Hi I've just got a small problem that needs to be cleared here to allow me to do the other questions. Statement: Two objects were detected in a 100-sec exposure with 3x4 CCD camera Now you are given 3 by 4 matrices for the raw image data, bias, dark count and flat field. I've provided a link to a screen shot of the excel spreadsheet of the data so you can actually see it clearly. LINK: http://img368.imageshack.us/img368/4324/ccddata7ev.jpg Anyway I found an expression in the lecturer's notes that says true count is given by [tex][(raw image_m/NLG_m)-bias_m-dark_m]/(FF_m*G)[/tex] Where the subscript m means matrix. Dark is the dark count matrix and FF is the flat field matrix and NLG and G are the non-linear gain correction and gain correction respectively. NLG and G aren't given though. Anyway the thing that confuses me is that you can't do that operation with matrices. Unless the division of the matrix is referring to the inverse. However, since the matrices aren't square they aren't invertible. I know you do have the right inverse and left inverse of a non-square matrix, so am I meant to calculate that and apply it to the above expression?? Or have I misinterpreted the equation here? Cheers for any help P.S. This is just an exercise, and it does violate Rayleigh's criterion.