基于激光散斑图像的目标识别研究

发布时间：2018-01-20 19:20

本文关键词： 激光散斑统计特性目标识别 SIFT算法　出处：《长春理工大学》2015年硕士论文　论文类型：学位论文

【摘要】：当激光照射物体表面上时,物面的散射光在光场随机叠加干涉形成的斑点状图案便是激光散斑图像。研究表明激光散斑图像包含被测物体的某些特征信息。激光散斑图像具有物体独特性、非接触、速度快、易采集、精度高,便于图像处理等多方面优点,应用散斑图像可以实现检测和识别,具有广阔的应用前景。本文对激光散斑图像进行分析与研究,总结出激光散斑图像与物体之间对应的关系,为目标识别实验奠定理论基础;深入研究SIFT算法,利用SIFT算法对激光散斑图像进行特征提取;通过搭建实验平台,实现激光散斑图像的获取,对同一物体进行小位移变量、不同材质物体、同一材质不同表面物体的激光散斑图像进行匹配识别。通过实验结果,初步验证基于激光散斑图像目标识别的可行性。
[Abstract]:When the laser hits the surface of the object. The speckle pattern formed by the random superposition of the scattering light on the surface of the object is the laser speckle image. The study shows that the laser speckle image contains some characteristic information of the object being measured. The laser speckle image is unique to the object. Non-contact, fast, easy to collect, high accuracy, easy to image processing and other advantages, the application of speckle image can achieve detection and recognition. This paper analyzes and studies the laser speckle image, summarizes the relationship between laser speckle image and object, and lays a theoretical foundation for target recognition experiment. The SIFT algorithm is studied deeply, and the feature extraction of laser speckle image is carried out by using SIFT algorithm. By setting up the experimental platform, the laser speckle image can be acquired, and small displacement variable and different material object are carried out on the same object. The laser speckle image of the same material with different surface is matched and identified. The feasibility of target recognition based on the laser speckle image is preliminarily verified by the experimental results.
【学位授予单位】：长春理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP391.41;TN249

【参考文献】