A little mystery.So I says, hey, I can...

A little mystery.

So I says, hey, I can check if the SSI cameras were pointing in the same direction for the sol 0 and sol 38 lander leg images.

Checking the metadata in the JPF files for the Left exposures:

sol 000, 17:07:41 : -63.4431, 39.9232
sol 038 17:39:16 : -64.4138, 38.8618

Well OK, now I can check the expected pixel difference between features in the images:

elevation difference: -0.9707 = -71 pixels
azimuth difference : -1.0614 = -78 pixels

distance difference: 105 pixels

So I load the L1 sol 0 and 38 inages into Stereophoto Maker and expect to see registration differences of about -78 pixels in X and -71 pixels in Y ( assuming Phoenix is sitting level and hasn't moved )

I start measuring and am more and more puzzled:

center : -38,-70 ( 80 pixels )
upper left : -45,-80 ( 92 pixels )
upper right: -48,-64 ( 80 pixels )
lower left : -30,-83 ( 88 pixels )
lower right: -29,-65 ( 72 pixels )

OK, I know neither of the SSI cameras pivots through their optical axis, so some of this may be due to that effect.

Lets see, 7.5 cm from the pivot point and 1 degree pointing difference -- mumble, mumble -- about 1/2 pixel difference can be due to the off axis rotation for a target at 1 meter -- so that's not it....

The elevation differences are within (-1 to +12 pixels of the expected differences -- but the azimuth differences are between (-49 to -30 )

Er, the lander rotated 1/2 degree???

Naaaah.

They must have changed the lander coordinate system a few days after landing...

But the elevation pixels differences...

Another coordinate system change?

Maybe. Maybe not.

Making sense of remote data: Damn queer business.