Urban regions are dynamic environments. Especially their road maps change by the expansion of the urban region. Therefore, automatic detection of roads from very high resolution aerial and satellite images is a very i...
详细信息
Urban regions are dynamic environments. Especially their road maps change by the expansion of the urban region. Therefore, automatic detection of roads from very high resolution aerial and satellite images is a very important research field. Unfortunately, the solution is not straightforward by using basic imageprocessing and computer vision algorithms. Therefore, advanced methods are needed for road network detection from aerial and satellite images. In this study, we propose a novel method for automatic detection of road segments from very high resolution color aerial and satellite images. Our method depends on choosing a training set from the input image manually. We use color chroma values of pixels as the discriminative features. Since road pixels have similar color characteristics, the distribution of color chroma feature values of the training region have a peak at a certain point in the feature space which shows the road class. Using this information and one-class classification methodology, we label road segments in a given remotely sensed image. Finally, we fit a road network shape on the detected segment. Experimental results on color aerial and Ikonos satellite images show the importance of color features in road detection applications.
image segmentation technology is an important part of the lower level of computer vision. It's also a basic precondition for image analysis and patternrecognition. It has been widely used in many fields such as m...
详细信息
image segmentation technology is an important part of the lower level of computer vision. It's also a basic precondition for image analysis and patternrecognition. It has been widely used in many fields such as medical images and remotesensingimages. Meanwhile, image segmentation is a difficulty in imageprocessing as well. Aiming at medical imagery, a novel variational domain approach to curve evolution for image segmentation is proposed based on a statistical active contour model using level sets. The essential idea is to re-define the computing domain in image repeatedly by separating the segmentation procedure into several individual phases. By our algorithm, the work can be done automatically without manual intervention. Moreover, compared with current methods, the rapidity can be enhanced effectively for the objects with complicated topology.
In material science and engineering, the grain structure inside a super-alloy sample determines its mechanical and physical properties. In this paper, we develop a new Multichannel Edge-Weighted Centroidal Voronoi Tes...
详细信息
In material science and engineering, the grain structure inside a super-alloy sample determines its mechanical and physical properties. In this paper, we develop a new Multichannel Edge-Weighted Centroidal Voronoi Tessellation (MCEWCVT) algorithm to automatically segment all the 3D grains from microscopic images of a super-alloy sample. Built upon the classical k-means/CVT algorithm, the proposed algorithm considers both the voxel-intensity similarity within each cluster and the compactness of each cluster. In addition, the same slice of a super-alloy sample can produce multiple images with different grain appearances using different settings of the microscope. We call this multichannel imaging and in this paper, we further adapt the proposed segmentation algorithm to handle such multichannel images to achieve higher grain-segmentation accuracy. We test the proposed MCEWCVT algorithm on a 4-channel Ni-based 3D super-alloy image consisting of 170 slices. The segmentation performance is evaluated against the manually annotated ground-truth segmentation and quantitatively compared with other six image segmentation/edge-detection methods. The experimental results demonstrate the higher accuracy of the proposed algorithm than the comparison methods.
The SIFT operator's success for computer vision applications makes it an attractive alternative to the intricate feature based SAR image registration problem. The SIFT operator processing chain is capable of detec...
详细信息
The SIFT operator's success for computer vision applications makes it an attractive alternative to the intricate feature based SAR image registration problem. The SIFT operator processing chain is capable of detecting and matching scale and affine invariant features. For SAR images, the operator is expected to detect stable features at lower scales where speckle influence diminishes. To adapt the operator performance to SAR images we analyse the impact of image filtering and of skipping features detected at the highest scales. We present our analysis based on multisensor, multitemporal and different viewpoint SAR images. The operator shows potential to become a robust alternative for point feature based registration of SAR images as subpixel registration consistency was achieved for most of the tested datasets. Our findings indicate that operator performance in terms of repeatability and matching capability is affected by an increase in acquisition differences within the imagery. We also show that the proposed adaptations result in a significant speed-up compared to the original SIFT operator.
This paper presents the framework for the navigation and target tracking system for a mobile robot. Navigation and target tracking are to be performed using a Microsoft Xbox Kinect sensor which provides RGB color and ...
详细信息
This paper presents the framework for the navigation and target tracking system for a mobile robot. Navigation and target tracking are to be performed using a Microsoft Xbox Kinect sensor which provides RGB color and 3D depth imaging data to an x86 based computer onboard the robot running Ubuntu Linux. A fuzzy logic controller to be implemented on the computer is considered for control of the robot in obstacle avoidance and target following. Data collected by the computer is to be sent to a server for processing with learning-based systems utilizing neural networks for patternrecognition, object tracking, long-term path planning and process improvement. An eventual goal of this work is to create a multi-agent robot system that is able to work autonomously in an outdoor environment.
The unified descriptive experiment design regularization (DEDR) method from a companion paper provides a rigorous theoretical formalism for robust estimation of the power spatial spectrum pattern of the wavefield scat...
详细信息
The unified descriptive experiment design regularization (DEDR) method from a companion paper provides a rigorous theoretical formalism for robust estimation of the power spatial spectrum pattern of the wavefield scattered from an extended scene observed in the uncertain remotesensing (RS) environment. For the considered here imaging synthetic aperture radar (SAR) application, the proposed DEDR approach is aimed at performing, in a single optimized processing, SAR focusing, speckle reduction, and RS scene image enhancement and accounts for the possible presence of uncertain trajectory deviations. Being nonlinear and solution dependent, the optimal DEDR estimator requires rather complex signal processing operations ruled by the fixed-point iterative implementation process. To simplify further the processing, in this paper, we propose to incorporate the descriptive regularization via constructing the projections onto convex sets that enable us to factorize and parallelize the reconstructive imageprocessing over the range and azimuth coordinates, design a family of such regularized easy-to-implement iterative algorithms, and provide the relevant computational recipes for their application to fractional imaging SAR. We also comment on the adaptive adjustment of the DEDR operational parameters directly from the actual speckle-corrupted scene images and demonstrate the effectiveness of the proposed adaptive DEDR techniques.
In recent years, the development of high-resolution remotesensingimage extends the visual field of the terrain features. Quickbird and other high-resolution remotesensingimage can show more characteristics such as...
详细信息
ISBN:
(纸本)9781424473021
In recent years, the development of high-resolution remotesensingimage extends the visual field of the terrain features. Quickbird and other high-resolution remotesensingimage can show more characteristics such as shape, spectral, texture and so on. Back Propagation neural network is widely used in remotesensingimage classification in recent years, it is a self-adaptive dynamical system which is widely connected by large amount of neural units, and it bases on distributing store and parallel processing. It study by exercise and had the capacity of integrating the information, synthesis reasoning, and rapid overall processing capacity. It can solve the regular problem arise from remotesensingimageprocessing, therefore, it is widely used in the application of remotesensing. This paper discusses the Back Propagation neural network method in order to improve the high resolution remotesensingimage classification precision. By analyzing the principle and learning algorithms of Back Propagation neural network, we utilize the Quickbird imagery of Beijing with high resolution as experimental data and do the research of road and simple building roof, In this paper, the use of remotesensingimageprocessing software Matlab, and then combined with Back Propagation neural network classifier for the high resolution remotesensingimages of their patternrecognition.
暂无评论