• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Asian conference on remote sensing;ACRS >An Open Source Image Segmentation Tool for Remote Sensing Images
【24h】

An Open Source Image Segmentation Tool for Remote Sensing Images

机译：一个用于遥感图像的开源图像分割工具

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

摘要

There is a growing interest in open source alternatives to expensive commercial packages for extracting information from remotely sensed images. The benefits of open source software go beyond cost savings on software licenses. For researchers, open source software provides the ability to explore existing algorithms as well as provide access to latest algorithms, by looking at the source code and making changes as required. Image Segmentation is a process of partitioning an image into homogenous non-overlapping regions of related content. The segmentation algorithms are based on two basic properties: discontinuity and similarity. An attempt is made here to develop an open source tool for image segmentation using python. The Python language is particularly beneficial in allowing the system to be expanded as there are a large number of Python libraries already available. It also provides good interoperability with other programming languages allowing existing code to be incorporated as part of the workflow. The tool is based on pull down menu driven graphical user interface. The tool provides basic file operations, filtering, segmentation, morphological operations and feature extraction routines. Laplacian edge detection. Gaussian smoothing. Canny, Gabor and Hough transform form a part of the filtering based segmentation. Erosion, Dilation. Opening and Closing are implemented under morphology. Segmentation menu includes Thresholding, Region Growing and Watershed segmentation techniques and work flow for road and building detection is implemented as feature extraction. The tool is particularly useful for feature extraction. One of the key advantages of the system proposed is that its modular nature allows other packages and routines to be incorporated alongside and to build new algorithms on top of the existing framework.
机译：人们对开放源代码替代昂贵的商业软件包的兴趣日益浓厚，这些软件包可从遥感图像中提取信息。开源软件的好处不仅限于节省软件许可成本。对于研究人员而言，开源软件可以查看源代码并根据需要进行更改，从而能够探索现有算法并提供对最新算法的访问。图像分割是将图像划分为相关内容的同质非重叠区域的过程。分割算法基于两个基本属性：不连续性和相似性。此处尝试开发一种开放源代码工具，以使用python进行图像分割。 Python语言在允许扩展系统方面特别有利，因为已经有大量的Python库可供使用。它还提供了与其他编程语言的良好互操作性，从而允许将现有代码合并为工作流的一部分。该工具基于下拉菜单驱动的图形用户界面。该工具提供了基本的文件操作，过滤，分段，形态学操作和特征提取例程。拉普拉斯边缘检测。高斯平滑。 Canny，Gabor和Hough变换构成了基于过滤的分段的一部分。侵蚀，膨胀。打开和关闭是在形态下实现的。分割菜单包括阈值分割，区域增长和分水岭分割技术，用于道路和建筑物检测的工作流程被实现为特征提取。该工具对于特征提取特别有用。提出的系统的主要优点之一是它的模块化特性允许将其他包和例程并入并在现有框架之上构建新算法。

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