改进的分形算法在弱小目标检测中的应用

针对单帧图像中的弱小目标检测问题,提出了一种改进的基于分形的快速检测方法。该方法首先利用图像的局部熵信息对目标进行粗定位,以得到一个包含目标的感兴趣区域,然后利用分形理论构造该区域的分维像,最后对分维像采用自适应阈值分割即可将弱小目标精确检测出来。与传统分形算法相比,提出的改进算法包含粗定位和细定位两部分,它将分形算法要处理的区域缩减到局部熵所估计的小范围内,从而克服了传统分形方法计算量大、抗噪性差的缺点。仿真实验结果表明,该方法能够稳健、快速、有效地检测弱小目标。

Hit accuracy analysis of guidance improved conventional aerial bomb

To satisfy the operational requirement of guidance improvement for conventional aerial bombs, the scheme of a closed-loop system for the guidance and control system utilizing the stable control loop and projectile is studied, and the design idea of the stability control loop is put forward. Firstly, the representation of the main errors and the model of distribution are analyzed which impacts the hit accuracy of the guidance improved conventional bomb. Secondly, the analytical method of hit accuracy based on Monte Carlo is given. Finally, A and B two set of experiments are used to analyze the influence of the distribution rule of drop point on the circular error probability (CEP) and hit accuracy. The simulation results show that the analysis method is scientific and reasonable, and the design scheme of the control system is feasible.