Plan_Tracks_CostSlope.R

#R Package gdistance: Distances and Routes on
# Geographical Grids

##Easy low coast calculation
#Tutorial http://www.projectpanormos.com/sacredway/routeanalysis/
#https://dyerlab.github.io/applied_population_genetics/ecological-distance.html
##load packages

#load packages

loadpackages <- function(pkg) {
  new.pkg <- pkg[!(pkg %in% installed.packages()[, "Package"])]
  if (length(new.pkg))
    install.packages(new.pkg, dependencies = TRUE)
  sapply(pkg, require, character.only = TRUE)
}

# usage
packages <-
  c(
    "osmdata",
    "tidyverse",
    "sf",
    "gmap",
    "gdistance",
    "raster",
    "rasterVis",
    "rgdal",
    "gmap",
    "rgeos"
  )
loadpackages(packages)

##Input Data

setwd("G:/R_Project/Data/")
Startpoint <- readOGR(dsn = ".", layer = "StartJossgrund")
Startpoint.df <- as(Startpoint, "data.frame")

MKK_shape <- readOGR(dsn = ".", layer = "MKK_Kreis")
Clip_Region <- MKK_shape[2,]
ext <- extent(MKK_shape[2,])

SRTM = raster::raster("SRTM.tif")
LULC = raster::raster("Corine2018.tif")


#Set Variable - Study Adrea

AOI <- "Jossgrund"


##Download DEM

###add!!! see 3D Mapping

#Create Raster

SRTM_resam <- resample(SRTM, LULC, method = 'bilinear')

# r12 has the minimal common extent to crop
r12 = SRTM_resam + LULC

SRTM_Masked <- mask(SRTM_resam,  r12)
SRTM_Cropped <- crop(SRTM_Masked,  Clip_Region)

slope <-
  terrain(SRTM_Cropped, opt = 'slope', unit = 'degrees')  #calculate slope

plot(slope)


aspect <-
  terrain(SRTM_Cropped, opt = 'aspect', unit = 'degrees') #calculate aspect

#Load LULC
LULC_Masked <- mask(LULC,  r12)
LULC_Cropped <- crop(LULC_Masked,  Clip_Region)

#disaggregate from 40x40 resolution to 10x10 (factor = 4)

SRTM_Cropped <- disaggregate(SRTM_Cropped, fact = 10)
res(SRTM_Cropped)

slope <- disaggregate(slope, fact = 10)
res(slope)

LULC_Cropped  <- disaggregate(LULC_Cropped , fact = 10)
res(LULC_Cropped)


plot(LULC_Cropped)

###Reclassify LULC

#Urban Fabric
values(LULC_Cropped)[values(LULC_Cropped) == 111] = 9
values(LULC_Cropped)[values(LULC_Cropped) == 112] = 9

#Industrial, commercial and transport units
values(LULC_Cropped)[values(LULC_Cropped) == 121] = 10
values(LULC_Cropped)[values(LULC_Cropped) == 122] = 10
values(LULC_Cropped)[values(LULC_Cropped) == 123] = 10
values(LULC_Cropped)[values(LULC_Cropped) == 124] = 10

#Mine,dump and construction sites
values(LULC_Cropped)[values(LULC_Cropped) == 131] = 10
values(LULC_Cropped)[values(LULC_Cropped) == 132] = 10
values(LULC_Cropped)[values(LULC_Cropped) == 133] = 10

#Artificial, non agricultural vegetated areas

values(LULC_Cropped)[values(LULC_Cropped) == 141] = 8
values(LULC_Cropped)[values(LULC_Cropped) == 142] = 8

#Agricultural areas
#Arable land
values(LULC_Cropped)[values(LULC_Cropped) == 211] = 10
values(LULC_Cropped)[values(LULC_Cropped) == 212] = 10
values(LULC_Cropped)[values(LULC_Cropped) == 213] = 10

#Permanet crops
values(LULC_Cropped)[values(LULC_Cropped) == 221] = 8
values(LULC_Cropped)[values(LULC_Cropped) == 222] = 8
values(LULC_Cropped)[values(LULC_Cropped) == 223] = 8

#Pastures
values(LULC_Cropped)[values(LULC_Cropped) == 231] = 5

#Heterogeneous agricultural areas
values(LULC_Cropped)[values(LULC_Cropped) == 241] = 10
values(LULC_Cropped)[values(LULC_Cropped) == 242] = 8
values(LULC_Cropped)[values(LULC_Cropped) == 243] = 8
values(LULC_Cropped)[values(LULC_Cropped) == 244] = 8

