Features Extraction and Classification of Cartridge Images for Ballistics Identification

Ballistics identification is a quite challenging task from both theoretical and practical point of view. The premise underlying firearms and toolmark identification is that each firearm owns its unique toolmarks, resulting in some unique characteristic markings left on the fired projectile and cartridge case. Although various techniques have been applied to firearm identification, the major problems are still unsolved and further improvement and development are required. In other words, the ability of extracting useful features and the intelligently identification is still a major concern. This paper addresses the difficulties with respect to feature extraction and intelligent ballistics recognition. In this paper, various image processing techniques are employed for image digitizing and preprocessing the ballistics images. A novel feature set known as Circle Moment Invariants has been proposed for extracting features in cartridge images. We utilize two feature sets to characterize the ballistics images. A neural network based intelligent system is designed for classifying the extracted features of ballistics images. The experimental results indicate that the proposed approaches and feature criteria are capable of classifying the cartridge images very efficiently and effectively.