一种基于Gabor深度学习的无人机目标检测算法

A Target Detection Algorithm for UAV Based on Gabor Deep Learning

随着智能控制技术的不断成熟,无人机给军事领域带来快速发展的同时也带来了威胁.因此针对空中飞行的无人机进行实时检测的任务需求,设计了一种基于Gabor深度学习的无人机目标检测算法.首先,搭建基于Gabor滤波器的深度神经网络,输入的图片经过该网络进行网格化划分,用以特征提取;然后,针对每个格子的特征利用回归算法计算其中物体的位置信息,并利用分类算法计算物体的类别信息,对以上得到的回归和分类结果进行筛选、融合得到最终的检测结果;最后,采集空中飞行的无人机真实数据构建数据集,在此基础上进行网络模型训练和算法验证.

With the continuous maturity of intelligent control technology,unmanned aerial vehicles(UAVs)bring rapid development to the military field and also pose threats simultaneously.Therefore,for the mission requirements of real-time detection of UAVs in flight,a target detection algorithm is designed for drones based on Gabor deep learning.Firstly,a deep neural network based on Gabor filter is built.And the input picture is meshed by the network for feature extraction.Then,the regression algorithm is implemented to calculate the position information of the contained object,the classification algorithm is used to calculate the category information of the object,and the final test results for targets detection are obtained by screening and merging the regression and classification results obtained above.Finally,the real image acquisition of the UAVs in the air is used to construct the data set,and the network model training and algorithm verification are performed on this basis.