三维枪弹痕点云数据处理及特征提取研究

为了解决传统算法中弹痕图像深度特征信息丢失的问题,本文提出一种能够计算出弹痕图像深度特征参数信息的三维点云图像特征提取算法。该方法根据棱线斜率变化从局部极值点中找到特征点,运用线性拟合法构建三角形对特征参数值进行计算,确定弹痕特征参数值的相似范围,通过利用相似范围对未知弹头进行判别,实现弹痕和枪的一致性确认。实验结果表明：所提出的算法在现有样本的条件下,弹痕比对的正确识别率达到90%以上,单组弹痕数据的转换、特征提取和参数计算共用时32.7 s。算法满足弹痕比对需求,可以对后续弹痕比对提供可信依据,具有一定的理论价值和实用意义。

Processing and feature extraction for three-dimensional bullet point cloud data

In order to solve the problem of bullet marks image depth feature information loss in traditional algorithm, we propose a three-dimensional bullet point cloud image feature extraction algorithm. The feature points are extracted from the local extremal points according to the change of ridge's slope. And the linear fitting method is utilized to construct the triangle in order to calculate the characteristic parameters and define the similar range of the characteristic parameters. Finally, the unknown bullet is discriminated through the similar range to ensure the consistency between the bullet and the gun. The experimental results show that correct recognition rate of bullet comparison reaches more than 90% under the condition of existing sample, and data processing of a single set of bullet data spends 32.7 s. The proposed algorithm satisfies the bullet comparison demand, provides credible basis for the following bullet comparison and has certain theory value and practical significance.