#Forest and seminatural areas
#Forest
values(LULC_Cropped)[values(LULC_Cropped) == 311] = 5
values(LULC_Cropped)[values(LULC_Cropped) == 312] = 5
values(LULC_Cropped)[values(LULC_Cropped) == 313] = 5

#Shrub and/or herbaceous vegetations associations

values(LULC_Cropped)[values(LULC_Cropped) == 321] = 0
values(LULC_Cropped)[values(LULC_Cropped) == 322] = 0
values(LULC_Cropped)[values(LULC_Cropped) == 323] = 3
values(LULC_Cropped)[values(LULC_Cropped) == 324] = 2

#Open spaces with little or no vegetation

values(LULC_Cropped)[values(LULC_Cropped) == 331] = 0
values(LULC_Cropped)[values(LULC_Cropped) == 332] = 3
values(LULC_Cropped)[values(LULC_Cropped) == 333] = 2
values(LULC_Cropped)[values(LULC_Cropped) == 334] = 10
values(LULC_Cropped)[values(LULC_Cropped) == 335] = 10

#Wetlands
#Inland wetlands
values(LULC_Cropped)[values(LULC_Cropped) == 411] = 3
values(LULC_Cropped)[values(LULC_Cropped) == 412] = 3

#Coastal wetlands

values(LULC_Cropped)[values(LULC_Cropped) == 421] = 10
values(LULC_Cropped)[values(LULC_Cropped) == 422] = 10
values(LULC_Cropped)[values(LULC_Cropped) == 423] = 10

#Waterbodies
#Inland waters
values(LULC_Cropped)[values(LULC_Cropped) == 511] = 200
values(LULC_Cropped)[values(LULC_Cropped) == 512] = 200

#Marine waters
values(LULC_Cropped)[values(LULC_Cropped) == 521] = 200
values(LULC_Cropped)[values(LULC_Cropped) == 522] = 200
values(LULC_Cropped)[values(LULC_Cropped) == 523] = 200

plot(LULC_Cropped)

rm(SRTM, SRTM_resam, r12, LULC_Masked, SRTM_Masked, LULC)

#coords <- c(xyFromCell(SRTM_Cropped, LULC_Cropped[i], spatial = FALSE))

#view <- viewshed(SRTM_Cropped, ext, coords, h1 = 0, h2 = 0)

#Download POIs, RoadNetwork

###DEM
###View
###Roads with different value

#availabel features in OSM
#the first five features
head(available_features())
available_features()

#amenities
available_tags("route")
available_tags("highway")
available_tags("amenity")
available_tags("water")

#shops
head(available_tags("shop"))

#########Get OSM Data

library(osmdata)

##Create a Loop

#OSM_Data (One Feature)


q1 <- opq(AOI) %>%
  add_osm_feature(key = 'waterway')
waterway <- osmdata_sp(q1)
waterway <- waterway$osm_lines

q1 <- opq(AOI) %>%
  add_osm_feature(key = 'natural', 'spring')
spring <- osmdata_sp(q1)
spring <- spring$osm_points

q1 <- opq(AOI) %>%
  add_osm_feature(key = 'highway')
highway <- osmdata_sp(q1)
highway <- highway$osm_lines

####Transfer to Raster
waterway <-
  spTransform(
    waterway,
    "+proj=utm +zone=32 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0"
  )

waterway <- mask(LULC_Cropped, waterway)
waterway[is.na(waterway[])] <- 0
values(waterway)[values(waterway) >= 0] = 100
plot(waterway)



spring <-
  spTransform(
    spring,
    "+proj=utm +zone=32 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0"
  )

spring <- mask(LULC_Cropped, spring)
values(spring)[values(spring) >= 0] = 0
spring[is.na(spring[])] <- 10
plot(spring)


highway <-
  spTransform (
    highway,
    "+proj=utm +zone=32 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0"
  )

highway <- mask(LULC_Cropped, highway)
values(highway)[values(highway) >= 0] = 0
highway[is.na(highway[])] <- 300
plot(highway)


#Transfer Raster in Dataframe

LULC_Cropped_temp <- rasterToPoints(LULC_Cropped)

#Make the points a dataframe for ggplot
LULC_df <- data.frame(LULC_Cropped_temp)
#Make appropriate column headings
colnames(LULC_df) <- c("Longitude", "Latitude", "Class")

ggplot() +
  geom_raster(data = LULC_df, aes(x = Latitude, y = Longitude, fill = Class)) +
  scale_colour_gradientn(colours = terrain.colors(10)) +
  geom_point(data = Startpoint.df,
             aes(x = coords.x2, y = coords.x1),
             color = "green") +
  theme_bw() +
  coord_equal() +
  scale_fill_gradient("LULC", limits = c(0, 10)) +
  theme(
    axis.title.x = element_text(size = 16),
    axis.title.y = element_text(size = 16, angle = 90),
    axis.text.x = element_text(size = 14),
    axis.text.y = element_text(size = 14),
    panel.grid.major = element_blank(),
    panel.grid.minor = element_blank(),
    legend.position = "right",
    legend.key = element_blank()
  )

