DISCRIMINATION OF MUNITION FILL TYPES BY K-NEAREST NEIGHBOR CLASSIFICATION TREE ANALYSIS OF ACOUSTIC-RESONANCE SIGNATURES

We have developed a noncontacting, long-standoff inspection system for munition fill type identification. The inspection system, based on us ing a laser vibrometer to measure the response of containers to acoust ic excitation, is capable of detecting subtle changes in vibration cha racteristics due to variations in the physical properties of different fill materials. Determination of fill type is achieved by applying a k-nearest neighbour classification tree analysis to the resulting acou stic resonance data. Application of the method to field trial measurem ents of chemical weapons shows the inspection system to be a practical and reliable means for discriminating between munitions with a variet y of chemical fills.