摘要
针对目前红外扫描成像引信在目标识别过程中处理时间慢、对图像尺度和旋转变换敏感的问题,提出了基于泽尔尼克矩和支持向量机的红外扫描图像识别算法。该算法利用不变矩重构了引信探测装置所扫描到的红外图像,求得图像的各阶泽尔尼克矩来构成描述扫描红外图像的特征因子m,最后利用支持向量机对目标与相似物进行分类,实现对目标的识别。仿真验证表明,基于泽尔尼克矩的图像识别算法能够较快实现目标特征的提取和识别,对低分辨率图像也有较好的处理效果,适用于扫描式引信探测系统的简易成像处理和目标识别。
In order to solve the problem of slow processing and sensitive to image scale and rotation transformation in infrared imaging fuze,a new method of infrared scanning image recognition based on Zernike moment and support vector machine was proposed.The method reconstructed the infrared image scanned by the fuze detection device by invariant moments and obtained the Zernike moment of each order of the image to form a feature factor m describing the scanned infrared image.Finally,the support vector machine was used to measure the target and the analogues to classify,to achieve the goal of identification.Simulation results showed that the image recognition algorithm based on Zernike moments could achieve the target feature extraction and recognition quickly,and also had better processing effect for low resolution images,which was suitable for simple imaging processing and target recognition of scanning fuze detection system.
作者
张晨
陈遵田
唐辉
ZHANG Chen;CHEN Zuntian;TANG hui(Science and Technology on Electromechanical Dynamic Control Laboratory,Xi'an 710065,China)
出处
《探测与控制学报》
CSCD
北大核心
2018年第3期56-60,共5页
Journal of Detection & Control
关键词
引信探测
红外扫描成像
目标识别
泽尔尼克矩
特征因子
fuze detection
infrared scanning image
target recognition
Zernike moment
feature factor