多模自适应目标图像跟踪

MULTIPLE MODELS ADAPTIVE TARGET IMAGE TRACKING

本文建立了图像序列中含加性和乘性两种噪声模型影响下的目标偏移值测量模型。并针对目标机动性，采用多模虑波器和两段卡尔曼虑波器对目标状态进行估值，有效抑制了各种噪声影响，又消除了显式的目标机动性检测，且有较高的实时性和较少的存储量等优点。

In this paper,the target image position offset measurement models in image sequence are developed under the effects of additive and multiplitive noise In consideration of the target maneuvering,the target states are estimated by multiple model filters and two stage Kalman filters,depressing the effects of noise and avoiding the explicit maneuvering detection The algorithm has good performance with high speed and small savings