ground_extarct_in_country_road

基于激光线性特征进行可行驶路径粗提取

• VLP32C mode(当前), dual VLP16 mode, VLP32C+ dual VLP16 mode

• dependency：

  ecl-geometry包

  sudo apt install ros-kinetic-ecl-geometry

1. VLP32C line feature mode (currently adopt)

   1. 只接受未补偿激光原始32线数据作为输入，因激光角度值用于激光线数判断和每点yaw值计算

   2. ground_detect.launch
      	所选用最低的线数
      	<arg name="low_lines" default="4" />
      	判断为路面的线段距离主线段的最大z轴差距
          <arg name="z_threshold" default="0.06" />
      	成为线段间相邻点的距离阈值（key param）
          <arg name="distance_threshold" default="0.20" />
      	采用输出的坐标系选择
      	<!-- map_tf:1. velo_middle  2. geometry_center-->
          <arg name="map_tf" default="geometry_center" />
      	成为线段的最小点数
          <arg name="lowest_segment_point_num" default="12" />
      	路径生成中噪点忽略阈值
          <arg name="abandon_min_points_num" default="6" />

   3. roslaunch ground_detect ground_detect.launch

   4. • /lidar_line_nanlyse_node:

        将原始激光点/lidar/vlp32_middle/PointCloud2进行线数判断(复用intensity字段整数部分)，yaw角计算(复用intensity字段小数部分)，距离选取，并根据上一帧路面点的每一线的角度范围/ground_detect/road_range，在当前帧中筛选潜在路面点/ground_detect/forward_points后发布

      • /feature_based_select_node:

        在潜在路面点/ground_detect/forward_points中寻找线段组并进行合并作为每一线的路面点，，计算每一线的路面点的角度范围/ground_detect/road_range并发布，以及选定路面点/ground_detect/ground_points

      • /path_provider_node:

        根据选定路面点/ground_detect/ground_points求取每一线点的重心，存入缓冲区，并依据多帧结果给出当前帧当前线的路面重心点作为路径点。对所选中线数的路径点进行三次样条拟合，并生成path后发布/ground_detect/Path_curved，拟合前path/ground_detect/Path_original

      

2. dual VLP16 mode

   1. roslaunch ground_detect ground_detect_2VLP16.launch

      

3. VLP32C+ dual VLP16 grid map mode

   1. roslaunch ground_detect ground_detect_grid_map.launch

      

4. VLP32C+ dual VLP16 ransac get_plane mode

   1. roslaunch ground_detect ground_detect_ransac.launch