基于最大类间方差法的信号弹图像分割算法研究

基于最大类间方差法的信号弹图像分割算法研究摘要信号弹是采用自身的光信号、声信号以及烟信号实现对相关人员的报警、定位的特殊弹药，因其本身具有携带方便、定位准确、传播距离远等特点而广泛的应用在各地的军队中。在白天光照强度较大时，发射信号弹升空时，常会捕捉不到，造成关键信息缺失的问题。本文针对信号弹图像，对采集的信号弹图像进行最大类间方差法进行处理，并对处理后图像进行特征值的提取，实现信号弹的识别。本文具体的研究工作如下：(1)详细分析了信号弹图像识别检测的重要意义，分别运用多种去燥方式实现对信号弹图像的去燥。利用多帧叠加以及选择合适的窗位和窗宽的方式实现对图像的优化。(2)图像经过优化增强后，利用最大类间方差法实现对信号弹图像的分割，选择合适阈值实现对于图像的阈值分割。(3)针对分割后的信号弹图像，根据信号弹图像的外形轮廓、纹理信息以及相关的特征实现信号弹图像特征值的提取。关键词：信号弹：最大类间方差：阈值分割：特征值提取abstractThe signal bomb is a special ammunition that uses its own optical signal,acoustic signaland smoke signal to realize the alarm and positioning of related personnel.It is widely used inthe military around the country because of its convenient carrying,accurate positioning,longtransmission distance and other characteristics.When the light intensity is high in the daytime,when the signal bomb is launched,it is often not captured,resulting in the problem of missingkey information.In this paper,the signal bomb image is processed by the maximum inter classvariance method,and the processed image is extracted with the eigenvalues to realize therecognition of the signal bomb.The specific research work of this paper is as follows:(1)The significance of signal bomb image recognition and detection is analyzed in detail,and various drying methods are used to achieve the signal bomb image drying.The image isoptimized by multi frame superposition and selecting appropriate window level and windowwidth.(2)After the image is optimized and enhanced,the maximum inter class variance method isused to segment the signal bomb image,and the appropriate threshold is selected to segment theimage.(3)For the segmented signal flare image,the feature value of the signal flare image isextracted according to the contour,texture information and related features of the signal flareimage.Keywords:signal flare;Maximum inter class variance;Threshold segmentation;Eigenvalueextraction目录摘要abstract...目录基于最大类间方差法的信号弹图像分割算法研究.2一、绪论….2(一)课题研究背景和意义.….2(二)国内外研究的现状3(三)信号弹图像检测发展趋势.4(四)论文的工作安排….4二、信号弹图像质量优化…(一)信号弹图像检测系统51.测量系统基本结构」2.信号弹图像噪声分析3.多帧叠加去噪.64.反锐化掩膜算法.9三、信号弹图像分割..11(一)图像边缘检测分割方法.11(二)传统阈值分割方法.….12(三)动态阈值分割方法…。.13(四)ROI的应用.15四、信号弹图像特征值提取….17(一)信号弹图像阈值分割结果重构17(二)信号弹图像特征值的提取.191.特征提取.192.特征参数选择】.19(三)信号弹烟雾轮廓的标记.20五、总结.22参考文献.23郑重声明….24