Here's what I didn't get around to doing, but should work:

1) Setup a DSLR in a fixed position pointing at a wall

2) Setup a grating in a fixed position near the DSLR

3) Shine a known wavelength through the center of the grating at a point on the wall that you label with a dot.

4) Make sure this point on the wall, as well as the second point generated by the grating, are within frame of the DSLR.

5) Shoot a photo.

6) Setup another known wavelength, point it at the same dot marked on the wall, and take another photo (do this for 3 or 4 total different known wavelengths, ie 473, 532, 589, 593.5)

7) Run through the same setup for your UNKNOWN wavelength.

8) In software, count the pixels between the two dots for each photograph, and plot them as X and Y values in Excel, like this:

PIXELS WAVELENGTH

313 473

731 532

1221 589

1271 593.5

8) Graph those as a line graph, and add a "Trendline" with Polynomial Order "3".

9) Display the equation on your chart, and then copy it down. You'll get something like "y = 2E-09x3 - 3E-05x2 + 0.1715x + 422.44"

10 Take the pixel count of your UNKNOWN wavelength as "x" and sub it into this equation.

11) Solve the equation, and "y" will be the wavelength of your unknown.

My feeling is that with a 3rd order polynomial, you're completely compensating for the tilt angle of the wall, the perspective skew of your camera's lensing, and the non-perfect angle of the actual diffraction grating projection. I probably wouldn't do it with 3 wavelengths, and in fact I'd probably try to avoid using both 589 and 593.5 as separate ones. If you had a HeNe, that would be a more appropriate 4th wavelength.