时空过采样系统及其在点目标检测中的性能仿真

从成像探测体制出发，对过采样扫描系统的光学系统、电子学系统等模块分别进行了数学建模，并对点目标和场景图像进行了仿真实验。仿真过程中提出了一种对有限图像大量分割代替连续图像的建模思想和方法，在matlab环境中对源点目标图像和场景以及探测器扫描积分过程进行仿真，并对过采样图像进行图像融合。仿真结果给出了不同过采样倍数以及点目标位于不同相位时的信噪比变化和信噪比随积分时间增长的变化曲线。仿真结果表明，过采样探测系统融合后的图像虽然能够增大点目标在图像上的尺寸，但融合后的图像会增大高频噪声，造成检测概率降低，虚警率增大。文章最后给出了利用小孔模拟点目标成像的实验，进一步证实了仿真结果的正确性，说明时空过采样系统并不适合用于点目标检测。

Temporal-spatial oversampling system and its performance analysis in point target detection

This paper focuses on the mathematical modeling of an oversampling scanning optical system and an electronic system as well as the imaging simulation for point targets and scene images. It proposes new modeling idea and method to divide the finite images into large numbers of simulate analog images. By using this method, the processes of point target imaging, scene imaging and scanning and energy integral are simulated in matlab, and reconstructed images are obtained. The simulation experiment presents the trends of Signal to Noise Ratios(SNRs) with different oversampling times and the SNRs of point target located at different phases, as well as the trend of SNR with integration time increasing. The result shows that the reconstructed image increases the high frequency noise so that to reduces the detection probability and enlarges the false alarm, although it increases the size of the point target in the reconstructed image. At last, it gives the result of the point target imaging experiment based on pinholes, which verifies further the validity of the simulation results, and points out that the oversampling system is unfit for the point target detection.