基于混合高斯和帧间差分的机场跑道入侵检测

Runway Incursion Detection Based on Multi-Gaussian and Frame Differencing

入侵机场跑道的威胁目标检测,难点在于算法的高精度和实时性。针对传统混合高斯背景差分运动目标检测算法自适应性较差的缺点,提出一种混合高斯背景差分与帧间差分相结合的运动目标检测算法,将帧间差分的结果反馈到混合高斯模型中,实现光线突变时高斯模型快速收敛,再进行图像后处理以获得精准的运动威胁目标。在Matlab仿真平台上进行实验,结果表明,提出的算法兼顾了检测的速度和精度,分别可达10-1秒级和像素级,满足了入侵机场跑道的威胁目标检测的需求,为机场终端区跑道入侵检测提供了有效的方法。

The difficulty of runway incursion lies in the accuracy and timeliness of the detection algorithm. A detection algorithm for motion objects proposed based on Gaussian mixes model and frame differencing. The increment of illumination supplied by the frames differencing is compensated to the Gaussian Mixture Model so the Gaussian Mixture Model can converge quickly, which can remedy the defect of the GMM when the illumination has jumping changes. Then, post-processing is done to get accurate motion objects. In the software development platform of Matlab7.1, comparative experiments Gaussian Mixture Mode[ background differencing method and the algorithm proposed by this paper are made. The simulation results show that this algorithm can detect the motion objects effectively and real- time.