基于改进RetinaNet的跑道边灯亮度检测

Runway Edge Lights Brightness Detection Based on Improved RetinaNet

为了解决机场跑道边灯亮度检测任务中小目标检测精度低、检测速度慢的问题,提出了一种基于改进RetinaNet的跑道边灯亮度检测方法。在RetinaNet的基础上,引入倒残差结构和深度可分离卷积,以提升网络的特征提取能力和检测速度。采用K-means聚类算法优化目标样本的锚点框尺寸,以提升网络的检测精度。实验结果表明,相比原始方法,本方法的性能有明显提升,平均检测精度达到97.2%,检测速度达到25.9 frame/s。

To solve the problems of low detection accuracy and slow detection speed of small targets in the task of detecting the brightness of the runway edge lights in the airport,a method for detecting the brightness of the runway edge lights based on improved RetinaNet is proposed in this paper.Based on the RetinaNet,the inverted residual structure and depth separable convolution are introduced to improve the feature extraction ability and detection speed of the network.The K-means clustering algorithm is used to optimize the size of the anchor box of the target sample to improve the detection accuracy of the network.The experimental results show that compared with the original method,the performance of the method is significantly improved,with the average detection accuracy of 97.2%and detection speed of 25.9 frame/s.