• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>International Conference on Computer Vision, Image and Deep Learning >An Island Remote Sensing Image Segmentation Algorithm Based on FC_U-Net Network
【24h】

An Island Remote Sensing Image Segmentation Algorithm Based on FC_U-Net Network

机译：一种基于FC_U-Net网络的岛遥感图像分割算法

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

With the increasing importance of islands in many fields, it has become the focus of research to obtain information from island remote sensing images efficiently by using image semantic segmentation algorithm. In recent years, deep learning methods based on convolutional neural network have been widely used in image segmentation. However, for remote sensing images with rich details and different target scales, there are still some problems such as insufficient segmentation accuracy. To handle these problems, we propose an island remote sensing image segmentation algorithm based on FC_U-Net network. In this network, U-Net and FCN (Fully Convolutional Network) are improved respectively to form two new modules, and then the semantic features and detail features from the two modules are fused by a fusion module. The testing and comparative experiments on NWPU-RESISC45 dataset show that the quantitative metrics and visual effects of FC_U-Net are better than other three state-of-the-art methods, U-Net, FCN (Fully Convolutional Network), SegNet, which indicates the effectiveness of our method.
机译：随着岛屿在许多领域的重要性越来越大，它已成为通过使用图像语义分割算法有效地从岛遥感图像获取信息的重点。近年来，基于卷积神经网络的深度学习方法已广泛用于图像分割。但是，对于具有丰富细节和不同目标尺度的遥感图像，仍存在一些问题，例如分割精度不足。为了处理这些问题，我们提出了一种基于FC_U-Net网络的岛遥感图像分割算法。在该网络中，U-NET和FCN（完全卷积网络）分别改进以形成两个新模块，然后通过融合模块融合两个模块的语义特征和细节特征。 NWPU-RESISC45数据集的测试和比较实验表明，FC_U-NET的定量度量和视觉效果优于其他三种最先进的方法，U-Net，FCN（全卷积网络），SEGNET，其表明我们方法的有效性。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号