基于卷积神经网络的毫米波图像目标检测

在公共安全检查领域中,研究毫米波图像目标检测的快速性和精准性的方法具有非常重要的实际应用价值。提出了基于Faster R-CNN深度学习的方法检测隐藏在人体上的危险物品。该方法将区域建议网络(region proposal network,RPN)和VGG16训练卷积神经网络模型相结合,接着通过在线难例挖掘(online hard example mining,OHEM)技术优化训练所提出的网络模型,从而构建了面向毫米波图像目标检测的深度卷积神经网络。实验结果证明所提的方法能高效地检测毫米波图像中的危险物品,并且目标检测的平均精度高达约94.66%,检测速度约为6帧/s,同时对毫米波安检系统的智能化发展有着极其重要的参考价值。

Millimeter Wave Image Object Detection Based on Convolutional Neural Network

In In the field of public security inspection, the method of studying the rapidity and accuracy of millimeter-wave image object detection has very important practical application value. It proposes a method based on Faster R-CNN deep learning to detect dangerous objects hidden in the human body. The method combines the Region Proposal Network (RPN) with the VGG16 training convolutional neural network model. Then, through the Online Hard Example Mining (OHEM) technology, the network model proposed by the training is optimized. Thus, a deep convolutional neural network for millimeter wave image object detection is constructed. The results show that the proposed method can efficiently detect dangerous objects in millimeter wave images, and the average accuracy of object detection is about 94.66%, and the detection speed is about 6fps. At the same time, it has extremely important reference value for the intelligent development of millimeter wave security inspection system.