########### Create Cost Raster

cost <- stack(LULC_Cropped,highway)
cost1 <- calc(cost, sum)
#cost1 <- mask(cost1, highway)

cost1_temp <- rasterToPoints(cost1)

#Make the points a dataframe for ggplot
cost1_df <- data.frame(cost1_temp)
#Make appropriate column headings
colnames(cost1_df) <- c("Longitude", "Latitude", "Class")


#Random Startpoint
#set.seed(123)

#Startpoint <- sampleRandom(cost1, size=1, cells=TRUE, sp=TRUE)
#Route <- as.data.frame(Startpoint.df)

Route <- raster::extract(cost1, Startpoint.df[1, 2:3], df = TRUE, cellnumbers =T)


Route <- cbind(Route, Startpoint.df[1, 2:3])

length_route <- 500

for (i in 1:length_route) {
  if (i <= (length_route)) {
    rowcol <- rowColFromCell(cost1, Route$cell[i])
    
    cell1 <-
      cellFromRowCol(cost1, c(rowcol[1, 1]) - 1, rowcol[1, 2])
    cell2 <-
      cellFromRowCol(cost1, c(rowcol[1, 1]) - 1, c(rowcol[1, 2]) -
                       1)
    cell3 <-
      cellFromRowCol(cost1, c(rowcol[1, 1]) - 1, c(rowcol[1, 2]) +
                       1)
    cell4 <-
      cellFromRowCol(cost1, rowcol[1, 1], c(rowcol[1, 2]) + 1)
    #cell5 <-
      cellFromRowCol(cost1, c(rowcol[1, 1]) + 1, c(rowcol[1, 2]) +
                       1)
    #cell6 <- cellFromRowCol(cost1, c(rowcol[1,1])+1, c(rowcol[1,2])-1)
    #cell7 <-
      cellFromRowCol(cost1, c(rowcol[1, 1]) + 1, rowcol[1, 2])
    cell8 <-
      cellFromRowCol(cost1, rowcol[1, 1], c(rowcol[1, 2]) - 1)
    
    cells <-
      as.data.frame(rbind(cell1, cell2, cell3, cell4, cell8))
    #cells <- as.data.frame(rbind(cell1,cell2,cell3,cell4,cell5,cell6,cell7,cell8))
    
    xy <- as.data.frame(xyFromCell(cost, cells$V1, spatial = FALSE))
    
    cells <- cbind(cells, xy)
    
    cent_max <-
      raster::extract(cost1, cells[, 2:3], df = TRUE, cellnumbers = T)
    
    if (is.na(cent_max$cells[1])) {
      cell3 <-
        cellFromRowCol(cost1, c(rowcol[1, 1]) - 1, c(rowcol[1, 2]) + 1)
      cell4 <-
        cellFromRowCol(cost1, rowcol[1, 1], c(rowcol[1, 2]) + 1)
      cell5 <-
        cellFromRowCol(cost1, c(rowcol[1, 1]) + 1, c(rowcol[1, 2]) + 1)
      cells <- as.data.frame(rbind(cell3, cell4, cell5))
      
      xy <-
        as.data.frame(xyFromCell(cost, cells$V1, spatial = FALSE))
      
      cells <- cbind(cells, xy)
      
      cent_max <-
        raster::extract(cost1, cells[, 2:3], df = TRUE, cellnumbers = T)
      
      cent_max = subset(cent_max, !(cent_max$cells %in% Route$cell))
      
      cent_max <-
        cent_max[order(cent_max$layer, decreasing = FALSE), ]
      
      Route[i + 1, 4:5] <- xyFromCell(cost1, cent_max$cells[1], spatial = FALSE)
      
      Route[i + 1, 2:3] <- cent_max[1, 2:3]
      
    } else {
      cent_max = subset(cent_max, !(cent_max$cells %in% Route$cell))
      
      cent_max = subset(cent_max, !is.na(cent_max$layer))
      
      cent_max <-
        cent_max[order(cent_max$layer, decreasing = FALSE), ]
      
      Route[i + 1, 4:5] <-
        xyFromCell(cost1, cent_max$cells[1], spatial = FALSE)
      
