Recognition of Ship Types from an Infrared Image Using Moment Invariants and Neural Networks

This is a NAVAL POSTGRADUATE SCHOOL MONTEREY CA report procured by the Pentagon and made available for public release. It has been reproduced in the best form available to the Pentagon. It is not spiral-bound, but rather assembled with Velobinding in a soft, white linen cover. The Storming Media report number is A248983. The abstract provided by the Pentagon follows: Autonomous object recognition is an active area of interest for military and commercial applications: Given an input image from an infrared or range sensor, find interesting objects in those images and then classify those objects. In this work, automatic target recognition of ship types in an infrared image is explored. The first phase segments the original infrared image in order to obtain the ship silhouette. The second phase calculates moment functions of those silhouettes that guarantee invariance with respect to translation, rotation and scale. The third phase applies those invariant features to a backpropagation neural network and classifies the ship as one of five types. The algorithm was implemented and experimentally validated using both simulated three-dimensional ship model images and real images derived from video of an AN/ AAS-44V Forward Looking Infrared (FLIR) sensor.