CODEWITHSUNDEEP
rmatrixgisdistance
Mery_evo
Mery_evo

Reputation: 1

Error in R package gdistance while calculating a matrix distance

I am trying to compute a matrix distance between points using the function costDistance is the R package gdistance, taking into account the altitude with a transition matrix with a code like this:

read.csv("subset.csv",h=T)->dat
llCRS <- CRS("+proj=longlat +ellps=WGS84")
dat_mat<- cbind(dat$Long, dat$Lat)
dat_sp <- SpatialPointsDataFrame(dat_mat, dat, proj4string = llCRS)
raster("subset.grd")->alt
alt[is.na(alt)] <- 0
heightDiff <- function(x){x[2] - x[1]}
tr <- transition(alt,heightDiff,8,symm=FALSE)
pC <- as.matrix(dat[c("Longitud", "Latitud")])
cosDist <- costDistance(tr, pC)

and I get this error message

Error in .Call("R_igraph_shortest_paths", graph, v - 1, to - 1, as.numeric(mode),  : 
  At structural_properties.c:5200 : cannot run Bellman-Ford algorithm, Negative loop detected while calculating shortest paths

If anyone knows why, it would be very helpful.

Upvotes: 0

Views: 750

Answers (1)

JacobVanEtten
JacobVanEtten

Reputation: 203

In your example, the transition function should throw a warning that negative values occur.

This is how to avoid it:

heightDiff <- function(x){abs(x[2] - x[1])}

Normally you need to use function geoCorrection, something you omit above.

Upvotes: 1

Related Questions