This tools is implemented by Qt, which is convenient to label ROIs in image using .shp format and crop these ROIs batchingly as data(256x256) in deep learning. Moreover, it could display image and segmentation correspondingly on linked widows and may be helpful to learn QGraphics(View, Scene, Item).
GDAL
OpenCV
Last tested successfully using gdal 2.0.3, opencv 3.3.0.
New folder in directory, for example: build. cd build in cmd or powershell, then cmake -G "Visual Studio 14 Win64" ..
Then you need to update gdal and opencv to include directory and library directory in VC++ directory(project properties), And update opencv(gdal_i.lib, opencv_world330d.lib, opencv_world330.lib) to Linker->input->additional dependencies.
Display: drag and drop single image file or both image and its corresponding segmentation files directly
Menu:
- Mode: Browse or Paint. you can label regions on Paint Mode(double clicks completes painting)
- Clip: First choose image file to be clipped, then shp file corresponding images to be clipped , Finally output file
- ...
You may need to learn about Qt firstly, Qt 实战一二三. Also GDAL(using patch-read to load larger raster data).
