期刊文献+

基于可变形卷积神经网络的遥感影像密集区域车辆检测方法 被引量:21

Vehicle Detection in Remote Sensing Images of Dense Areas Based on Deformable Convolution Neural Network
下载PDF
导出
摘要 车辆检测是遥感图像分析领域的热点研究内容之一,车辆目标的智能提取和识别,对于交通管理、城市建设有重要意义。在遥感领域中,现有基于卷积神经网络的车辆检测方法存在实现过程复杂并且对于车辆密集区域检测效果不理想的缺陷。针对上述问题,该文提出基于端到端的神经网络模型DF-RCNN以提高车辆密集区域的检测精度。首先,在特征提取阶段,DF-RCNN模型将深浅层特征图的分辨率统一并融合;其次,DFRCNN模型结合可变形卷积和可变形感兴趣区池化模块,通过加入少量的参数和计算量以学习目标的几何形变。实验结果表明,该文提出的模型针对密集区域的车辆目标具有较好的检测性能。 Vehicle detection is one of the hotspots in the field of remote sensing image analysis. The intelligent extraction and identification of vehicles are of great significaalce to traffic management and urban construction. In remote sensing field, the existing methods of vehicle detection based on Convolution Neural Network (CNN) are complicated and most of these methods have poor performance for dense areas. To solve above problems, an end-to-end neural network model named DF-RCNN is presented to solve the detecting difficulty in dense areas. Firstly, the model unifies the resolution of the deep and shallow feature maps and combines them. After that, the deformable convolution and RoI pooling are used to study the geometrical deformation of the target by adding a small number of parameters and calculations. Experimental results show that the proposed model has good detection performance for vehicle targets in dense areas.
作者 高鑫 李慧 张义 闫梦龙 张宗朔 孙显 孙皓 于泓峰 GAO Xin;LI Hui;ZHANG Yi;YAN Menglong;ZHANG Zongshuo;SUN Xian;SUN Hao;YU Hongfeng(Key Laboratory of Technology in Geo-spatiaI Information Processing and Application System fnstitute of Electronics,Chinese Academy of Sciences,Beijing 100190,China;University of Chinese Academy of Sciences,Beijing 100049,China;University of Central Lancashire(UCLan),Preston PR1 2HE,United Kingdom)
出处 《电子与信息学报》 EI CSCD 北大核心 2018年第12期2812-2819,共8页 Journal of Electronics & Information Technology
基金 国家自然科学基金(41501485)~~
关键词 遥感影像 车辆检测 密集区域 端到端卷积神经网络 Remote sensing images Vehicle detection Dense areas End-to-end convolution neural network
  • 引文网络
  • 相关文献

参考文献3

二级参考文献6

共引文献125

同被引文献179

引证文献21

二级引证文献129

;
使用帮助 返回顶部