      Route[i + 1, 2:3] <- cent_max[1, 2:3]
    }
  } else {
    cell1 <- cellFromRowCol(cost1, c(rowcol[1, 1]) - 1, rowcol[1, 2])
    cell2 <- cellFromRowCol(cost1, c(rowcol[1, 1]) - 1, c(rowcol[1, 2]) -1)
    #cell3 <- cellFromRowCol(cost1, c(rowcol[1, 1]) - 1, c(rowcol[1, 2]) +1)
    #cell4 <- cellFromRowCol(cost1, rowcol[1, 1], c(rowcol[1, 2]) + 1)
    #cell5 <- cellFromRowCol(cost1, c(rowcol[1, 1]) + 1, c(rowcol[1, 2]) +1)
    cell6 <- cellFromRowCol(cost1, c(rowcol[1,1])+1, c(rowcol[1,2])-1)
    cell7 <-cellFromRowCol(cost1, c(rowcol[1, 1]) + 1, rowcol[1, 2])
    cell8 <-cellFromRowCol(cost1, rowcol[1, 1], c(rowcol[1, 2]) - 1)
    
    cells <-
      as.data.frame(rbind(cell1,cell2, cell6, cell7, cell8))
    #cells <- as.data.frame(rbind(cell1,cell2,cell3,cell4,cell5,cell6,cell7,cell8))
    
    xy <- as.data.frame(xyFromCell(cost, cells$V1, spatial = FALSE))
    
    cells <- cbind(cells, xy)
    
    cent_max <-
      raster::extract(cost1, cells[, 2:3], df = TRUE, cellnumbers =T)
    
    if (is.na(cent_max$cells[1])) {
      cell3 <-
        cellFromRowCol(cost1, c(rowcol[1, 1]) - 1, c(rowcol[1, 2]) + 1)
      cell4 <-
        cellFromRowCol(cost1, rowcol[1, 1], c(rowcol[1, 2]) + 1)
      cell5 <-
        cellFromRowCol(cost1, c(rowcol[1, 1]) + 1, c(rowcol[1, 2]) + 1)
      cells <- as.data.frame(rbind(cell3, cell4, cell5))
      
      xy <-
        as.data.frame(xyFromCell(cost, cells$V1, spatial = FALSE))
      
      cells <- cbind(cells, xy)
      
      cent_max <-
        raster::extract(cost1, cells[, 2:3], df = TRUE, cellnumbers =
                          T)
      
      cent_max = subset(cent_max, !(cent_max$cells %in% Route$cell))
      
      cent_max <- cent_max[order(cent_max$layer), ]
      
      length_cent_max <- length(cent_max)
      
      Route[i + 1, 4:5] <-
        xyFromCell(cost1, cent_max$cells[length_cent_max], spatial = FALSE)
      
      Route[i + 1, 2:3] <- cent_max[length_cent_max, 2:3]
      
    } else {
      cent_max = subset(cent_max, !(cent_max$cells %in% Route$cell))
      
      cent_max = subset(cent_max, !is.na(cent_max$layer))
      
      cent_max <- cent_max[order(cent_max$layer, decreasing = FALSE), ]
      
      Route[i + 1, 4:5] <-
        xyFromCell(cost1, cent_max$cells[1], spatial = FALSE)
      
      Route[i + 1, 2:3] <- cent_max[1, 2:3]
    }
  }
}

ggplot() +
  geom_raster(data = cost1_df,
              aes(x = Latitude,
                  y = Longitude,
                  fill = Class)) +
  scale_colour_gradientn(colours = terrain.colors(300)) +
  geom_point(data = Route,
             aes(x = coords.x2, y = coords.x1),
             color = "green")


xy <- Route[,c(4,5)]

Route_Spatial <- SpatialPointsDataFrame(
  coords = xy,
  data = Route,
  proj4string = CRS(
    "+proj=utm +zone=32 +datum=WGS84 +units=m +no_defs +ellps=WGS84 +towgs84=0,0,0"
  )
)

Route_Spatial2 <- gBuffer(Route_Spatial,width=500)


Route_Spatial3 <- mask(LULC_Cropped, Route_Spatial2)
Route_Spatial3[is.na(Route_Spatial3[])] <- 0
values(Route_Spatial3)[values(Route_Spatial3) > 0] = 1000
plot(Route_Spatial3)


#writeOGR(
#  obj = spdf,
#  dsn = "tempdir",
#  layer = "spdf",
#  driver = "ESRI Shapefile"
#)
#plot(spdf)

cost2 <- stack(cost1,Route_Spatial)
cost2 <- calc(cost2,sum)

