基于上三邻域连续点计数的激光光斑实时检测  

Real-Time Laser Facula Detection Based on Top-Three-Neighborhood- Region Continuous-Pixels Numbering

作　　者：徐文宇[1] 白廷柱[1] 刘洋[1] 梁镇[2] 

机构地区：[1]北京理工大学光电成像技术与系统教育部重点实验室 [2]中国人民解放军第68128部队

出　　处：《光学学报》2013年第12期130-135,共6页Acta Optica Sinica

摘　　要：研究了实时高精度激光光斑检测方法。利用高帧频、高灵敏度CCD采集14位激光光斑视频;分析了激光光斑的特征,在使用阈值分割出光斑区域后,通过上三邻域连续点计数算法检测了激光光斑区域;分析了激光光斑中余光斑存在的原因,利用平均阈值法滤除了余光斑,在剩余的主光斑中计算获得了更为精确的光斑中心(含质心与形心),制定了以参考帧为基准的视频帧序列的操作序列法光斑检测流程,解决了传统相邻帧相减法无法检测逆光斑帧及光斑中心位置不同的连续相邻光斑帧的问题。实验结果表明,算法可实践用于在线实时与离线实时的高精度激光光斑检测。The real-time and high precision detection algorithm of the laser facula is proposed. This work has digitized 14 bit laser facula video with the high frequency and high sensitive CCD; analyzed the feature of the laser facula is analyzed, the laser facula region by top-three-neighborhood-region continuous-pixels numbering algorithm is detected after segmenting the facula region by the threshold; the reason of surplus facula in the whole laser facula region is analyzed. The surplus facula region with the average threshold algorithm is subtracted, the more accurate facula centroid （including the center of mass and the center of figure） is gotten, and made the flow of the facula detection from the video frame, named operation-serial algorithm, which solve the problem of that tranditional neighbored-frame-subtraction algorithm can not dectect the inverse facula frame and the continous facula frame when the centroids of the facula are different. The experimental result shows that the proposed algorithm can be used to the online and offline detection of real-time and high precision laser facula practically.

关 键 词：光学设计 光斑检测 光斑中心 激光光斑 上三邻域连续点计数 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]