红外激光特性远小弱特征的高精度增强分割技术研究  

Research on high precision enhanced segmentation of infrared laser characteristics

作　　者：孙艳君[1] 刘正光[2] 王磊[3] 

机构地区：[1]天津天狮学院信息与自动化学院,天津301700 [2]天津大学电气与自动化工程学院,天津300072 [3]天津城建大学,经济与管理学院,天津300384

出　　处：《激光杂志》2017年第3期133-136,共4页Laser Journal

基　　金：国家自然科学基金青年基金项目(71303170);天津市教委科技发展战略研究计划项目(15ZLZLZF00670);天津市普通高校本科教学质量研究项目(C03-0809)

摘　　要：针对传统的图像增强分割算法抗干扰性能不强,容易出现白平衡偏差和失真的问题,提出一种新的红外激光特性远小弱特征目标图像增强分割算法。对采集的红外激光图像远小弱特征进行最大似然比检测滤波处理,对降噪输出的图像进行仿射不变矩特征提取,在仿射不变区域进行模板特征匹配,提高图像分割能力。仿真结果表明,采用该算法进行远小弱特征红外激光成像增强处理后,输出信噪比较高,图像分割的特征配准性能较好,各项技术指标优于传统方法。In view of the traditional segmentation algorithms for image enhancement have the anti-jamming ability is not strong,it＇s easy to have a white balance deviation and distortion of the problem,put forward a new infrared laser feature far small weak target image segmentation algorithm. For collection of far infrared laser image small weak characteristics to filtering processing of maximum likelihood ratio detection,the noise of the output image affine moment invariant feature extraction,in the region of the affine invariant feature template matching,improve the ability of image segmentation. The simulation results show that this algorithm is far small weak characteristic infrared laser imaging enhancement processing,output,high signal noise ratio（ SNR）,the characteristics of the image segmentation registration performance is good,various technical indicators are superior to traditional methods.

关 键 词：红外激光 远小弱特征目标 图像增强 分割 降噪 

分 类 号：TN249[电子电信—物理电子学]