viewsheds: Computes cumulated viewsheds

In Blecigne/viewshed3d: Compute Viewshed in 3D Point Clouds of Ecosystems

Computes cumulated viewsheds

Description

Computes cumulated viewsheds within a 3D point cloud and return a voxel cloud accounting for the number of times each voxel was visible.

Usage

viewsheds(data, positions, angular_res, vox_res, cut_off, pb)

Arguments

data	LAS class object containing the xyx coordinates of a 3D point cloud.
positions	data.frame or data.table with 3 columns containing the xyz coordinates of the animal locations from which the viewsheds will be computed.
angular_res	numeric. The angular resolution of a single sightline. Default = 1.
vox_res	numeric. The resolution of the output voxel cloud. Default = 0.2.
cut_off	(optional) numeric. Defines a cut-off distance for each individual viewshed. Speeds up the process when viewsheds is applied to big datasets.
pb	logical. If FALSE, desables the progress bar.

Details

Sightline directions in each viewshed are computed from the method described by Malkin (2016). This ensures that every sightline explores a similar portion of the 3d space.

Value

A LAS class object containing the coordinates of the voxel cloud (X, Y, Z), and the number of times each voxel was visible from any position (N_visible).

Note

In most cases, a ground reconstruction should be performed before viewsheds computation. This can be done with the classify_ground and grid_terrain functions from the lidR-package.

References

Malkin, Z. (2016). A new method to subdivide a spherical surface into equal-area cells. arXiv:1612.03467.

Examples

#- import the tree_line_plot dataset
file = system.file("extdata", "tree_line_plot.laz", package="viewshed3d")
tls = lidR::readLAS(file,select="xyz")

#- remove noise to avoid visibility estimates error
tls_clean = viewshed3d::denoise_scene(tls,method="sd",
                                       filter=6)

#- RECONSTRUCT THE GROUND
#- classify ground points
class=lidR::classify_ground(tls_clean, lidR::csf(rigidness = 1L,
                                           class_threshold = 0.1,
                                           sloop_smooth = TRUE), FALSE)

#- reconstruct the ground. No need for a very fine ground reconstruction.
ground = lidR::grid_terrain(class, 0.05, lidR::knnidw())

#- build the final scene
reconstructed = na.omit(raster::as.data.frame(ground, xy = TRUE))
names(reconstructed)=c("X","Y","Z")
recons=rbind(lidR::LAS(na.omit(reconstructed)),tls_clean)

#- CREATE THE POSITIONS WITH RANDOM POINTS
N_positions = 10 #- how many points ?
height = 2 #- points height relative to the ground
positions=data.table::data.table(reconstructed[runif(N_positions,
                                               1,nrow(reconstructed)),])
positions[,Z:=Z+height]

#- compute the cumulated viewsheds from the positions
cumulated=viewshed3d::viewsheds(data=recons,
                                positions = positions ,
                                angular_res = 1,
                                vox_res = 0.2)

#- plot the result
x=lidR::plot(cumulated,color="N_visible",size=3,
             colorPalette=viridis::cividis(nrow(positions)+1),trim=6)

#- add the positions
lidR::add_treetops3d(x,sp::SpatialPointsDataFrame(positions,positions),
                     radius=0.5,col="red",add=TRUE)

Blecigne/viewshed3d documentation built on Nov. 5, 2022, 9:08 p.m.