论文部分内容阅读
以激光雷达三维点云对弹头目标形状、姿态、位置估计与目标识别为应用背景。针对导弹弹头的形状特点,将其近似建模为一个圆锥,并可通过6个参数来表征。提出了基于激光成像雷达得到的三维点云与Levenberg-Marquardt(L-M)算法的圆锥状目标参数估计方法,并指出这种方法也可用于对圆锥状目标的识别。通过仿真实验,给出了弹头参数估计误差与三维点云的距离分辨率、点云规模的关系。仿真和实测数据的实验结果表明:该方法在一定的观测距离分辨率下可以较精确地估计出圆锥状弹头目标的参数,并可以根据得到的参数以及对目标函数的优化结果,对目标进行识别。
Based on the lidar 3D point cloud, the warhead target shape, attitude, position estimation and target recognition are used as background. According to the shape characteristics of the missile warhead, it is approximately modeled as a cone and can be characterized by six parameters. A conical target parameter estimation method based on 3D point cloud and Levenberg-Marquardt (L-M) algorithm based on laser imaging radar is proposed. It is also pointed out that this method can also be used to identify conical targets. Through the simulation experiments, the relationship between the warhead parameter estimation error and the distance resolution of 3D point cloud and the point cloud size are given. The experimental results show that this method can estimate the parameters of cone-shaped warhead targets more accurately at a certain resolution of observation distance and can identify the targets according to the obtained parameters and the optimization results of the objective function .