一种红外图像快速目标检测方法  

A Fast Target Detection Method for Infrared Images

作　　者：王宁 周铭[1] 杜庆磊[1] 王冰 WANG Ning;ZHOU Ming;DU Qinglei;WANG Bing(Air Force Early Warning Academy,Wuhan 430019,China)

机构地区：[1]空军预警学院,武汉430019

出　　处：《弹箭与制导学报》2020年第4期24-28,33,共6页Journal of Projectiles,Rockets,Missiles and Guidance

摘　　要：为解决红外图像目标检测问题,着重从红外图像预处理、背景抑制和图像分割展开工作。预处理中采用灰度分级调整灰度变换权值的方法;背景抑制采用根据成像距离调整Top-hat变换滤波算子的尺寸;图像分割借鉴二维灰度直方图中阈值划分的相关思路,构建截距直方图,根据最大类间方差准则计算出最优门限分割红外图像;最终实现红外图像快速目标检测。该方法有噪声抑制能力,对单目标、多目标都实现了较好的红外目标检测效果,且精度好,有一定的应用前景。In order to realize infrared image target detection,this paper focuses on infrared image preprocessing,background suppression and image segmentation.In preprocessing,gray scale grading is used to adjust the weights of gray scale transformation;in the background suppression,the size of Top-hat transforms filter operator is adjusted according to the imaging distance;in image segmentation,the intercept histogram is constructed by referring to the relevant idea of threshold division in two-dimensional gray scale histogram,and the optimal threshold segmentation is calculated according to the criterion of maximum inter-class variance.Ultimately infrared image fast target detection is achieved.This method has the ability of noise suppression,and achieves a good infrared target detection effect on single target and multi-target,and has a certain application prospect of good accuracy.

关 键 词：红外图像 TOP-HAT 阈值选取 最大类间方差法 

分 类 号：TN911.73[电子电信—通信与信息系统]