Dissertation > Excellent graduate degree dissertation topics show

Unsupervised multi-channel remote sensing image change detection methods

Author: ZhaoLei
Tutor: WangBin
School: Fudan University
Course: Circuits and Systems
Keywords: Change detection SashForms Threshold segmentation Semi-supervised support vector machine Fuzzy C-Means Clustering Neighborhood Information Multi- channel remote sensing image
CLC: TP751
Type: Master's thesis
Year: 2010
Downloads: 160
Quote: 0
Read: Download Dissertation

Abstract


Remote sensing image change detection analysis of remote sensing images of the same area in different periods, from which detected the change over time between the surface features. Have a wide range of applications of remote sensing image change detection techniques in land and natural resources surveys, environmental monitoring, natural disaster prevention, military reconnaissance, which is an important aspect of the remote sensing information science. Without the supervision of remote sensing image change detection is the direct comparison of images for different periods, detect changes in information, without additional prior information, and therefore get a wide range of applications. How unsupervised method to improve the accuracy and applicability of the change detection, and become a research hotspot. Unsupervised existing multi-channel remote sensing image change detection, a large number of studies, solution. Its main contents include the following three parts: 1. Propose a multi-channel remote sensing image change detection method based on unsupervised segmentation window. This method the difference image is divided into sub-images, and to determine the difference in the overall image threshold through Qiuzi image local threshold. Solve the problems in remote sensing image change detection, when a relatively large proportion of the changes in the area of ??the region in a whole lot of remote sensing images or small, change detection methods can not accurately detect the change information. Experimental results show that the proportion of the change region in the parcel sensing image relatively large or small, with respect to the general change detection method, the method detection accuracy significantly improved. 2. Propose a split window combined with semi-supervised support vector machine classification method of remote sensing image change detection. The method combines semi-supervised support vector machine (Semisupervised Support Vector Machines, SSSVM) taxonomy first difference image for the sub-image instead of the whole image variation by sub-images to determine the classification of the whole image super optimal hyperplane plane. This method combines the advantages of semi-supervised support vector machine classification vector classification, to take full advantage of the remote sensing image band information, while addressing the changes in the regional area of ??the whole image is relatively larger or smaller, semi-supervised support vector machine classification method can not accurately detect changes in information. The experimental results show that this method can solve the remote sensing image change area relatively larger or smaller semi-supervised support vector machine classification method can not accurately detect the problem, change detection performance. 3. Propose a multi-channel remote sensing image change detection method based on fuzzy C-means clustering (Fuzzy C-Means, FCM) and neighborhood analysis. Usually multi-channel remote sensing image change detection methods will change information from multiple-band compression to a band, loss of remote sensing image band information. To solve this problem, this chapter presents a multi-band changes and FCM clustering-based approach to achieve multi-channel remote sensing image change detection. FCM sensitive outlier detection results easily affected by noise. This chapter on the basis of the FCM, proposed a combination of remote sensing image spatial information of multi-channel remote sensing image change detection methods, improved FCM sensitive to outliers. The experimental results show that, relative to the other change detection methods and flow cytometry (FCM) methods, the proposed multi-channel analysis of remote sensing image change detection method based on FCM and neighborhood remote sensing image change detection to eliminate noise, change detection performance .

Related Dissertations

  1. Citrus Image Segmentation Based on Genetic Algorithm,TP391.41
  2. Reduction and Rule Acquisition in Fuzzy Decision System Based on Neighborhood Relationship,TP18
  3. Research of Facial Expression Recognition Algorithm,TP391.41
  4. Study on Nonwoven Fabric Defects Inspection,TP391.41
  5. Research on Cutting Tool Wear Detection Based on Computer Vision,TP274
  6. Research on Orange Quality Detection Algorithm Based on DSP,TP391.41
  7. Polarimetric SAR coherent information on registration and change detection algorithm,TN957.52
  8. Automatic Registration of Terristrial Lidar Data Using Planar Targets,P225.2
  9. Habitation Extraction and Edge Optimization,P237
  10. Research on Change Detection for Remote Sensing Images Based on Discriminant Analysis,TP751
  11. Effect of Scenes Presentation on Change Detection in Ficker Paradigms,B842.3
  12. Density-based clustering algorithm,TP311.13
  13. Research and Implementation of CAD on the Pathological Slices of Gastric Adenocarcinoma,TP391.41
  14. Remote Sensing Detection for Subsidence-Resulted Water Body and Solid-Waste Dump in Yanzhou Coal Mining Area,P627
  15. The Research on Automatic Identification System for the Microscopic Images of No Chaetoceros Harmful Algae,TP391.41
  16. Research on Algorithm of Lung Nodule Detection Based on the Image Processing Technology,TP391.41
  17. The Monocular Vision Target-location Methods Based on Cross-ratio Invariants,TP391.41
  18. Research on Moving Targets Detection and Tracking in Video Sequences,TP391.41
  19. Research on Change Detection Technology of SAR Image,TN957.52
  20. Research and Implementation of Web Information Content Change Detection,TP393.092
  21. Video Retrieval for H.264 Compressed Video,TP391.41

CLC: > Industrial Technology > Automation technology,computer technology > Remote sensing technology > Interpretation, identification and processing of remote sensing images > Image processing methods
© 2012 www.DissertationTopic.Net  Mobile