SQLite as an alternative to shapefiles, and some GPS fun in R

Submitted by dylan on Tue, 2010-05-25 04:15.

Finally made it out to Folsom Lake for a fine day of sailing and GPS track collecting. Once I was back in the lab, I downloaded the track data with gpsbabel, and was ready to import the data into GRASS.

# import GPX from GPS:
gpsbabel -t -i garmin -f /dev/ttyS0 -o gpx -F trip1.gpx

I was particularly interested in how fast we were able to get the boat going, however my GPS does not keep track of its speed in the track log. Also, I had forgotten to set GPS up for a constant time interval between track points. Dang. In order to compute a velocity between sequential points from the track log I would need to first do two things: 1) convert the geographic coordinates into projected coordinates, and 2) compute the associated time interval between points.

The GDAL/OGR tools provide one of the simplest means of simultaneously converting between formats AND coordinate systems. The GPX file contains both a line and point representation of the log; the point version seems simpler to work with. Initially I used the default output format (ESRI shapefile), however I quickly discovered that the datetime information from the GPX file was silently truncated to just the date part. Turns out that GDAL/OGR can now write vector data inSQLite format-- storing both the geometry and attributes into a single file, with a table structure very similar to PostGIS. Seems like a great replacement for shapefiles, that can deal with complex data types, long field names, and probably very large geometries.

# convert into sqlite DB with geometry-- better preservation of attributes
ogr2ogr -F "SQLite" -t_srs '+proj=utm +zone=10 +datum=NAD83' trip1_track_points.db trip1.gpx track_points

# import, we are using a sqlite DB back-end
v.in.ogr dsn=trip1_track_points.db out=folsom_trip1_pts

Next, we import the data (in projected coordinates) into a GRASS location/mapset, with a SQLite database back-end. The data are now in a format that can be easily imported into R for velocity calculations, and then sent back to GRASS for further processing. An example R session is presented below. Track points are in chronological order, so we can use the diff() function for computing sequential time and distance intervals.

Velocity vs. BearingVelocity vs. Bearing

 
Further rocessing in R

library(spgrass6)
library(RColorBrewer)
library(plotrix)


# read vector points
# alternatively we probably could have read in the GPX format directly
x <- readVECT6('folsom_trip1_pts')

# convert date/time field
x@data$date_time <- as.POSIXct(strptime(x@data$time, format="%Y/%m/%d %H:%M:%S+00"))

# distance intervals (meters)
dx <- diff(x@coords[,1])
dy <- diff(x@coords[,2])
x.dist <- sqrt(dx^2 + dy^2)

# time interval (seconds)
x.dt <- diff(x@data$date_time)

# compute velocity and convert to mph
x.meters_per_second <- x.dist / as.numeric(x.dt)
x.mph <- x.meters_per_second * 2.2369

# compute bearing
x.bearing <- atan2(dy, dx) * 180/pi

# add back to original DF
# note that first point has no velocity or bearing
x@data$mph <- NA
x@data$bearing <- NA
x@data$mph[-1] <- x.mph
x@data$bearing[-1] <- x.bearing

# check visually
spplot(x, 'mph', col.regions=brewer.pal('Spectral', n=5), cuts=5)
spplot(x, 'bearing', col.regions=brewer.pal('Spectral', n=5), cuts=5)

# keep only interesting information in the attribute table
# note that this destroys the original object's table
x@data <- x@data[, c('cat','ele','time','date_time','mph','bearing')]

# save data back to GRASS
writeVECT6(x, 'folsom_trip1_pts_dt')

# fancy plots
png(file='speed_vs_bearing.png', width=1000, height=500)
par(mfcol=c(1,2))
plot(mph ~ bearing, data=x@data, xlim=c(-180,180), xlab="Angle from East (positive angles are CCW)",
ylab="Velocity (mph)", main="Velocity Field", col=color.scale(mph,c(0,1,1),c(1,1,0),0), pch=16, cex=1.5)

## polar plot:
polar.plot(x@data$mph[-1], x@data$bearing[-1] ,main="Sailing Direction from East (0 Deg)", line.col='blue', lwd=2)
dev.off()