Because of the characteristics of low signal-to-noise ratio and small contrast ratio existed in the air target, it’s difficult to separate the target from background for the traditional mixed Gauss algorithm moreover with a big influ-ence of illumination.To solve the problem, an improved algorithm was put forward, which the model was established, initialized and updated;the aspects of generating new Gaussian component and deleting invalid Gaussian component and reducing the redundancy link are optimized;the influence of illuminance abrupt variation was reduced by changing the standard deviation threshold at the same time.The average time of final algorithm single frame operation is close to 1.3 times of the single gaussian model algorithm and under 0.5 times of gaussian mixture model algorithm. The real-time tracking speed and stability of the ground-to-air,air to air target have been effectively improved by the algorithm.%由于空中目标存在信噪比小，对比度低的特点，传统混合高斯背景建模较难地把目标从背景中分离出来，且受光照影响较大。针对此问题，提出一种改进算法，对模型建立、初始化和更新，生成新高斯分量，以及删除无效高斯分量等方面进行优化，减少中间的冗余环节；同时通过改变标准差阈值减小光线突变的影响。最终算法单帧运算所需要的平均时间接近单高斯模型算法的1.3倍，达到混合高斯模型算法的0.5倍以下。算法有效提高了地对空、空对空目标的实时跟踪的速度与稳定性。
展开▼