###Least Cost Path
trCost <- transition(cost2, transitionFunction = min, 8)
trCost2 <- geoCorrection(trCost, type = "c")


Start_Point <- c(Route[length_route,4:5])
End_Point <- c(Startpoint.df[1,2:3])
cost2 <- accCost(trCost2, c(Start_Point$coords.x1,Start_Point$coords.x2))

cost3 <- costDistance(trCost2, c(Start_Point$coords.x1,Start_Point$coords.x2), 
                      c(End_Point$coords.x1,End_Point$coords.x2))

path.1 <- shortestPath(trCost2, c(Start_Point$coords.x1,Start_Point$coords.x2), 
                       c(End_Point$coords.x1,End_Point$coords.x2), output = "SpatialLines")
plot(path.1)

path.1.df <-
  SpatialLinesDataFrame(path.1, data = data.frame(ID = 1))

path.1.df_fortify <- fortify(path.1.df)


###Plot Least Coth Path on Cost Raster
#Transfer Raster in Dataframe

cost1_temp <- rasterToPoints(cost1)

#Make the points a dataframe for ggplot
cost1_df <- data.frame(cost1_temp)
#Make appropriate column headings
colnames(cost1_df) <- c("Longitude", "Latitude", "Class")

ggplot() +
  geom_raster(data = cost1_df,
              aes(
                x = cost1_df$Latitude,
                y = cost1_df$Longitude,
                fill = cost1_df$Class
              )) +
  scale_colour_gradientn(colours = rainbow) +
  geom_point(data = Startpoint.df,
             aes(x = Startpoint.df$coords.x2, y = Startpoint.df$coords.x1),
             color = "black") +
  geom_point(data = path.1.df_fortify,
            aes(x = path.1.df_fortify$lat, y = path.1.df_fortify$long),
            color = "red") +
  geom_point(data = Route,
             aes(x = coords.x2, y = coords.x1),
             color = "red")+
  theme_bw() +
  coord_equal() +
  scale_fill_gradient("Cost", limits = c(0, 500)) +
  theme(
    axis.title.x = element_text(size = 16),
    axis.title.y = element_text(size = 16, angle = 90),
    axis.text.x = element_text(size = 14),
    axis.text.y = element_text(size = 14),
    panel.grid.major = element_blank(),
    panel.grid.minor = element_blank(),
    legend.position = "right",
    legend.key = element_blank()
  )

##merge Dataframe Lines, extract values, change cost, that the existing route won't be used for
#wayback, length, extract landuses, slope, include slope in calculation, plot 3D -> Gif and Img,buffer 
#um Route 



#####################

#data(volcano)
#library(spatstat)
#LLC <- data.frame(E = 174.761345, N = -36.879784)
#coordinates(LLC) <- ~ E + N
#proj4string(LLC) <- CRS("+proj=longlat +datum=WGS84")
#LLC.NZGD49 <- spTransform(LLC, CRS("+init=epsg:27200"))
#volcano.r <- as.im(list(
#  x = seq(
#    from = 2667405,
#    length.out = 61,
#    by = 10
#  ),
#  y = seq(
#    from = 6478705,
#    length.out = 87,
#    by = 10
#  ),
##  z = t(volcano)[61:1,]
#))
#volcano.sp <- as(volcano.r, "SpatialGridDataFrame")
#proj4string(volcano.sp) <- CRS("+init=epsg:27200")
#str(volcano.sp)
#spplot(volcano.sp,
#       at = seq(min(volcano.sp$v), max(volcano.sp$v), 5),
#       col.regions = topo.colors(45))

#r <- raster(volcano.sp)

#altDiff <- function(x)
#  x[2] - x[1]
#hd <- transition(r, altDiff, 8, symm = FALSE)

#slope <- geoCorrection(hd)

#adj <- adjacent(r,
#                cells = 1:ncell(r),
#                pairs = TRUE,
#                directions = 8)
#speed <- slope
#speed[adj] <- 6 * exp(-3.5 * abs(slope[adj] + 0.05))

#Conductance <- geoCorrection(speed)

#A <- c(2667670, 6479000)
#B <- c(2667800, 6479400)
#AtoB <- shortestPath(Conductance, A, B, output = "SpatialLines")
#BtoA <- shortestPath(Conductance, B, A, output = "SpatialLines")


#plot(r,
#     xlab = "x coordinate (m)",
#     ylab = "y coordinate (m)",
#     legend.lab = "Altitude (masl)")
#lines(AtoB, col = "red", lwd = 2)
#lines(BtoA, col = "blue")
#text(A[1] - 10, A[2] - 10, "A")
#text(B[1] + 10, B[2] + 10, "B")