一种基于飞行时间相机的点云目标提取方法

基于飞行时间(Time of Flight,TOF)原理的深度相机成像方法不同于二维图像来计算三维信息,而是通过光在空气中的飞行时间,来计算出目标的距离,从而直接获取场景目标的三维点云信息。本文通过研究基于飞行时间红外相机的三维重建技术,设计了一种基于飞行时间红外相机的点云目标提取方法。利用飞行时间相机直接获得场景的三维点云数据,提出一种双阈值空间滤波算法,对点云数据进行空间滤波,并对滤波效果进行了对比评价。在双阈值空间滤波算法的基础上提出了一种改进的基于法向量的随机抽样一致性(RANSAC)算法,实现了对三维点云数据的目标提取,为基于飞行时间相机的场景目标三维重建奠定了基础。

A point cloud object extraction method based on time-of-flight camera

The depth camera imaging method based on time of flight (TOF) principle is different from the two-dimensional image to calculate the three-dimensional information.It calculates the distance of the target through the time of light flying in the air,so as to obtain the three-dimensional point cloud information of the scene target directly.In this paper,a point cloud target extraction method based on time-of-flight infrared camera is designed by studying the three-dimensional reconstruction technology based on time-of-flight infrared camera.Using the time-of-flight camera to directly obtain the three-dimensional point cloud data of the scene,a double threshold spatial filtering algorithm is proposed to filter the point cloud data in space,and the filtering effect is compared and evaluated.An improved Random Sampling Consistency (RANSAC) algorithm based on Normal Vector is proposed on the basis of double threshold spatial filtering algorithm,which realizes the target extraction of 3D point cloud data and lays the foundation for 3D reconstruction of scene target based on time-of-flight camera.