A laser-based research system for gated viewing is currently being developed at FOI One of its possible applications is long-range automatic target recognition. This report treats an implementation of a viable target recognition algorithm based on one-, two- and three-dimensional (1D, 2D and 3D) comparisons of target features. The work starts with a strategy for the construction of three-dimensional objects from gated image sequences, and covers the synthesis of a candidate library. The next step is a course pre-screening test based on one-dimensional projections, followed by a more thorough test based on surface and boundary comparisons. The algorithm is tested on real gated image sequences, and robustness tests are also performed with synthesized objects. An approach to automatic target recognition from airborne scanning 3D laser radar data is under development, and a brief summary of the main steps of the method is given. Finally, a method of generating adaptive 3D mesh surface representations from 3D laser radar data is implemented and tested. The implemented method is found to produce acceptable results, but to suffer from computational efficiency problems. However, a constituent surface-matching function is found to have an application to target recognition based on airborne scanning 3D laser radar data.