Dissertation > Excellent graduate degree dissertation topics show

Complex environment of road detection technology research

Author: HaoZhiShuai
Tutor: TangZhenMin
School: Nanjing University of Technology and Engineering
Course: Applied Computer Technology
Keywords: Visual navigation Road detection Unstructured road kalman filtering Feature Fusion
CLC: U416.06
Type: Master's thesis
Year: 2009
Downloads: 391
Quote: 4
Read: Download Dissertation

Abstract


The road detection is one of the key technologies of vision-based vehicle navigation. Papers to the complex environment of the vision-based road detection techniques for the study of object using color road image segmentation algorithm presented, analyzed and discussed on the basis of prior art, road detection algorithm is able to adapt to the complex environment. Solved to some extent, the unstructured road in the presence of shadow traces of water, light and the the leaves coverage under the conditions of the road area recognition. The paper introduces the research status of the road detection technology, presents and analyzes from the path of structured and unstructured road detection algorithms already several well-known in the world, pointed out that the current structural road-based detection technology has matured, unstructured road detection technology is still in the research stage. The paper discusses the characteristics and typical color road image detection algorithm. First, based on the analysis of the road image features and road testing requirements of unstructured road fuzzy classification, the width of the gradient of the road and the width of the frequency change of the road two categories, laid the targeted road detection algorithm design foundation. Second, comprehensive analysis of a variety of image pre-processing technology, and by experiment designed image preprocessing algorithm for this topic. Finally, the comparative analysis of the various color road image edge detection algorithm and region segmentation algorithm. The paper studies the road detection algorithm based on the model. Designed an improved road detection algorithm based on a straight-line model with kalman filter tracking. Algorithm as the visual region near the detection region, and using linear modeling road boundary, the more real-time. The algorithm uses a recursive loop structure, the test results depends on the previous test results. First, within a certain region near the center line of the road in a previous detection sampling averaging seed pixels, and then use the region growing method the initial road segmentation based on HSI space. Select number of observations on the left and right borders of the road, road boundary line kalman filter estimates. Finally, according to the left and right boundary line projected road centerline. At the initial condition, the algorithm uses an unsupervised road model matching method detecting road boundary. In addition, the algorithm is fully taken into account the interference of the complex environment of the shadows, water stains and other road detection, designed to identify targeted remove links to enhance the accuracy of road detection. Thesis detection algorithm based on the characteristics of the road. The fusion road detection algorithm based on edge characteristics and regional characteristics. Making the region segmentation process, consider only the H and S component data, so to some extent alleviate the shadow environmental factors on the road detection. Edge feature extraction using an improved based on the direction of the equalization of local information entropy algorithm sobel operator, to improve the accuracy of edge detection. Based on the analysis of experimental data, regional characteristics and edge feature fusion algorithm that regional segmentation-based edge detection results of the correction method to improve the accuracy of detection of the road. Finally, using chain code algorithm to track the road boundary. Experimental results show that the road detection algorithm does not depend on the width of the road gradient hypothesis, to be able to adapt to the width of the road gradient or frequency-dependent environment. Algorithm able to detect a certain extent there are shadows, water stains, leaves covering a variety of complex environments unstructured road.

Related Dissertations

  1. The Research of Navigation Multi-Sensor Data Fusion Technology Based on Micro Unmanned Platform,V249.32
  2. Feature Extraction, Selection and Combination in Lipreading,TP391.41
  3. The Compression and Fusion Technique Research of Underwater Target Feature,TN911.7
  4. Study on Wireless Sensor Networks with Communication Constraints,TN929.5
  5. Research on Multiple Moving Target Detection and Tracking,TP391.41
  6. Abandoned Object Detection Based on Video Image Sequences,TP391.41
  7. The Research for Visual Surveillance System Based on Fusion of Multiple Model Data,TP391.41
  8. Study on GPS High Accuracy Deformation Monitoring Data Processing and Analysis Method,P228.4
  9. Attitude Determination and Finite Time Control Algorithms for a Satellite,V448.222
  10. GPS Single Epoch Algorithm and Its Application in Disaster Monitoring,P228.4
  11. Based on dual-source multi-feature fusion technology research gait,TP391.41
  12. Integration into high-level semantic feature of medical image retrieval technology,TP391.3
  13. Features and relevance feedback based on the underlying medical image retrieval,TP391.41
  14. Human Motion Captrue System Based on Monte Carlo Algorithm,TP391.41
  15. Vision-based Navigation Route Recognition for Agricultural Wheeled Mobile Robot,TP242.62
  16. Research on Flight Parameter for Autonomous Landing of UAV Based on Visual Navigation,V249.1
  17. Research on INS/SAR Integrated Navigation System,V249.328
  18. Based on feature fusion method of single- word aligned,TP391.2
  19. Research on Airborne Guidance Weapon Transfer Alignment,TJ762.2
  20. Small unmanned autonomous helicopter landing visual navigation system design and simulation,TN966
  21. Tracking moving objects using multi-feature fusion,TP391.41

CLC: > Transportation > Road transport > Road works > Roadbed, pavement engineering > General issues > Acceptance inspection
© 2012 www.DissertationTopic.Net  Mobile