光子成像静止点目标的管道滤波探测方法

构建了具有单光子成像能力的光子成像系统,提出应用管道滤波方法实现对光子受限静止点目标的探测。分析了光子图像静止点目标与点状噪声特征,根据静止点目标在序列图像中位置的确定性以及噪声点的不相关性,研究了基于管道滤波的光子成像静止点目标探测方法。为降低目标探测的虚警概率,优化了管道滤波直径。以实验采集得到的多组光子图像序列为样本,获得了探测概率、虚警概率与信号光子数、噪声光子数、管道长度以及检测阈值的关系。检测结果显示,对信号与噪声的平均发生率为0.4和5.215的序列图像,当管道长度为9、检测阈值为2时,探测概率达0.9以上且虚警概率〈0.08。对比多组图像序列的检测结果表明,影响探测概率的主要参数是信号光子数,而影响虚警概率的主要参数是噪声光子数。

Pipeline filtering detection of stationary point targets in photon images

A photon imaging system with single photon detecting and imaging capabilities is constructed and a detection method based on pipeline filtering is proposed to detect the stationary point target under a photon limited condition. The stationary point target detection method based on pipeline filtering is studied on the characteristics of position and noise for stationary point targets in photon images. The diameter of pipeline filter is optimized to reduce the false alarm rate. The relationships of detection performance with the number of signal photons and noise photons, the pipeline length and the detection threshold are explored through photon image sequences. For the image sequence with the average incidences of signal and noise of 0.4 and 5.215 respectively, the detection rate can be more than 0.9 and the false alarm rate less than 0.08 when the pipeline length is 9 and detection threshold is 2. The conclusion from experimental results is that the detection rate is mainly affected by the number of signal photons and the false alarm rate is dependent on the number of noise photons.