Dissertation > Excellent graduate degree dissertation topics show

Robust Image Corner Detection Based on Curvature Product in Direct Curvature Scale Space

Author: LiChang
Tutor: ZhongBaoJiang
School: Nanjing University of Aeronautics and Astronautics
Course: Computational Mathematics
Keywords: Computer vision corner detection shape matching CSS DCSS MSCP
CLC: TP391.41
Type: Master's thesis
Year: 2009
Downloads: 49
Quote: 0
Read: Download Dissertation

Abstract


Corner detection is one of the most important tasks which has been widely studied in computer vision and pattern recognition. Among various corner detection techniques, the idea of multi-scale corner detection can get the features of a signal from high scales, while keeping the details of the signal at low scales. It is therefore attract much attention in the past few decades. Recently two new multiscale corner detectors, i.e., the direct curvature scale space (DCSS) algorithm and the multi-scale curvature product (MSCP) algorithm, have been proposed. The two algorithms have been respectively studied in this paper. By combining their advantages, a new corner detection algorithm, call product curvature scale space (P-DCSS) algorithm is introduced.The P-DCSS algorithm starts with extracting the contour of the object to be analyzed form real-world images, and then the curvature function of the curve is evolved. Finally, the curvature product of each point is computed through multiple scales, and local extremes of the curvature product are selected as corners when the value exceeds a threshold. To overcome the sensitivity of P-DCSS algorithm in the presence of noise, a small amount of Gaussian smoothing is suggested to be a preprocessing step. Compared to the DCSS algorithm, the P-DCSS algorithm omits a parsing process of the DCSS image, and hence it is much simpler in algorithm coding. Compared to the MSCP algorithm, the P-DCSS algorithm can work equally well at much less computational cost.In the last part of the paper, we discuss the application of the P-DCSS algorithm in shape matching. To solve the problem of different sizes of the object shapes to be matched, the data resulting from P-DCSS corner detection are simplified and normalized at first. Then a matching value is evaluated which measures the similarity between different object shapes. Numerical experiments show that the P-DCSS algorithm has a good performance both for corner detection and shape matching.

Related Dissertations

  1. Research of Image Mosaic Technology,TP391.41
  2. Research on Learning-Based Low-Level Vision Problem,TP391.41
  3. Camera Calibration and Position and Pose Detecting on Vision Measurement System of PCB,TP391.41
  4. Research on Nondestructive Detection Technology for External Qualities of Papayas Based-on Vision,S667.9
  5. Study on Detection and Grading of ’Jiro’ Persimmon’s External Quality Based on Computer Vision,S665.2
  6. Prediction of Pork Processing Functionalities,TS251.1
  7. Research on Inspection Technology of Dehydrated Garlic Slice Based on Computer Vision,TP391.41
  8. Moving target trajectory analysis based Intelligent Traffic Monitoring System,TP277
  9. Research on Several Technologies of Image Analysis for the Objectification of Tongue Diagnosis,TP391.41
  10. The application of wavelet analysis in the feature extraction of palmprint,TP391.41
  11. The Design and Implementation of On-line Strip Surface Defect Inspection System Based on Computer Vision,TP391.41
  12. Single Camera Calibration Method Based on Dual-mirror Imaging,TP391.41
  13. Target Tracking System Design and Implementation,TP391.41
  14. Infrared and visible image registration and fusion research,TP391.41
  15. Camera calibration Related Issues,TP391.41
  16. Coaxial different field picture image registration algorithm,TP391.41
  17. Real-time modeling spatial augmented reality technology research related,TP391.41
  18. Based on Image Processing Method of automatic target,TP391.41
  19. Application Kinect augmented reality and gesture recognition educational support system,TP391.41
  20. Research on Fast Corner Extraction Algorithm in Camera Calibration,TP391.41
  21. Research on Key Technologies of Monocular Visual Odometry Based on DSP,TP391.41

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net  Mobile