基于数学形态学的弱点状运动目标的检测

提出了一种新的基于数学形态学的红外图像序列中弱点状运动目标的非参数检测算法。采用数学形态学抑制背景杂波干扰和增强目标,用沿时间轴投影和二维空域搜索代替复杂的时空三维搜索形成组合帧,然后在每条可能的轨迹上将进行目标能量累加,实现了一种快速检测前跟踪(TBD)检测算法。仿真实验表明:在恒虚警概率条件下,该检测算法能高效地检测信噪比约为2的弱点状运动目标,检测性能对噪声分布不敏感,能精确地得到目标的即时位置和速度信息,适合于实时图像处理和目标探测,具有很高的实用价值。

Algorithm based on mathematical morphology for dim moving point target detection

A new nonparametric algorithm based on mathematical morphology is proposed for the detection of dim moving point target in infrared image sequence. The mathematical morphology method is used for background clutter suppression and target enhancement. In order to get a quick tracking-before-detection algorithm, the 3D spatio-temporal scanning for target is reduced to 2D space hunting by means of integration of the image sequence output after background suppression which can produce a composite frame of images. From the simulation results of real infrared images, the algorithm can successfully detect dim moving point target with the signal-to-noise ratio equaling 2 and obtain its real-time position and velocity information accurately under the condition of constant false-alarm probability(CFAR). The algorithm useful in many practical applications is insensitive to the changes of noise's statistical distribution and adaptable to real-time image processing and target detection.