Introduction. Automated ballistic identification systems allow automation of inspections by array firing pin traces containing thousands of similar objects. However, sometimes the system allow «mistakes», ie can not find ‘doubles’ trail in database. In addition, quite often «doubles» trail of test array is placed on the priority list is far from its beginning. This is due, primarily, large morphological variety and high variability of individual attributes of weapons displayed in the firing pin tracks, as well as uneven lighting traces due to their complex forms. Theoretical analysis. Studies have shown that non-uniformity of brightness of digital images traces of the firing pin can be smoothed by applying the homomorphic image processing method. Analysis of the morphology of individual signs weapons displayed in the tracks of the strikers 30 models of weapons, allowed to identify six main morphological types of individual attributes. Experimental investigation. Efficient algorithms for feature extraction of the form of large irregularly shaped spots by applying the Wiener filter and method Niblek developed. For feature extraction in the form of circles, we propose a method based on the use of the Canny filter. Conclutions. Studies have shown: homomorphic method of digital image processing can be recommended for pre-processing raw images; classification of morphological types of individual signs developed; algorithms binarization images features in the form of large irregularly shaped areas and in the form of circles developed.