Dissertation > Excellent graduate degree dissertation topics show

Research and Implementation of Visual Analysis Algorithm Based on Da Vanci Platform

Author: WangZuo
Tutor: YuLi；ZhengYaYu
School: Zhejiang University of Technology
Course: Control Theory and Control Engineering
Keywords: Moving target detection Human Recognition Ada-Boost DaVinci technology
CLC: TP391.41
Type: Master's thesis
Year: 2010
Downloads: 73
Quote: 0
Read: Download Dissertation

Abstract


The intelligent visual analysis system is a forward direction in the field of computer vision, the core of which is the use of computer vision technology to detect moving objects in the dynamic scene, tracking and recognition. At present, the the intelligent visual analysis system toward large-scale development direction of diversification of applications, PC-based video surveillance system has been unable to adapt to this trend, so the development of large-scale application, and the applicability of the embedded intelligent vision analysis The system has important theoretical significance and application value. Pedestrians is an important object of attention by the the intelligent visual analysis system, pedestrian detection, tracking and recognition is one of the core issues of the intelligent vision systems research, is also a concern in recent years, the field of computer vision hot research direction. Traditional human motion analysis algorithm often requires high-performance hardware support in order to achieve real-time requirements. How to improve human motion analysis algorithm, so that it applies to the embedded platform is the bottleneck in the development of the intelligent visual analysis system, this article is in this context, expand the study of intelligent visual analysis system based on Leonardo da Vinci embedded platform . The first chapter section first expounded the significance of the topic, followed by the development of intelligent visual analysis system on the basis of the overview of the status of domestic research outside the review and summary, and finally introduce the content and structure of the paper. The second chapter the problem of poor the traditional moving target detection algorithm scenes adaptability, a moving target detection algorithm block background modeling based on probability and statistics. Firstly, the original image data, transformed the way through the index map compression and block statistics through the background, framing updated to improve the detection performance of the algorithm. Meanwhile, the histogram modification method of moving objects extracted shadow suppression and moving objects, which laid the foundation for subsequent tracking recognition algorithms on embedded platforms achieve separation of adhesions. Finally, the use of fusion tracking algorithm based on the characteristics of a moving object, through two feature matching method to improve tracking accuracy. Chapter Ada-Boost algorithm based on Haar features for human recognition. Extracting sample characteristics, a more diverse Haar features flexible combination of sparse particles, thus enriching the sample set of features. Classifier offline training weak and strong classifier cascade of weak classifiers to improve the identification efficiency and reduce the complexity of the algorithm, in order to meet the real-time requirements of embedded intelligent vision systems. Chapter DaVinci platform optimization algorithm, computing hardware accelerator integral image, reducing processor computational burden, and the algorithm package to make it independent of the hardware platform has become to be a Linux application software alone call The software module. The fifth chapter summarizes the results of this research, and some further research.

Related Dissertations

  1. Research on the Classification Based on the Reconstruction of Solder Joint,TP391.41
  2. Tongue Feature Extraction and Research of Fusion Classification,TP391.41
  3. The Fatigue State Recognition of the Driver Based on Eye Detection,TP391.41
  4. Feature Extraction, Selection and Combination in Lipreading,TP391.41
  5. The Research on Paper Currency Classification Method Based on Harr-Like Feature and Minimal Ball Including Samples,TP391.41
  6. Research on Predicting Intrinsic Disorder Protein Structure Based on Supervision Manifold Learning Algorithm,Q51
  7. Face Recognition Method Based on DE,TP391.41
  8. Research on Face Recognition Based on AdaBoost Algorithm,TP391.41
  9. The Design and Implementation of Digital Video Server Based on DM6467,TP393.05
  10. Based on wavelet transform and linear subspace face recognition technology,TP391.41
  11. Embedded Wireless Video Surveillance System Based on TI’s DaVinci Platform,TP277
  12. Study on the Key Algorithms of AVS Video Coding,TN919.81
  13. 3G Mobile Video Surveillance System Based on TI DaVinci Technology,TP277
  14. Design of Traffic Data Collection System Based on Video,TP274.2
  15. Design and Implementation of Network Surveillance System for Multi-platform Based on DaVinci Technology,TN919.8
  16. Research on Multi-Channel Visual Platform of Intelligent Robot Based on DAVINCI Technology,TP242.6
  17. Hardware Design of Multi-Video Composite and Playback Device,TP277
  18. Design of Digital Multimedia Player System Based on DM6446,TP368.12
  19. Research on Moving Object Detection and Tracking in Complex Condition,TP391.41
  20. A Study on Moving Target Detecting and Tracking in Intelligent Video Surveillance Systems,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