摘要
由于旋转角度多样性、极端的尺度差异的影响,遥感图像中的飞机检测目前仍存在挑战。为了解决旋转和尺度的问题,目前的策略是将现有的自然场景下的目标检测算法(如Faster R-CNN、SSD等)直接迁移到遥感图像中。这些算法的主干网络复杂,模型占用空间大,难以应用到低功耗和嵌入式设备中。为了在准确率不降低的情况下提高检测速度,本文提出了一个仅包含9层的卷积神经网络来解决飞机检测问题。该网络采用了由粗到细的策略,通过级联两个网络的方式减少计算开销。为了评估方法的有效性,我们建立了一个针对飞机检测的遥感数据集。实验结果表明,该方法超越了VGG16等复杂的主干网络,达到了接近主流检测方法的性能表现,同时显著降低了参数量并使检测速度提高了2倍以上。
Detecting airplanes from remote sensing images remains a challenging task,since the images of airplanes always have the characteristics of multiple rotation angles and severe scale change.In order to solve these problems,the most commonly used strategies are to transfer the existing mainstream object detection algorithms based on natural scenario into the remote sensing images directly,such as Faster R-CNN or SSD.However,the backbones of such networks are generally heavy and occupying large space,which are difficult to be applied to low-power consumption devices or front-end embedded systems.To this end,we designed a simple convolutional neural network architecture with only 9 convolutional layers for airplane detection.Our method adopted a coarse-to-fine strategy by cascading a two-stage network,which further reducing the computation cost of detection.Finally,we built a remote sensing dataset for airplane detection to verify our proposed method.The experimental results show that compared with heavy backbone networks such as VGG16,the performance of our method is close to popular methods,but with much less parameters and more than 2 times higher detection speed.
作者
王晓林
苏松志
刘晓颖
蔡国榕
李绍滋
WANG Xiaolin;SU Songzhi;LIU Xiaoying;CAI Guorong;LI Shaozi(Intelligent Science&Technology Department,Xiamen University,Xiamen 361005,China;Computer Engineering College,Jimei University,Xiamen 361005,China)
出处
《智能系统学报》
CSCD
北大核心
2020年第4期697-704,共8页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金项目(61806172,41971424)
厦门市海洋与渔业局海洋科技成果转化与产业化示范项目(18CZB033HJ11).
关键词
飞机检测
遥感图像
级联
深度学习
卷积神经网络
两阶段
由粗到细
嵌入式设备
airplane detection
remote sensing images
cascade
deep learning
convolutional neural network
two-stage
coarse-to-fine
embedded device
