New Shapemodel!

Here is my latest shapemodel:

It is as good as I believe can be done with the current limited data-set. It is built by correlating 150 surface features over 120 navcam and OSIRIS frames. Each image has had its viewing geometry reconstructed using sparse bundle adjustment. Then a select number of images where used to derive dense depth from stereo data. To combat the noise in the stereo data I used shape from shading to create high resolution local data.

For the parts of the comet currently in darkness I have used images of the limb against the slightly brighter dust cloud to constrain the volume that the surface is inside.

I have smoothed out the “unknown” areas to make the model a bit prettier.

There are areas especially around the neck with artifacts and seams. I have not fixed them because the model is continuously made obsolete by the steady stream of new images being released.

As soon as new data becomes available I try to update the model. And I will post a new one here if there are meaningful changes to the model.

The model is Copyright Mattias Malmer CC BY SA 3.0

I would like to be credited for the creation of the model and a link to this blog would be appreciated.

Image source credits:





CIVA Depth Cues

Added some depth cues to the CIVA panorama. It really helps to create a sense of depth and scale.

I segmented the image boulder for boulder looking for overlapping sillhouettes. That created the initial sorting.

Then I used the shadows and the scale of the surface grain to estimate distance.

I coloured the surface redish to match with some information on colour that I have seen drifting around on the web..

I added the contrast loss that an atmosphere/fog would give based on the distance estimates.

CIVA Sun direction

I spent some time reconstructing the vector to the sun in the Philae CIVA panorama images.

I mostly used the CIVA camera 4 image. It has a lot of objects casting shadows. onto other surfaces

If one can see an object and its corresponding shadow in an image one can construct a plane using the cameras position and the two vectors pointing from the camera to the shadow and its caster. The sun is located somewhere on that plane.

If one has more than one object with an corresponding shadow one can create multiple planes. And the line that is created in the intersection of those planes will point at the sun.

This approah worked very well for that image.

My normalized vector pointing at the sun in the Philae spacecraft coordinate system is [0.0449037,0.68573,0.726469]

Those numbers obviously do not say much so I created this QuicktimeVR:


This image is looking down at philae from the sun. The sun is in the intersection of the three coloured lines that are the intersecting planes seen edge on.

Looking from the sun

This image shows which shadows I used. (I found the shadows in the vastly better resolved mosaic released by ESA. I then transferred them to this square image.) The green cross gizmos are the shadow/caster pairs and the red green and blue lines are actually the planes constructed seen edge on.

shadow Planes

Navcam November 17

Navcam images registered on my shapemodel. I am very happy with how well the model matches. There are very few alignment errors.



3DTV (side by side)

Image credits: ESA/Rosetta/NAVCAM

Image processing/shapemodel: Mattias Malmer




The Mountains of Agilkia

I just made this digital flyby of the mountains of Agilkia. OSIRIS frame draped ofer my latest digital terrainmodel of the landing area.

Sleep well little Philae. What a grand view you have!

Image credits:


The slopes and roughness of the Philae landingsite

I wanted to get a feel for the landscape in which Philae is set to land. I had already built a resonably accurate terrainmodel of the area. So I decided to play a little with the data.

The result are these two maps.

First a map showing the slope of the terrain. There are a few really steep places in the landing area but most of it is thankfully rather flat.

Then a slightly more worrying map showing the surface roughness. There are a few boulder fields that scare. but much of the area seem to be smooth (at least on the scales resolved in my dataset.)

Obviously my methods are not as refined as the ones employed by the mission teams but these maps should give an idea of the challenges facing the little lander.


I also made this Anaglyph: