Tip:
Highlight text to annotate it
X
This tutorial will cover Image Rectification, Clipping and Reprojection with MapInfo Discover.
The Rectify Image tool allows imagery (such as aerial photography and satellite imagery, or scanned geological maps)
to be accurately georeferenced using ‘rubber- sheeting’.
This rectifies images that have been stretched, skewed and/or rotated
so that they will display in the correct geographical location, by warping portions of the image to fit the provided control points.
In contrast, Registering such problematic images with MapInfo Pro
will result in the distortion of both the map window projection, as well as any overlying vector data.
In order to rectify an image a number of accurate control points need to be added
for which the real world coordinates/locations are known.
Ideally control points should be placed over the entire image and the more control points the better the final rectification.
Two primary methods of control point specification are available;
either manually using known sets of coordinates
or via a spatially-coincident MapInfo Pro map layer with corresponding features.
We will cover both methods in this tutorial.
From MapInfo Discover’s IMAGES tab, select the Rectify Image tool.
Use the Load button to browse to and directly open the source image.
Select the ‘No’ button on the Loading Control Points message (otherwise it will create control points at the 4 corners of your image).
Let’s look at manual control point entry method first.
To start, use the ‘Assign Projection’ button at the bottom to set the projection of the known control points.
In this case, the geological map is in latitude/longitude AGD 66, so I’ll specify this.
Note that even though I actually work daily in the more recent GDA94 cartesian system,
I need to use the projection of the known data/control points;
later in this tutorial I will demonstrate how to Reproject from this source projection into a desired ‘working’ projection.
If working in latitude/longitude, the adjacent DMS tick box will allow control point entry
in the more useful Degrees:Minutes:Seconds format, rather than decimal degrees.
Pan and zoom to the first control point location at the top left of the data area,
noting in this example that the coordinates of the location are 31 degrees latitude and 138 degrees longitude.
Note that as this Australian map sheet is in the southern hemisphere, the latitude will need to be negative.
Press the Image button and note that the cursor has changed to a cross-hair.
With this, click on the desired control point location: a red cross symbol will be created,
and a new row added to the Rectify Image dialog.
In the Map_X and Map_Y fields for this new row enter the control point’s coordinates.
When complete, press the ‘New Point’ button to add the next control point row (it will be highlighted green when selected/active).
Pan and zoom to the next control point location, again select the Image button, and target the location.
Once captured, enter the coordinates.
Repeat this process for each additional control point, ensuring you FIRST press the New Point button each time,
otherwise you will simply overwrite the last entered control point.
PROTIP: I’ve configured my middle mouse wheel such that in addition to the default ‘Pan’ action,
the mouse wheel also ‘Zooms’ on its current location, which removes the need to activate any tools for these functions.
Set this via PRO>Options>Map Window and enable the ‘Mouse Wheel Zoom…..’ option).
Once you have finalized entry of control point information (you will need 3 points as a minimum,
but the more accurate control points the better), note the RMS error.
This should be less than 20, and ideally less than 5.
You can easily visually check each existing control point whilst zoomed-in, by selecting its row and pressing the ‘View GCP’ button.
With the row selected, pressing and applying the ‘Image’ button will update the Image_X and Image_Y coordinates for that control point.
When satisfied, press the Rectify button to start the Rectification process.
Select ‘Yes’ to first save the control points (which is very useful if they need to be updated or corrected later on),
leaving the default table name ( a ‘_gcp’ suffix for ‘ground control points’).
Then press ‘Save’ for the new rectified image, again leaving the default new table name and location.
A progress bar will display at the bottom of the dialog.
When complete, the resulting image will open into a new map window.
Note the coordinates and projection.
Obviously this geological mapsheet image contains numerous extraneous components,
such as legends, cross-sections, etc, which we’d like to remove.
I’ll use the 250K mapsheet boundaries for Australia to clip-out this non-relevant information.
You’ll note that the Parachilna mapsheet polygon coincides perfectly with the geology map area;
I’ll first select this polygon, and then open up the Clip Image tool from the IMAGES tab.
In the dialog, ensure that the correct source image table is selected,
as well as the correct clipping polygon (in this case the ‘Selection’, rather than the whole mapsheet table).
Leaving the other options as their defaults, press OK.
The final step in this process is to convert the resulting clipped image into the cartesian projection I work in daily: GDA94.
Opening the Reproject Image tool, again ensure that the new clipped image is set as the target, and set the New Projection as desired.
I’ll also set the Interpolation to Bicubic for the most accurate (though slowest) interpolation process.
Returning to the rectification process, let’s try the second method:
capturing control points from a coincident MapInfo Pro spatial dataset.
We are going to rectify this aerial photo on the left, but I’d like to populate my control points
from this coincident vector table on the right of the visible stream network, with these correlating stream features clearly visible.
Having opened my airphoto into the Rectify Image tool, I’ll first press the Image button to position my first control point.
Then, with the new row selected, I’ll make the ‘Streams’ mapper active, enable Snapping (via the keyboard 'S' key),
and press the ‘Map’ button in the dialog.
I’ll locate the stream junction correlating with this control point location in the image, and select.
It will prompt us to use the Streams map projection (because I have not already assigned one).
Select Yes, and the control point's Map_X and Map_Y coordinates are automatically filled from the spatial dataset.
Repeat this process for each clearly identifiable control point in the two datasets,
ensuring you first press the ‘New Point’ button each time, to prevent overwriting the previous control point.
Press Rectify when satisfied, and open the resulting rubber-sheeted image
into the same mapper as you source spatial dataset for validation.
That concludes this tutorial on Image Processing with MapInfo Discover.
Thankyou for watching.