The Research And Implementation Of The Key Algorithms Of The Integrated Bullet Identification Software System Based On 3-D Laser Color Scanner

In this thesis we discuss the research and implementation of the key algorithms of computer-aided bullet identification system which is based on the Three- Dimensional Laser Color Scanner ( 3DLCS ) developed by our research group.At first, we define and introduce some basic conceptions of 3DLCS Integrated Bullet Identification System (3DLCS-IBIS), to convenient readers comprehend the whole thesis.Then, concerning the image processing technologies for bullet trace detecting, we raise an unprecedented method integrated with 2D/3D model and database: in the case of warhead trace detecting, we propose a new Unwrapped-Spread Curve (USC) method to judge whether they are from the same certain gun, and give a mapping from the boundary curve of the cross section (BCCS) of the bullet to the USC. Based on the USC method, rather than directly compare the BCCS of different bullets in order to judge whether they are from the same certain gun, we invent a new detecting algorithm by using the comparability measurement of the USC instead. To ensure the practicability of the USC method, the conformal property of the mapping, which is closely related to the effectiveness of the method, is proved. During bullet marks measuring, this method can reduce the quantity of data and the difficulty of measuring. With the automatic detecting algorithm of computer, the workload of human is considerably alleviated. Besides, based on human vision mechanism we raise the Statistical Algorithm Based on Grain Distribution (SABGD), and give the human vision principle and algorithm flow, strongly show the superiority of SABGD comparing to Minimum Absolute Distance (MAD) algorithm by using digital image processing methods; In the case of cartridge case trace detecting, we propose the Color Contour Plot (CCP) algorithm, research the meaning and give the slow of CCP algorithm.In the end of the thesis, we give the method of bullet information database design, including database design and search design. Then raise a way of bullet trace detecting using this database subsystem.From the experiment results, we prove that these new algorithms can preferably suite the practical require of bullet trace detecting, considerably alleviate the workload of human trace detectors and enhance the efficiency of related police jobs.