摘要
针对基于单特征不能准确识别复杂环境下空域目标问题,利用偏最小二乘法(PLS)预处理飞机图像特征,通过变量投影重要性分析法(VIP)选取有效图像特征,并将有效特征融合到粒子滤波算法中,实现了颜色与形状特征融合的粒子滤波跟踪算法。该方法在雾霾和遮挡情况下的跟踪结果准确,且跟踪速度得到提高。
Under complex environment about the detetcing and tracking of aircraft,it is impossible to track target accurately by only a feature. The partial least square(PLS) method was introduced to preprocess images of features,the VIP was applied to select valid images of features,and the effective features are integrated into the particle filter algorithm To achieve a particle filter tracking algorithm color and shape feature fusion. The experimental results show that the method is accurate in tracking results under the environmental changes such as fog and occlusion, and also verifies the improvement of tracking speed.
引文
[1]汪鸿翔,柳培忠,骆炎民,等.单目标跟踪算法综述[J].海峡科学,2016(7):28-32.Singh P,Deepak B B V L,Sethi T,et al.Real-time object detection and Tracking using color feature and motion[C]//International Conference on Communication and Signal Processing.2015:1236-1241.
[2]吴世东,鲍华,张陈斌,等.基于视觉显著性特征的粒子滤波跟踪算法[J].中国科学技术大学学报,2015(11):934-942.Hossain K,Oh C M,Lee C W,et al.Multi-Part SIFT feature based particle filter for rotating object tracking[C]//International Conference on Informatics,Electronics&Vision.IEEE,2012:1016-1020.
[3]唐雨操.基于车辆识别的车辆追踪与计数[J].信息与电脑,2016(13).
[4]凌永国,胡维平.复杂背景下的车辆检测[J].计算机工程与设计,2016,37(6):1573-1578.
[5]李万臣,刘康.基于支持向量机的高光谱遥感图像分类[J].信息通信,2015,27(2):10-10.
[6]何力,曲仕茹.基于PLS-VIP特征降维的车辆检测[J].中国公路学报,2014,27(4):98-105.Barker M,Rayens W.Partial least squares for discrimination[J].Magnetic Resonance Imaging,2012,30(3):446-452.