mapmixer & rectifier

Yahoo! has announced Map Mixer, something that at first look, feels a lot like Metacarta’s Map Rectifier that I’ve blogged about before – so I thought I would give it a go, and compare the two.

This time, I wanted to put on an aerial photograph of the area where I live, and thanks to it being on the flightpath for Leeds Bradford Airport, and Flickr, I found a suitable photo, by Imran no less! He had already done some photo stitching on this photo too. But you can see the main roads and prominent buildings. Its also not too oblique, so should be able to “warp” well. I would expect the edges to suffer the most, however.

Next comes uploading to the two sites, and registering the image to the map. Here is where the main difference is, with map Rectifier using multiple Ground Control Points (GCPs) which gives a much better fit than just two pairs of GCPs (Point Matching) that Map Mixer works with.

With MapMixer, After setting the two GCPs for “point matching”, it rotates the image pretty well. You are then given a change to rotate, and linearly stretch the image to match. Then after giving a title and tags it’s saved, however, you are unable to go back and edit it, and unable to download the image for use elsewhere. But you can easily embed this new map.

Heres the url for the map i created using Map Mixer: http://maps.yahoo.com/mapmixer?lid=f9618ad7&pg=view

With Metacarta’s Map Rectifier, I was able to add in 12 GCP’s, and it warps it for you, good results were had with quadratic fit. If you select “Linear fit” the map looks pretty much identical as the map mixer result.

You can then download the warped image, or use the url in a WMS query to put on your own website, and map of choice. You can change GCP’s and add in new ones, and experiment with different ways of warping to give a satisfactory result.

the more warped cubic fit. http://labs.metacarta.com/rectifier/rectify/289

MapMixer works well, and is intuitive, but for maps and images with slightly off projections or warping, it should incorporate more point matching, but of course there would be a trade off. It’s nice to see them taking a step forward with enabling user contributed geodata to be published and consumed online.