Great to see so many replies.
I indeed thought about using a lens and did some research on it.
Germanium and polyethylene seemed very good materials, block visible and have a high transmission in the infrared, but Ge is very expensive.
For focusing the heat radiation into the bolometer I need 2 plano convex lenses, which would make it rather expensive.
And either way, the amount of radiation I would get from a 1 mm^2 spot would be very small and as far as I see it, it would be the same as placing an 1 mm^2 aperture in front of the bolometer, and putting it very close to the area of interest. Isn´t this true? Also, using lenses I expect that the bolometer would have to be further away from the sensing area, which would make it more prone to temperature fluctuations...
You can check more info about the sensor I have here: http://www.melexis.com/Infrared-Thermometer-Sensors/Infrared-Thermometer-Sensors/MLX90614-615.aspx