Papers by Keyword: Region Segmentation

Abstract: Segmentation algorithm of 3D point cloud data based on region growing is proposed, the main idea is as follows: First, seed points in each region of object surface are searched, and then, starts from the seed point, the process of regional growing is done, which all the point cloud data belong to same surface are included until some discontinuous set of points appear. The algorithm is implemented under C, and the 3D point cloud data are showed by OPENGL software.

Abstract: Fast and accurate segmentation of the facial feature regions is of great significance to research on objectification of complexion diagnosis. In this paper, we used the active appearance model (AAM) system to accurately locate 68 key points, and segmented the face region simultaneously. According to the Chinese medicine complexion-viscera principle, five face regions representing the five internal organs were segmented by using 68 key points. Experiments have shown that this method was efficient and fast for facial feature region segmentation, and laid a foundation for further research of objectification of complexion diagnosis.

Abstract: The image fusion technique is applied to the field of Terahertz (THz) image processing to solve the problem that THz cameras offered low-resolution images , which do not match the human visual habits. A fusion algorithm is proposed based on region segmentation and Nonsubsampled Contourlet (NSCT). The THz image is divided into target area and background area ,which are mapped into the visible images in next step, and the different rules are designed in NSCT domain. Experimental results show that the proposed algorithm displays the target, retains the background and detail information effectively. This algorithm is more feasible than traditional fusion algorithm based on pixel and regional energy in the human eye perceives, and can be used effectively for machine vision and terahertz security area.

Abstract: To automate present Automatic Optical Inspection (AOI) systems, an intelligent method based on lightness analysis for solder joints detection is proposed in this paper. Firstly, the image taken from real printed circuit board (PCB) is preprocessed and segmented. Then, we compute the image similarity by calculating the Euclidean distance. If the similarity exceeds the threshold, we extract the lightness of solder joints by converting between RGB color mode and HLS color mode. Lastly, all the defects of solder joints are classified into several different types according to their lightness appearances. Experiment results show that the method deduced in this paper can detect accurately solder defects from different patches.

Abstract: An automatic detection algorithm of bridges above water in aerial images is proposed. Firstly, aerial image is binarized based on gradient mean square variance, and binary image is denoised based on pixel density. Then connective areas are labeled in the binary image based on pixels to extract water area. Finally, bridge area is detected based on gray information. Experiments show this algorithm is effective in automatic detection of bridges above water in aerial images.

Abstract: The paper analyzes the Region Partitioning Method of highway pavement distress image and gives a new region partitioning method and its specific operation steps.

Abstract: Image segmentation is one focus of digital image processing. In this paper, fourteen different kinds of classical image segmentation algorithms are studied and compared using corn image and simulating in MATLAB based on HSI color model. The result reveals that the method that using H component based on HSI color model to deal with the histogram threshold algorithm and Laplace edge detection algorithm is effectively extract the plant from the corn image

Abstract: Currently CAD/CAM systems have become more common in automated manufacturing to enhance accuracy and efficiency. However, machining efficiency and accuracy are mainly relied on the experience of the engineer. Therefore, in this research the main aim is to propose the system which automatically analyzes the characteristics of a CAD model and to extract manufacturing features. The method for analyzing shapes of a CAD model represented by triangle mesh is studied. The system can extract planar, spherical, cylindrical, hole, arc, and other curved regions. This research will improve the integration of CAD, CAM system, and is supposed to have a great potential in the application of manufacturing and other engineering fields.

Abstract: An approach of regions segmentation algorithm in triangle meshes based on threshold angle of the normal vectors is proposed. In this paper, a set of threshold angles is used to identify features and segment different surfaces, and then the non-feature regions based on least squares method are fitted and the corresponding fitting precision are obtained. Finally, the best threshold angle of each region is determined according to the comparison the above fitting precisions and each non-feature region points are obtained exactly. This method can improve the precision of feature extraction and region segmentation in some hybrid blend region. Examples are given to prove the effectiveness and the processing flexibility and stability of the proposed approach.

Abstract: The surface topography characterization of MEMS device is very important to bonding technology of MEMS device. Motif characterizing method is a characterizing method of surface topography by graph. Aiming at the diversity and regionality of surface topography of MEMS device, in this study we have sampled the surface of MEMS device by 3-dimentinal grids using the surface profiler developed by us and characterizes the surface topography of MEMS device by the extended Motif characterizing method. The surface of MEMS device is divided into several Motif regions; the surface topography of every divided region can be evaluated respectively; the details of every region can be zoomed and these regions as a whole or every region can be revolved and projected; one of these regions can be as the reference of other regions. So the height, gradient and other characteristics of others regions of the whole MEMS device surface can be analyzed relative to the reference region; the whole and local surface topography of whole MEMS device can be analyzed.