Simulation of Lidar Target Orientation Estimation Based on Point Clouds Model Matching

摘要

Thee-dimensional (3-D) laser active imaging technology widely used in military and civilian applications for its advantages of high energy density, coherence and directionality. 3-D pose estimation using laser array imaging system involves finding the rotation and translation between image point clouds and model point clouds. ()wing to limitations such as detector technology and other factors, it is difficult to achieve high-resolution 3-D laser imaging systems. This article introduces a laser array imaging simulation method, and the pose estimation process is simulated. In the simulation, different range and spatial resolutions of the imaging system are discussed to analyse the influence of these imaging system factors on the accuracy of 3-D pose estimation. The results show that for a target size of 4 x 4 x 6m(3), the accuracy of orientation estimation could be 0.10 degrees when the ranging resolution of the system is 0.35 m. For a ranging resolution of 1.50 m and spatial resolution of 0.15 m, the accuracy of orientation estimation could be 0.50 degrees. The simulation results further show that the impact of the range resolution on the accuracy of pose estimation is greater than that of the spatial resolution. We can, therefore, reduce the spatial resolution and increase the range resolution of the imaging system to improve the accuracy of the pose estimation for certain special cases. The results of this analysis could be used in the design of systems with different needs, which makes the design more efficient in terms of the use of resources. The simulation method and analysis can be used in other applications, such as point clouds registration and lidar target recognition based on point clouds-model matching.