A long range target recognition system has been developed using a high range resolution mode in a surveillance radar. The radar performs normal air surveillance but on targets of interest a wideband waveform is used in a few pulses of the normal dwell. These pulses are processed to yield high resolution range profiles of the target. Combining range profiles with tracking data makes it possible to classify uncooperative targets into the following perceptual classes: large jet, small jet, large propeller, small propeller, and helicopters. Target recognition algorithms were developed and tested with a database obtained by recording the signatures of a mix of identified civilian and military aircraft. The single look recognition accuracy is better than 80 per cent for all classes except helicopters, which was 68 per cent.