How to extract X Y coordinates and RGB from pixels using python in QGIS?

Question

I'm trying to white a Python plugin in QGIS (2.2) to get, for each pixel of a raster image, its X Y coordinates, as well as, its RGB (3 bands) values.

At first I opened the raster file using: rlayer = QgsRasterLayer(rfileName, rbaseName)

Now I don't know how to get, for example, for pixel (1,1) its coordinates (X,Y) and its RGB color values.

Anybody could help me?

I know that I'll need to implement a DO WHILE, but I don't know the commands to extract this information of each pixel.

Thanks in advance, Mateus

Answer 1

As a geologist, I make geological cross section using the elevations values from a DEM and the colors from a geological map, look at the pure Python solution with osgeo.gdal in Python Script for getting elevation difference between two points

But now, since PyQGIS 2.x, it is easier with PyQGIS and the QgsRaster.IdentifyFormatValuefunction

• example with a DEM (one band):

  myDEM = qgis.utils.iface.activeLayer()  
  print myDEM.bandCount()  
  1   
  print myDEM.dataProvider().identify(QgsPoint(229774,111171), QgsRaster.IdentifyFormatValue)  
  {1: 221.0}
  

• example with a classic raster (three bands -> R,G,B values):

   print myraster.dataProvider().identify(QgsPoint(229774,111171), QgsRaster.IdentifyFormatValue).results()
   {1: 180.0, 2: 142.0, 3: 125.0}
  

The result is a dictionary (key= band number) and you can create a simple function that returns the values:

def val_raster(point,raster):
    return raster.dataProvider().identify(point, QgsRaster.IdentifyFormatValue).results().values()

elevation = val_raster(QgsPoint(229774,111171),myDEM)
print elevation
221.0
R,G,B = val_raster(QgsPoint(229774,111171),myraster)
print R,G,B
180.0, 142.0, 125.0

I presented a complete solution with PyQGIS in French PyQGIS (QGIS 2): geological cross-sections (colorization of a topographic profile build from a DEM with the colors of a raster and placement of the intersection points of the geological layers boundaries )

Answer 2

I don't have enough points to comment so I am posting in answer section.

I came from MATLAB background and always like to have map2pix and pix2map function as giving in mapping toolbox. Using above comments and posted pdf link, I wrote both of them. It might help someone.

After this layer.dataProvider().identify(QgsPoint(x, y), QgsRaster.IdentifyFormatValue) can be used to get appropriate values

def map2Pix(longitude =0,
            latitude =0,
            transform =None):
    """
    converts map coordinates to image pixel [col, row]
    :param longitude:
    :param latitude:
    :param transform:
    :return: xOffset, yOffset
    """
    if transform:
        # compute pixel offset
        xOffset = int((longitude - transform[0]) / transform[1])
        yOffset = int((latitude - transform[3]) / transform[5])
        return xOffset, yOffset

def pix2Map(xOffset =0,
            yOffset =0,
            transform =None):
    """
    converts image pixel [row, col] to map coordinates
    :param xOffset:
    :param yOffset:
    :param transform:
    :return: longitude, latitude
    """
    if transform:
        longitude = transform[0] + transform[1]*xOffset
        latitude = transform[3] + transform[5]*yOffset
        return longitude, latitude

Answer 3

The support for accessing raster data via QgsRasterLayer is fairly limited:

http://lists.osgeo.org/pipermail/qgis-developer/2008-September/004790.html

You can access the value of an individual pixel using the identify tool:

x, y = -1.535, 54.634
res = layer.dataProvider().identify(QgsPoint(x, y), QgsRaster.IdentifyFormatValue)
print res.results()

You can access the width, height, extent and coordinate system too:

width = layer.width()
height = layer.height()
extent = layer.extent()
crs = layer.crs()

However, you're probably better off accessing the raster directly using GDAL, for which there are many examples. See this PDF as a general introduction to reading raster data with Python/GDAL: http://www.gis.usu.edu/~chrisg/python/2009/lectures/ospy_slides4.pdf