基于仿射不变离散哈希和条件随机场的遥感图像目标检测
作者:
作者单位:

1.南京理工大学计算机科学与工程学院,南京 210094;2.江苏科技大学计算机学院,镇江 212100

作者简介:

通讯作者:

基金项目:

国家自然科学基金(61673220)资助项目。


Object Detection of Remote Sensing Image Based on AIDH and CRF
Author:
Affiliation:

1.School of Computer Science and Engineering, Nanjing University of Science and Technology, Nanjing 210094, China;2.School of Computer, Jiangsu University of Science and Technology, Zhenjiang 212100, China

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
    摘要:

    建立了一种结合仿射不变离散哈希(Affined-invariant discrete hashing, AIDH)和条件随机场(Confidential random field, CRF)的模型,实现遥感图像的目标检测。对遥感图像进行超像素分割,构建适用于CRF的以超像素块为顶点的无向图结构。以超像素块作为测试样本,使用AIDH学习方法作为CRF一元势函数,生成初始类别标签。采用Potts模型构建CRF的二元势函数进行标签的再学习,平滑目标邻域信息,解决目标检测中的漏判问题。最后,使用基于凸壳边界的方法生成最小外接目标框作为目标检测结果。实验表明,本文方法在目标检测的精度和效率上取得了较好的平衡。

    Abstract:

    By constructing a model combining affine-invariant discrete hashing (AIDH) and confidential random field (CRF) , the object detection of remote sensing image is achieved. Firstly, the remote sensing image is reconstructed by superpixel segmentation, and the undirected graph structure with superpixel block as vertex is constructed for CRF. Then, the superpixel block is used as the test sample for AIDH learning which is used as CRF unary potential function to generate the initial category label. Then, the pairwise potential function of CRF is constructed by using Potts model for label re-learning, while the object neighborhood information is smoothed and the missing area of object detection is resolved. Finally, the convex hull boundary method is used for generating minimum external rectangular frame as object detection result. Experimental results demonstrate that the proposed method achieves the tradeoff of accuracy and efficiency for objection detection tasks.

    表 2 各算法的算法用时对比Table 2 Comparison of time consuming between each algorithm
    图1 基于AIDH-CRF的目标检测流程图Fig.1 Flow chart of object detection based on AIDH-CRF
    图2 基于AIDH的遥感图像目标检测Fig.2 AIDH-based remote sensing image object detection
    图3 凸壳边界法处理前后目标检测结果对比Fig.3 Comparison of target detection results before and after convex hull boundary method
    图4 本文方法各阶段结果示意图Fig.4 Results of each stage of the proposed method
    图5 本文方法目标检测结果图Fig.5 Object detection results of the proposed method
    图1 Derivative of magnetic biasFig.1
    图2 Flow chat of FRUKFFig.2
    图3 Position errors of EKF and UKFFig.3
    图4 Magnetic biases of EKF and UKFFig.4
    图6 Memberships of the fuzzy regulatorFig.6
    图7 Position error of FRUKFFig.7
    表 1 各算法在NWPU VHR-10数据集下的AP和mAP对比Table 1 AP and mAP comparison of each algorithm in the NWPU VHR-10 dataset
    参考文献
    相似文献
    引证文献
引用本文

孔颉,孙权森.基于仿射不变离散哈希和条件随机场的遥感图像目标检测[J].数据采集与处理,2021,36(4):769-778

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
历史
  • 收稿日期:2019-05-28
  • 最后修改日期:2021-06-28
  • 录用日期:
  • 在线发布日期: 2021-09-23