Dissertation > Excellent graduate degree dissertation topics show

Motion Object Detection Based on MPEG-4 Video

Author: DongWenTao
Tutor: TianYuMin
School: Xi'an University of Electronic Science and Technology
Course: Computer System Architecture
Keywords: MPEG-4 Target detection DC coefficient Motion vector Global motion estimation
CLC: TP391.41
Type: Master's thesis
Year: 2008
Downloads: 95
Quote: 0
Read: Download Dissertation


Moving object detection is the field of image processing and computer vision in a very active branch , has a wide range of applications in many aspects in the field of video surveillance and military . This paper focuses on the compressed domain moving target detection technology , has made ??a detailed study based on MPEG-4 , single - target detection . MPEG-4 video streams captured by a stationary camera moving target detection problem , this paper presents an improved domain moving target detection algorithm based on the MPEG-4 compression . The algorithm directly to obtain the motion vector in the compressed code stream , and use the good characteristics of the iterative filter as much as possible to restore the real motion vector information , using adaptive threshold selection method identifies the motion area by the motion vector . On this basis , the first calculation of the difference of the DC coefficient of the DC coefficient and the center of the the eight neighborhood within the sliding window block mold , and then according to the selected threshold value to determine the movement characteristics of the central block , and then extended motion area , calculate the background blocks in the new zone the average DC coefficient, and , with the edge block difference value calculation , and then adjust the edge block . The algorithm takes advantage of the MPEG-4 stream contains the motion information , overcome only by the motion vector detection of moving targets error big disadvantage . Experimental results show that the algorithm does not need to fully decode the compressed stream , can greatly improve the detection efficiency and improve the detection results on the different video sequences with good universality remains well under the influence of the light intensity, luminance and other factors robustness. Oriented sports background , global motion estimation is a widely used method to extract moving targets . This paper presents a direct use of the compressed video stream improved algorithm for global motion estimation . The algorithm starting from the basis of the global motion estimation , using the similarity of the characteristics of background macroblock motion quickly establish background macroblock collection and four parameter commonly used in the global motion estimation model estimation motion parameters . Finally, to calculate the motion vector residual, and the detection of moving targets . Experimental results show that the proposed algorithm only half decoding of compressed video stream to improve the detection efficiency , and improve the detection effect .

Related Dissertations

  1. Preparation of Enrofloxacin Sustained-release Preparations and Its Pharmacokinetics,R96
  2. The Design and Implementation of the Single-chip MPEG2 Video Encoder,TN919.81
  3. The Research and Application of the Transmission and Distribution for Multiple-Concurrent Flows Based on IPv6,TN949.292
  4. 3D Facial Animation Generation Design and Implementation,TP391.41
  5. Design and implementation of monitoring vehicle with multimedia terminals,TP37
  6. High Efficiency 3D Video Representation and Coding,TN919.81
  7. The Design of Network Video Monitor System Based on MPEG-4,TP277
  8. Embedded video decoder motion compensation process data layout optimization,TN919.81
  9. Contains curcumin MPEG-P (CL-co-PDO) polymeric micelles research,R283
  10. TILE Pro64 based on multi-core processors for 3G video transcoding server software design,TP393.05
  11. Temporal integration of visual selective attention model and applied research in the MPEG compressed domain,TP391.41
  12. Combination of multi-core based on the generalized Gaussian image classification method,TP391.41
  13. Research of Damage Efficiency by Livin/Survivin RNA Interference on Colorectal Cancer Cell Mediated by mPEGylated-chitosan Nanoparticles,R735.34
  14. Application and Research of Video Surveillance System in Highway Based on ASON,TP277
  15. The Research and Implementation of Streaming Media Player Based on DSP,TN919.8
  16. Feature-based Video Retrieval in Compressed Domain,TP391.41
  17. Motion Object Detection Based on MPEG-4 Video,TP391.41
  18. Research on Ontology-based Description and Retrieval of Surveillance Video and Development of Streaming Media Release Platform,TN919.8
  19. Research on Violent Video Detection Algorithm Based on Bag of Audio Words and MPEG-7 Features,TP391.41
  20. Design of Wireless Video Transmission System Based on 802.11,TN919.8
  21. The Realization of Audio/Video Synchronization for MPEG-2,TN949.197

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