剪切痕迹激光检测信号自适应匹配算法研究

发布时间：2019-07-02 11:56

【摘要】：在刑事技术中,剪切工具痕迹是其研究的重点方向之一。工具痕迹是盗窃、抢劫、杀人等许多案件中最常出现的一种痕迹,对于认定案件性质,确定作案工具,证实犯罪嫌疑人具有重要意义。据不完全统计,70%以上的刑事案件现场有工具痕迹,有些地区的工具痕迹出现比例达80%左右。而且工具痕迹还具有不易破坏、难以伪装、出现率高,鉴定价值好的特点,这些优势都是其它类型痕迹难以比拟的。因而对工具痕迹的提取分析进行深入研究,具有重大的实际意义。本文在痕迹信号的采集上面,采用基于LabVIEW设计的工具痕迹激光检测装置。通过该装置可以有效的采集到工具痕迹上面的痕迹信号。在采集信号的过程中由于反光以及装置等不可控因素的存在,所以扫描采集得到的信号中会有异常数据以及噪声干扰的存在。在对异常数据处理上面,重点研究了运用K-Means算法对信号中的异常数据进行修复,并通过试验仿真来验证该算法对于剪切工具痕迹激光检测信号中的异常数据具有很好的修复效果。在对信号降噪处理上面,重点研究了通过LOWESS(局部加权回归散点平滑法)算法对数据进行平滑处理,通过该算法能够最大程度的消除扫描数据中的噪声。然后通过试验仿真来验证该算法的有效性。在相似度比对上面,首先通过对平滑后的信号进行特征信号的提取,对提取得到的特征信号进行特征向量处理。将信号间的比对转化成空间距离的计算。最后利用动态规划进行逐个相似度匹配,得到最终相似度的大小,进而判断出剪切工具。在理论研究与试验仿真的基础上,通过工具痕迹激光检测装置对痕迹信号进行采集。然后软件实现和试验分析测试相结合,来对本文所提出来的算法进行验证分析,进而判断出该算法的有效性和正确性。
[Abstract]:In the criminal technique, the trace of the shear tool is one of the key directions of its research. The tool mark is one of the most frequently occurring marks in many cases such as theft, robbery, and killing. It is of great significance to identify the nature of the case, determine the case and confirm the criminal suspect. According to incomplete statistics, more than 70% of the criminal cases have tool marks in the field, and some areas have some 80% of the tool marks. And the tool trace has the characteristics of being difficult to be damaged, difficult to camouflage, high in appearance rate and good in identification value, and all of these advantages are difficult to compare with other types of trace. Therefore, it is of great practical significance to study the extraction and analysis of the tool marks. In this paper, on the acquisition of trace signal, a tool-trace laser detection device based on LabVIEW is used. The device can effectively collect the trace signal on the tool trace. In the process of acquiring the signal, due to the existence of non-limiting factors such as the reflection and the device, the presence of the abnormal data and the noise interference in the signal obtained by the scanning acquisition can be present. In this paper, the method of K-Means algorithm is used to repair the abnormal data in the signal, and the test simulation is used to verify that the algorithm has good repair effect on the abnormal data in the laser detection signal of the shear tool. In this paper, we focus on the smoothing of the data through the LOWESS (local weighted regression point smoothing method) algorithm, which can eliminate the noise in the scan data to a maximum extent. And then the validity of the algorithm is verified by the test simulation. The method comprises the following steps of: firstly, carrying out characteristic signal extraction on the smoothed signal at a similarity ratio, and carrying out feature vector processing on the extracted feature signal. The ratio of the signals to the space distance is calculated. And finally, the dynamic programming is used for matching the similarity degree one by one, so that the size of the final similarity is obtained, and then the shearing tool is judged. On the basis of the theoretical research and the test simulation, the trace signal is collected by means of the tool trace laser detection device. And then the software implementation and the test analysis test are combined to carry out the verification and analysis on the algorithm presented in the paper, so as to judge the validity and the correctness of the algorithm.
【学位授予单位】：昆明理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：D918.91

【相似文献】