I was thinking of using a linear CCD, from a scanner since those are abundant and cheap.
I don't know if the linear CCD would work though, since a lot of times the beams are asymmetric and with linear CCD you will only get the cross section of the beam. Unless youscan the across the beam. The filter issue is fairly easy to resolve since you can have a combination of filters so for higher power laser you can just add on more filter. Or one of those adjustable one might work, but be aware that there will be difference in OD across the beam because the filter density change is continuous.