly affected by water medium, atmosphere medium, pressure and temperature. This emphasizes the necessity of image segmentation, which divides an image into parts that have strong correlations with... more
The quality of underwater images is directly affected by water medium, atmosphere medium, pressure and temperature. This emphasizes the necessity of image segmentation, which divides an image into parts that have strong correlations with objects to reflect the actual information collected from the real world. Image segmentation is the most practical approach among virtually all automated image recognition systems. Feature
Segmentation of a 3-dimensional (3D) polygonal mesh is a method of breaking the mesh down into\ meaningful" connected subsets of vertices called regions. The method used here is based on the watershed segmentation scheme which... more
Segmentation of a 3-dimensional (3D) polygonal mesh is a method of breaking the mesh down into\ meaningful" connected subsets of vertices called regions. The method used here is based on the watershed segmentation scheme which appears prominently in ...
The present paper develops a general methodology for the morphological segmentation of hyperspectral images, i.e., with an important number of channels. This approach, based on watershed, is composed of a spectral classification to obtain... more
The present paper develops a general methodology for the morphological segmentation of hyperspectral images, i.e., with an important number of channels. This approach, based on watershed, is composed of a spectral classification to obtain the markers and a vectorial gradient which gives the spatial information. Several alternative gradients are adapted to the different hyperspectral functions. Data reduction is performed either by Factor Analysis or by model fitting. Image segmentation is done on different spaces: factor space, parameters space, etc. On all these spaces the spatial/spectral segmentation approach is applied, leading to relevant results on the image.
Segmenting a MRI images into homogeneous texture regions representing disparate tissue types is often a useful preprocessing step in the computer-assisted detection of breast cancer. That is why we proposed new algorithm to detect cancer... more
Segmenting a MRI images into homogeneous texture regions representing disparate tissue types is often a useful preprocessing step in the computer-assisted detection of breast cancer. That is why we proposed new algorithm to detect cancer in mammogram breast cancer images. In this paper we proposed segmentation using vector quantization technique. Here we used Linde Buzo-Gray algorithm (LBG) for segmentation of
we present in this paper an automatic system of urban road extraction from satellite and aerial imagery. Our approach is based on an adaptive directional filtering and a watershed segmentation. The first stage consists of an automatic... more
we present in this paper an automatic system of urban road extraction from satellite and aerial imagery. Our approach is based on an adaptive directional filtering and a watershed segmentation. The first stage consists of an automatic procedure which adapts filtering of each block band to the dominant direction(s) of roads. The choice of the dominant direction(s) is made from a criterion based on the calculation of a factor of direction of detection. The second stage is based on watershed algorithm applied to a Shen-Castan gradient image. This process provides a decision map allowing correcting the errors of the first stage. A ratio of surface on perimeter is used to distinguish among all segments of the image those representing probably roads. Finally, in order to avoid gaps between pieces of roads, the resulting image follows a treatment, based on proximity and colinearity, for linking segments. The proposed approach is tested on common scenes of Landsat ETM+ and aerial imagery of...
- by G. Seetharaman and +2
- •
- Multimedia, Kalman Filter, Computer Software, Motion Detection
The objective of this paper is to present a method to characterize a brain tumor. The authors developed a tumor characterization technique using Marker Controlled Watershed Segmentation method and region property functions using image... more
The objective of this paper is to present a method to characterize a brain tumor. The authors developed a tumor characterization technique using Marker Controlled Watershed Segmentation method and region property functions using image processing toolbox. The parameters extracted are area, major and minor axis length, eccentricity, orientation, equivdiameter, solidity and perimeter. This method is quite versatile, fast and simple to use. This can be applied to all type of 2D MR Images representing all tumors irrespective of their location in human body and their size. The proposed technique has been simulated on Matlab and results are compared with experimental data obtained from diagnostic centre.
Based on a series of empirical studies beginning in the 1970’s, it was noted that remote sensing data suffered from the scale and aggregation problem. It was further recognized that there was no unique or ‘optimal’ spatial resolution for... more
Based on a series of empirical studies beginning in the 1970’s, it was noted that remote sensing data suffered from the scale and aggregation problem. It was further recognized that there was no unique or ‘optimal’ spatial resolution for detecting the different sized, shaped, and spatially arranged entities represented in a remote sensing image of a complex scene. Today within the Earth sciences, it is strongly recognized that landscapes exhibit distinctive spatial patterns associated to different processes at different scales. Consequently, multiscale approaches are required for modern landscape analysis. It is within this context that the Multiscale Object-Specific Analysis (MOSA) framework was developed. In this paper we review the background, foundations, and recent developments of MOSA. We begin with the original definition of Object-Specific Analysis (OSA) and Object-Specific Upscaling (OSU), and continue with the recent integration of Marker Controlled Watershed Segmentation ...
Brain tumour detection is one of the challenging tasks in medical image processing. The present study discusses in detail the segmentation process by means of histogram clustering, Global thresholding, Watershed segmentation and edge... more
Brain tumour detection is one of the challenging tasks in medical image processing. The present study discusses in detail the segmentation process by means of histogram clustering, Global thresholding, Watershed segmentation and edge based segmentation. Six MRI images from radiologists were collected and the experiments were conducted for statistical analysis also. A comparative study is made and the results are of great interest and practical utility.
This study is part of a regeneration program of the coconut grove of French Polynesia where most coconut palm trees of the Tuamotu archipelago were planted in the 1980s following the various hurricanes that had struck islands. The French... more
This study is part of a regeneration program of the coconut grove of French Polynesia where most coconut palm trees of the Tuamotu archipelago were planted in the 1980s following the various hurricanes that had struck islands. The French Polynesia government acquired one-meter pansharpened RGB Ikonos images over the Tuamotu archipelago. To exploit these data, a pilot study is conducted on the island of Tikehau, well-known from the specialists and easily accessible from Tahiti. A maximum likelihood (ML) classification is performed to segment the high vegetation in images. Thus, a support vector machines (SVM) classification allows the high vegetation to be classified in different patterns. And finally, a robust segmentation process based on markers controlled watershed segmentation is proposed to extract tree crowns. Through the ground mission, the trees detection accuracy is estimated which is then used to compute the number of trees the closest to the reality by applying a weighted factor to the number of trees located in each class.
Plumes are a mixture of fresh water and river sediment load, with some dilution caused by currents. Spatial and temporal variation of the river plumes can be studied by remote sensing techniques. The main objectives of this work were... more
Plumes are a mixture of fresh water and river sediment load, with some dilution caused by currents. Spatial and temporal variation of the river plumes can be studied by remote sensing techniques. The main objectives of this work were modeling the Douro River Plume (DRP) dimension based on image segmentation of MERIS data and to establish a relationship between the DRP dimension and different input parameters. Two different segmentation techniques were applied (watershed and region-based) in order to estimate the DRP dimension of twenty-five MERIS scenes (from 2003 to 2005). Firstly, we considered a simple linear regression model of the DRP dimension on the water volume, considering seasonal effects (summer period and the rest of the year), where a significant correlation of 0.664 was found (watershed segmentation) ignoring summer period. The second proposed model consisted in the incorporation of several parameters (last available plume, water volume, tide height and wind speed), presumed to be related to the DRP dimension. A determination coefficient of 62.2% was found for watershed segmentation excluding the summer period, regarding the multiple linear regression branch of the second proposed model.