Abstract: This paper proposes a space location method of infrared targe based on UKF according to the large error in instantaneous location using one observation of single station.Using active vision for reference，the infrared thermal camera is controlled to move and shoot the target to acquire its positions，and the world coordinates of the camera are acquired by the GPS receiver at the same time.The nonlinear system model of the space location is established using WGS-84 coordinate system and topocentric coordinate system for reference based on transformation equation of perspective projection.The UKF filtering algorithm is introduced to space location.Result of the experiment shows the feasibility and the effectiveness of the proposed method.

Key words: Unscented Kalman Filtering（UKF）, nonlinear filtering, infrared target, space location

摘要： 针对单站系统利用一次观测对目标进行定位，瞬时定位误差较大，提出了基于UKF的红外目标空间定位方法。基于针孔成像模型，借鉴主动视觉思想，通过控制红外热像仪运动拍摄目标图像，获取目标位置，并用GPS接收机测得拍摄点世界坐标。根据透视投影变换方程，借助站心地平坐标系和WGS-84坐标系建立红外目标空间定位的非线性系统模型，在此基础上引入UKF（Unscented Kalman Filtering）滤波算法进行空间定位。实验结果证明了该方法的可行性和有效性。

关键词: 不敏卡尔曼滤波, 非线性滤波, 红外目标, 空间定位