Comparing matrices with different length.

I want to compare several matrices from kinematic gps (moving antenna) observations to see the differences from several processing strategies. The matrices look like this

 X              Y           h    m d h m  s
7475331,566  604211,547 53,728   5 7 2 38 49

with one day of observations (one for every second). The problem is that the matrices miss some of the lines because of the observations always ain't that good. So somewhere in the file it might be a leap from ex 5 7 2 38 49 to 5 7 2 38 55, missing 5 seconds.

I need a way to sort the matrices so i can make the subtractions at the right time.

Example:

From strategy 1,

 X              Y        h      m d h m  s
7475331,566  604211,547 53,728   5 7 2 38 49
7475341,575  604223,523 53,543   5 7 2 38 50
7475351,453  604254,556 53,765   5 7 2 38 51
7475361,345  604277,523 53,897   5 7 2 38 52

From strategy 2,

 X               Y        h      m d h m  s
7475332,523  604210,565 53,768   5 7 2 38 49
7475343,556  604224,674 53,578   5 7 2 38 50
7475352,230  604253,345 53,345   5 7 2 38 55  <---- gap of 4 seconds
7475364,456  604276,123 53,789   5 7 2 38 56

I was thinking i should try to fill up the rows with the missing data with only zeroes. But i can't find a way to do it. These files are also to big to do it manually 24 hours with one row for every second.

Anyone got any ideas on how i could solve this?

Answer 1

[Bjorn Gustavsson]

You could fill it in with interpolation of some sort, see: interp1. Or you could use inpainting, see google - if you dont find something useful on the matlab file exchange. You could also compare only the observations from times that exist in both sets, see: intersect.

HTH.