viewshed3d (version 3.2.0)

visibility: Computes the visibility from a single location in a 3D point cloud

Description

Computes visibility from a user-defined location with user-defined sightline angles and returns the visibility as function of distance and, optionally, the 3D point cloud classified as visible and non-visible points.

Usage

visibility(data, position, angular_res, scene_radius, store_points)

Arguments

data

LAS class object containing the xyz coordinates of a 3D point cloud

position

vector of length 3 containing the xyz coordinates of the animal location. Default = c(0,0,0).

angular_res

numeric. The angular resolution of a single sightline. Default = 1.

scene_radius

(optional) numeric. Defines the radius of the scene relative to the animal position. Can be used to apply a cut-off distance to visibility analyses.

store_points

logical. If TRUE, the 3D point cloud is returned with visible and not visible points classified (see details).

Value

If store_points = FALSE, a data.table of the visibility (Visibility) as a function of distance to the animal location (r) is returned. If store_points = TRUE, a list containing two objects is returned. The first object is similar to the data.table returned when store_points = FALSE. The second object is a LAS class object containing the coordinates of the point cloud (X, Y, Z), the distance of each point to the animal position (r) and the class of each point: visible or not visible from the animal position (Visibility = 2 or 1, respectively).

Details

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

References

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

Examples

# NOT RUN {
#- import the tree_line_plot dataset
file <- system.file("extdata", "tree_line_plot.laz", package="viewshed3d")
tls <- lidR::readLAS(file)

center <- c(0,0,2) # defines the scene center for the entire process
angle <- 1 # defines the angular resolution for the entire process

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


#- class ground and vegetation points
class <- lidR::classify_ground(tls_clean, lidR::csf(rigidness = 1L,
                                              class_threshold = 0.2,
                                              sloop_smooth = FALSE))

#- reconstruct the ground with the optimal resolution
recons <- viewshed3d::reconstruct_ground(data=class,
                                         position = center,
                                         ground_res = 0.05,
                                         angular_res = angle,
                                         method="knnidw",
                                         full_raster = TRUE)

#- compute the visibility and store the output point cloud.
#- As the input file is a LAS object, the output
#- point cloud is also stored in a LAS file.
view.data <- viewshed3d::visibility(data = recons,
                                    position = center,
                                    angular_res = angle,
                                    scene_radius = 17, # apply cut_oof distance
                                    store_points = TRUE)

#- 3D plot with visible points in white and non-visible points in darkgrey
x=lidR::plot(view.data$points,color="Visibility",colorPalette = c("grey24","white"))

#- add animal position to the plot
position=data.frame(X=center[1],Y=center[2],Z=center[3])
lidR::add_treetops3d(x,sp::SpatialPointsDataFrame(position,position),
                     radius=0.2,col="red")

#- plot the visibility as function of distance
plot(view.data$visibility$r,view.data$visibility$visibility,
     type="l",ylim=c(0,100),lwd=4)
# }
# NOT RUN {
# }