NormalEstimationIntegralImage

What is this?

Compute normals for an organized point cloud using integral images.

Subscribing Topics

• ~input (sensor_msgs/PointCloud2)

  Input point cloud. (point type is pcl::PointXYZRGB)

Publishing Topics

• ~output (sensor_msgs/PointCloud2)

  Output normals. (point type is pcl::Normal)

• ~output_with_xyz (sensor_msgs/PointCloud2)

  Output point cloud with normal field. (point type is pcl::PointXYZRGBNormal)

Parameters

• max_depth_change_factor (Double, default: 0.02)

  max depth change factor

• normal_smoothing_size (Double, default: 20.0)

  normal smoothing size parameter

• estimation_method (Int, default: 1)

  Estimation method.

  • 0: AVERAGE_3D_GRADIENT
  • 1: COVARIANCE_MATRIX
  • 2: AVERAGE_DEPTH_CHANGE

• depth_dependent_smoothing (Boolean, default: false)

  Use depth dependent smoothing.

• border_policy_ignore (Boolean, default: true)

  Ignore border policy.

Parameters above can be changed by dynamic_reconfigure.

Sample

roslaunch jsk_pcl_ros sample_normal_estimation_integral_image.launch