基于形态学和高阶统计量的弱小运动目标检测

Moving Dim Target Detection Based on Morphology and High-Order Statistics in Infrared Image

为检测复杂背景中的红外弱小运动目标,提出了一种基于形态学滤波和三阶累积量分析的检测方法。该方法采用形态学背景估计方法对红外图像序列进行预处理,提高图像信噪比,并消除高亮度大尺寸背景和图像边缘对高阶累积量估计造成的影响。利用高阶累积量对高斯噪声是"盲的"这一特点,基于红外图像序列构造三阶累积量作为统计判据检测图像中的弱小运动目标。仿真实验表明,该方法能够连续有效地检测红外弱小运动目标。

An algorithm based on morphology and third-order cumulant is proposed to detect moving dim target under complex background. The algorithm uses morphological pretreatment to suppress the highlight background and the edge and improve the signal-to-noise ratio （SNR）. Based on the fact that the high-order cumulant is blind to Gaussian processes, the moving dim targets are detected as per third- order cumulant binary criterion. The simulations verify the effectiveness of the algorithm.