侧扫声呐SeabedObjects-KLSG海底声呐数据集(1190)介绍

该数据集出自河海大学霍冠英团队SeabedObjects-KLSG

https://github.com/huoguanying/SeabedObjects-Ship-and-Airplane-datasethttps://github.com/huoguanying/SeabedObjects-Ship-and-Airplane-dataset

 ( 612张有目标图像 + 578张海底图像 =1190
A   385 wreck images, 残骸
B   36 drowning victim images,溺水者
C   62 airplane images, 飞机
D  129 mine images,雷

E   578 seaflfloor images.海底图片

385张残骸图片,

36张溺水者的照片,

62张飞机图片,

129张地雷图像,

578张海底图片。

侧扫声呐SeabedObjects-KLSG海底声呐数据集(1190)介绍_第1张图片

SeabedObjects船舶和飞机数据集(溺水人员&雷&海底图像没有开源)

SeabedObjects(船舶和飞机)数据集目前已开放供acedemic使用。
这个开放的数据集包含385张船舶图像和62张飞机图像,可用于学术目的。

我们要感谢L-3 Klein Associates、EdgeTech、Lcocean、Hydro-techMarine和Tritech在过去几年中为提供有价值的实时侧扫声纳图像提供的大力支持,这对我们数据集的建立做出了很大贡献。
如果您有任何问题,请不要犹豫,通过[email protected][email protected].顺致敬意,

MERSF-Sonar-Image-Segmentation

A Robust and Fast Method for Sidescan Sonar Image Segmentation Using Nonlocal Despeckling and Active Contour Model. This software is a matlab version which implemented the method in the paper "A Robust and Fast Method for Sidescan Sonar Image Segmentation Using Nonlocal Despeckling and Active Contour Model "(IEEE TCYB,2017). To use the software, you simply following the steps below:

  1. Unzip the “MERSF Segmentation Sidescan sonar image.zip”into your desired directory;
  2. Start the Matlab program(a version of R2017B or later is ok);
  3. Open the "\MERSF Segmentation Sidescan sonar image\MERSF" directory;
  4. You will find five files in the current directory;
  5. Simply type "SonarSegmentation" in the cmd window, and press enter key, then a GUI will be loaded;
  6. Press the "OpenImage" button, and choose an image file in the "sonar image set" directory;
  7. The image to be segmented will be loaded and shown, then press the "ReadParameters" button, and the default parameters for that image will be loaded;
  8. You can modify all the parameters to achieve maybe even better results, or you can just press the "TargetDetection" button to start the segmentation;
  9. The Segmentation results with different iterations will be displayed in different images;
  10. The segmentation will automatically stop if it reaches the convergence, and you can press the "SaveResult" button to save the final segmentation results. Note:Parameters should be selected according to the noise level of the image, the denosing method and the regions to be segmented can also be switched.
    If you have any problem in using this software, please donnot hesitate to contact me via emai: [email protected]. Best regards.

提出了一种基于非局部去斑点和活动轮廓模型侧扫声纳图像分割方法。
该软件是一个matlab版本,实现了论文《使用非局部去斑点和活动轮廓模型的侧扫声纳图像分割的鲁棒快速方法》(IEEE TCYB,2017)中的方法。要使用该软件,只需执行以下步骤:

将“MERSF Segmentation Sidescan sonar image.zip”解压缩到所需目录中;

启动Matlab程序(可以使用R2017B或更高版本);

打开“\MERSF分割侧扫声纳图像\MERSF”目录;

您将在当前目录中找到五个文件;

只需在cmd窗口中键入“Sonar Segmentation”,然后按enter键,即可加载GUI;

按“打开图像”按钮,在“声纳图像集”目录中选择一个图像文件;

将加载并显示要分割的图像,然后按“ReadParameters”按钮,将加载该图像的默认参数;

您可以修改所有参数,以获得更好的结果,或者只需按下“TargetDetection”按钮即可开始分割;

不同迭代次数的分割结果将显示在不同的图像中;

如果达到收敛,分割将自动停止,您可以按“SaveResult”按钮保存最终分割结果。注:应根据图像的噪声水平选择参数,也可以切换去噪方法和要分割的区域。

如果您在使用本软件时遇到任何问题,请随时通过emai与我联系:[email protected].顺致敬意,

你可能感兴趣的:(声呐)