基于连续波雷达的无人机检测

UAV detection based on continuous wave radar

为了检测和跟踪空中无人机,并且节约成本,本文使用连续波雷达。针对连续波雷达的特点,提出了改进算法。该方法可分为3个阶段:背景去除、可疑目标提取,目标轨迹获取。在第1阶段,使用改进的GMM迭代方法,使用两个高斯混合模型交替迭代来移除背景。第2阶段,获得可疑的无人机目标。在第3阶段,通过改进的TBD方法组合多帧信息来获得可疑目标轨迹,并确认无人机轨迹。实验表明,该方法可以有效地检测并跟踪无人机轨迹,且误报概率较低。

In order to detect and track airborne drones and to save costs, this paper uses continuous wave radar. According to the characteristics of continuous wave radar, an improved algorithm is proposed. The method can be divided into three stages: background removal, suspicious target extraction, and target trajectory acquisition. In the first phase, we use the improved GMM iterative method to remove the background a using improved Double-GMM-Iteration model. In the second phase, the suspicious drone target was obtained. In the third stage, we combined the multi-frame information with the improved track-before-detect(TBD) method to obtain the suspicious target trajectories and determine the drone trajectory. Experiments show that this method can effectively mark the UAV trajectories and the probability of false alarm is low.