Skip to content

Change detection annotation tool built on the well-used labelme, so called labelme_cd.

Notifications You must be signed in to change notification settings

shinianzhihou/labelme_cd

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

LABELME_CD

这个作者很懒,名字都没改,代码没咋变,只增加了个变化检测的标注功能。

labelme_cd_start.mp4

基于labelme构建的用于变化检测的标注工具。

安装

conda create -n labelme_cd python=3.9 -y
conda activate labelme_cd
conda install pyqt -y

cd labelme_cd
python setup.py install

使用

  1. 首先将待标注数据以 assets/examples/change_detection/ 的方式放好(支持软连接)

  2. A/B/ 表示待标注的两个时相目录,下面的命名保持一致, label/ 用来存储标注结果

  3. 通过如下方式启动软件进行标注:

    # option-1
    labelme_cd
    # then chose the specific dir (A/ or B/)
    
    
    # option-2
    labelme_cd ./assets/examples/change_detection/A/
    # or
    # labelme_cd ./assets/examples/change_detection/B/
    
  4. 转化为 mask (可选)

    python3 tools/labelme2mask.py -i assets/examples/change_detection/label/ -o assets/examples/change_detection/label/ -f 255

特性

这个作者很懒,下面的东西下次再整理放上来

  • 支持两张图分别进行亮度/对比度调节、频域对齐和直方图校正。

  • 支持lazy_cd中目标/像素级算法辅助标注

  • 理论上可以标注所有的 A+B 图像,具体可以参考 ./assets/examples/ 下面的其它例子,如 RGB-IR 的场景标注

声明

About

Change detection annotation tool built on the well-used labelme, so called labelme_cd.

Topics

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages