9

I have an ASCII file with latitude, longitude, and data_val in the following format.

35-13.643782N, 080-57.190157W, 118.6
...

I have a GeoTiff image file, and I can easily view it.

I want to place a "pin" (can be a dot/flag/star or whatever is easiest) on the image at the specific latitude/longitude position found in the ASCII file.

Here is what I've managed to do so far:

My source image looks like this:

Driver: GTiff/GeoTIFF
Files: /tmp/Charlotte SEC 100.tif
Size is 16867, 12358
Coordinate System is:
PROJCS["Lambert Conformal Conic",
    GEOGCS["NAD83",
        DATUM["North_American_Datum_1983",
            SPHEROID["GRS 1980",6378137,298.2572221010042,
                AUTHORITY["EPSG","7019"]],
            AUTHORITY["EPSG","6269"]],
        PRIMEM["Greenwich",0],
        UNIT["degree",0.0174532925199433],
        AUTHORITY["EPSG","4269"]],
    PROJECTION["Lambert_Conformal_Conic_2SP"],
    PARAMETER["standard_parallel_1",38.66666666666666],
    PARAMETER["standard_parallel_2",33.33333333333334],
    PARAMETER["latitude_of_origin",34.11666666666667],
    PARAMETER["central_meridian",-78.75],
    PARAMETER["false_easting",0],
    PARAMETER["false_northing",0],
    UNIT["metre",1,
        AUTHORITY["EPSG","9001"]]]
Origin = (-365041.822331817995291,240536.419747152860509)
Pixel Size = (42.334586069440391,-42.334898968590878)
Metadata:
  AREA_OR_POINT=Area
  TIFFTAG_DATETIME=2016:06:24 12:46:45
  TIFFTAG_RESOLUTIONUNIT=2 (pixels/inch)
  TIFFTAG_SOFTWARE=Adobe Photoshop CS5 Windows
  TIFFTAG_XRESOLUTION=300
  TIFFTAG_YRESOLUTION=300
Image Structure Metadata:
  COMPRESSION=LZW
  INTERLEAVE=BAND
Corner Coordinates:
Upper Left  ( -365041.822,  240536.420) ( 82d48'55.43"W, 36d13' 4.92"N)
Lower Left  ( -365041.822, -282638.262) ( 82d35'10.11"W, 31d30'17.00"N)
Upper Right (  349015.641,  240536.420) ( 74d51'46.40"W, 36d13'26.16"N)
Lower Right (  349015.641, -282638.262) ( 75d 4'55.60"W, 31d30'36.99"N)
Center      (   -8013.091,  -21050.921) ( 78d50'12.11"W, 33d55'36.35"N)
Band 1 Block=16867x1 Type=Byte, ColorInterp=Palette
  Color Table (RGB with 256 entries)
    0: 255,255,255,255
...

Here is what I've managed to cobble together in Python:

from osgeo import gdal, osr

src_filename = '/tmp/Charlotte SEC 100.tif'
dst_filename = '/tmp/foo.tiff'

# Opens source dataset
src_ds = gdal.Open(src_filename)
format = "GTiff"
driver = gdal.GetDriverByName(format)

# Open destination dataset
dst_ds = driver.CreateCopy(dst_filename, src_ds, 0)

# Specify raster location through geotransform array
# (upperleftx, scalex, skewx, upperlefty, skewy, scaley)
# Scale = size of one pixel in units of raster projection
# this example below assumes 100x100
gt = [-365041.822, 100, 0, 240536.420, 0, -100]

# Set location
dst_ds.SetGeoTransform(gt)

# Get raster projection
epsg = 4269            # http://spatialreference.org/ref/sr-org/lambert_conformal_conic_2sp/
srs = osr.SpatialReference()
srs.ImportFromEPSG(epsg)
dest_wkt = srs.ExportToWkt()

# Set projection
dst_ds.SetProjection(dest_wkt)

# Close files
dst_ds = None
src_ds = None

But, I can't quite figure out how to place a "red dot" at 35-13.643782N, 080-57.190157W

I'm having to learn some new details here (nomenclature about GIS).

  • The topic you may need to investigate is Georeferencing. – PolyGeo Nov 20 '16 at 3:59
  • Thanks.. I did some Google searching using the term Georeferencing. That was helpful. Half the battle is knowing what technical terms to use.. – Brad Walker Nov 20 '16 at 4:09
  • I'm sure I'm missing something, but have you considered converting the data to KML or something? – barrycarter Nov 20 '16 at 17:59
  • 1
    You may need to convert your DD-MM.mmmmH coordinates into decimal degrees. You'll need to parse the Hemisphere information W or S means a negative value (do that as the last step). The minutes need to be divided by 60 and added or concatenated with the degrees portion. – mkennedy Nov 21 '16 at 21:02
7

Your gdalinfo output shows you have a single band GeoTIFF with a colour table (AKA palette). I can't see the values in that colour table so the commands below convert the single band + colour table to a three band RGB GeoTIFF. The commands also assume your ASCII file has a header row and has coordinates in decimal degrees, you may need to modify your file if it doesn't.

Inputs:

$ cat testlonlat.csv
LON,LAT
143.798425,-15.551485
143.827437,-15.535119
143.84561,-15.530017
143.859107,-15.54819
143.812347,-15.523641
143.853581,-15.534694
143.883337,-15.537669
143.885356,-15.561687
143.887694,-15.588468

$ gdalinfo testutm.tif
Driver: GTiff/GeoTIFF
Files: testutm.tif
Size is 1102, 959
Coordinate System is:
PROJCS["WGS 84 / UTM zone 54S",
    GEOGCS["WGS 84",
        DATUM["WGS_1984",
            SPHEROID["WGS 84",6378137,298.257223563,
                AUTHORITY["EPSG","7030"]],
            AUTHORITY["EPSG","6326"]],
        PRIMEM["Greenwich",0,
            AUTHORITY["EPSG","8901"]],
        UNIT["degree",0.0174532925199433,
            AUTHORITY["EPSG","9122"]],
        AUTHORITY["EPSG","4326"]],
    PROJECTION["Transverse_Mercator"],
    PARAMETER["latitude_of_origin",0],
    PARAMETER["central_meridian",141],
    PARAMETER["scale_factor",0.9996],
    PARAMETER["false_easting",500000],
    PARAMETER["false_northing",10000000],
    UNIT["metre",1,
        AUTHORITY["EPSG","9001"]],
    AXIS["Easting",EAST],
    AXIS["Northing",NORTH],
    AUTHORITY["EPSG","32754"]]
Origin = (798741.168775000027381,8282084.855279999785125)
Pixel Size = (10.000000000000000,-10.000000000000000)
Metadata:
  AREA_OR_POINT=Area
Image Structure Metadata:
  INTERLEAVE=BAND
Corner Coordinates:
Upper Left  (  798741.169, 8282084.855) (143d47' 4.96"E, 15d31'16.22"S)
Lower Left  (  798741.169, 8272494.855) (143d47' 9.15"E, 15d36'27.98"S)
Upper Right (  809761.169, 8282084.855) (143d53'14.43"E, 15d31'11.47"S)
Lower Right (  809761.169, 8272494.855) (143d53'18.78"E, 15d36'23.20"S)
Center      (  804251.169, 8277289.855) (143d50'11.83"E, 15d33'49.74"S)
Band 1 Block=1102x7 Type=Byte, ColorInterp=Palette
  Color Table (RGB with 256 entries)
    0: 120,112,136,255
    1: 96,104,88,255
    ...
    254: 76,124,140,255
    255: 232,228,236,255

Process:

