Atomic line filters take advantage of the sharp spectral features offered by electronic transitions in free atoms. The basic idea is to convert a narrow atomic absorption profile into an equally narrow optical transmission filter by detecting the fluorescence radiation that is emitted after a photon has been absorbed. We developed the metastable thallium atomic line filter from the basic physical idea to a laboratory prototype in order to evaluate the usefulness of the technology for satellite borne backscatter LIDARs. Atomic thallium provides for an unsurpassed combination of features for this purpose. The input wavelength of the filter is 535 nm, which matches the frequency doubled Nd:BEL laser. The filter is an active device, it upconverts the 535 nm input into 378 nm output that is detected by a PMT. The experimental results were as expected concerning the characteristics directly related to atomic properties, i.e. the filter has only 10 ns response time, its optical bandwidth is 0.004 nm, and its acceptance angle is only limited by the device geometry. An optimized setup, with the size constraints of a space-borne system, displays a total quantum efficiency of 2%, i.e. from input photons to detected output photons. This constitutes a remarkable and unsurpassed value for atomic line filters but is too low for the application in mind. (A ground-based, scaled-up version of our prototype would reach about 10% quantum efficiency). In addition, the 500 degree(s)C operating temperature of the vapor cell requires sophisticated thermal layout and thermal shielding, which means a lot of added mass and volume. In summary, we found that atomic line filters can be built that offer characteristics not found with other technologies, but their applicability for space borne systems is questionable.