Dissertation > Excellent graduate degree dissertation topics show

A typical three-dimensional laser-based radar target recognition technology research ground

Author: LiuYao
Tutor: MaJie
School: Huazhong University of Science and Technology
Course: Pattern Recognition and Intelligent Systems
Keywords: Imaging Lidar 3D object recognition Spin Figure Point cloud Surface Reconstruction
CLC: TP391.41
Type: Master's thesis
Year: 2011
Downloads: 111
Quote: 0
Read: Download Dissertation

Abstract


Imaging laser radar is a high distance resolution, angular resolution and accuracy of the new high-precision imaging sensors. Study is based on a three-dimensional laser radar Automatic Target Recognition technology has important military application value and theoretical significance. In this paper, laser imaging radar as the background to the three-dimensional laser imaging radar simulation model for the experimental data in order to spin Pictured target features, a similar measure for the matching algorithm to study the three-dimensional surface graph based on the spin Automatic Target Recognition. The main contents are as follows: (a) in the three-dimensional target data acquisition and processing, respectively, for ground truth data and loaded three-dimensional model of the laser beam intersection point coordinates and scene solving process, the simulation of three-dimensional point cloud data ; discuss laser imaging radar data of expression, and select the appropriate spin on this graph algorithms irregular triangular mesh represent three-dimensional model; proposed target 3D point cloud data extraction methods and three-dimensional surface reconstruction method. (2) spin-based object recognition algorithms graph, diagram describes the generation of the spin, the target scene and the model proposed target recognition four discrimination index, and the index for the identification, recognition criterion proposed combination; combination recognition criteria can be different imaging conditions for adaptive adjustment, can achieve the balance of performance and efficiency. The target recognition algorithm main idea is: the three-dimensional laser imaging radar point cloud data into two-dimensional spin-drawing space, then three-dimensional object recognition problem is converted into two-dimensional drawings and models spin spin scene graph matching problem. (3) analysis of the three-dimensional movement of the imaging model target and the target type of geometric distortion, and derive the target size and geometric distortion between the target motion parameters, image geometric distortion of the feature extraction. (4) In the three-dimensional object recognition experiments, first established a typical ground target database, followed by the spatial resolution of the target, the laser radar ranging noise, target occlusion rates and geometric distortion of the four aspects of the imaging target recognition results. Experimental results demonstrate that the proposed target recognition algorithm. This study can provide three-dimensional laser radar target recognition system is designed to provide a reference.

Related Dissertations

  1. Based on the geometry of multiple images and textures automatically rebuild,TP391.41
  2. Large spiral bevel gear error analysis and correction techniques,TH132.41
  3. Study on Algorithms of Airborne LiDAR Point Cloud Data Filtering,TN713
  4. Research and Implementation on Surfaces Reconstruction,TP391.41
  5. Scattered point cloud model triangle mesh processing algorithms to achieve,TP391.41
  6. Research on Skeletonization of Low-quality Chinese Characters,TP391.41
  7. Theoretical analysis of the crystal surface reconstruction and relaxation,O469
  8. The Surface Modeling of Partial Car Body Based on Augment Reality,TP391.41
  9. Volumetric Stereo and Silhouette Fusion for Image-based Modeling,TP391.41
  10. The Properties of Superconductor, Reconstruction and Surface Alloy of Semiconductor Surface,TN304.01
  11. The Research of Three-dimension Face Data Fusion,TP391.41
  12. Research on Point Cloud Matching Method Based on Dual Quaternion Algorithm,TP391.41
  13. Research on Tunnel Surface Reconstruction Based on Laser Scan Data,U452.1
  14. Based on the the RE technology of plastic products forming,TQ320.6
  15. Estimation of Differential Quantities on Point Clouds and Applications,TP391.41
  16. Research about Surface Reconstruction Based on CMM Measuring Technology,TH164
  17. The Calibration and Point Cloud Processing in 3D Reconstruction,TP391.41
  18. STL-based spiral bevel gear cutting motion visualization,TH132.41
  19. Research of Tractor Modeling Method Based on Reverse Engineering,S219.02
  20. Research on the Application of 3D Laser Scanner and Infrared Thermography in the Rock Weathering,P575

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