$ gdal_translate -expand rgb testutm.tif testutm_rgb.tif

$ ogr2ogr -f "GeoJSON" -dialect sqlite                      \
  -sql "select ST_buffer(Geometry,0.001) from testlonlat"   \
  -s_srs EPSG:4326 -t_srs EPSG:32754                        \
  /vsistdout/ CSV:testlonlat.csv -oo X_POSSIBLE_NAMES=Lon   \
  -oo Y_POSSIBLE_NAMES=Lat |  gdal_rasterize -b 1 -b 2 -b 3 \
  -burn 255 -burn 0 -burn 0 /vsistdin/ testutm_rgb.tif

The last command does the following:

  • buffers the Lon/Lat point to a bigger polygon so it shows up better (you can skip that if you just want a single pixel coloured red)
  • converts from WGS84 Lat/Lon (EPSG:4326) to the same coordinate system as the raster (EPSG:32754 which is WGS 84 UTM Zone 54S, your CRS will be different)
  • writes the output polygon as GeoJSON to STDOUT and pipes it to gdal_rasterize
  • burns RGB 255,0,0 into the RGB raster bands 1, 2 & 3

Result:

enter image description here

3

You have a good start. gdal.CreateCopy will take care of the georeferencing, so you don't have to set that a second time using the geotransform and the projection.

The complete process will transform the lon/lat coords into the XY coordinates of the raster spatial reference. Then these XY coords will be transformed into the row,col indices of the raster using the inverse geotransform. Some pixel value will be written at that position.

from osgeo import gdal, osr
import numpy as np

src_filename = '/tmp/Charlotte SEC 100.tif'
dst_filename = '/tmp/foo.tiff'

# Opens source dataset
src_ds = gdal.Open(src_filename)
format = "GTiff"
driver = gdal.GetDriverByName(format)

# Open destination dataset
dst_ds = driver.CreateCopy(dst_filename, src_ds, 0)

# Get raster projection
epsg = 4269         # http://spatialreference.org/ref/sr-org/lambert_conformal_conic_2sp/
srs = osr.SpatialReference()
srs.ImportFromEPSG(epsg)

# Make WGS84 lon lat coordinate system
world_sr = osr.SpatialReference()
world_sr.SetWellKnownGeogCS('WGS84')

# Transform lon lats into XY
lonlat = [[0.,30.], [20., 30.], [25., 30.]]
coord_transform = osr.CoordinateTransformation(world_sr, srs)
newpoints = coord_transform.TransformPoints(lonlat) # list of XYZ tuples

# Make Inverse Geotransform  (try:except due to gdal version differences)
try:
    success, inverse_gt = gdal.InvGeoTransform(gt)
except:
    inverse_gt = gdal.InvGeoTransform(gt)

# [Note 1] Set pixel values
marker_array_r = np.array([[255]], dtype=np.uint8)
marker_array_g = np.array([[0]], dtype=np.uint8)
marker_array_b = np.array([[0]], dtype=np.uint8)
for x,y,z in newpoints:
    pix_x = int(inverse_gt[0] + inverse_gt[1] * x + inverse_gt[2] * y)
    pix_y = int(inverse_gt[3] + inverse_gt[4] * x + inverse_gt[5] * y)
    dst_ds.GetRasterBand(1).WriteArray(marker_array_r, pix_x, pix_y)
    dst_ds.GetRasterBand(2).WriteArray(marker_array_g, pix_x, pix_y)
    dst_ds.GetRasterBand(3).WriteArray(marker_array_b, pix_x, pix_y)

# Close files
dst_ds = None
src_ds = None

Note 1:

The command gdal.RasterBand.WriteArray(array, xoff, yoff) operates from the top left corner. In this example I provide a 1x1 array with the value 255, so xoff and yoff are the actual row,col indices for the lon/lat position. If you want to write a 3x3 square, you need to adjust xoff and yoff by subtracting 1. You should also make sure the array datatype matches the raster's. Because you said you want a "red dot", I am assuming there are three bands of uint8.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.