SoftwareImpacts / SIMPAC-2023-375

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

An-image-segmentation-method-for-the-physical-touching-soybean-seeds

This algorithm is designed for create individual soybean seed images (227x227 pixels) from the images of the physical touching soybean seeds (3072x2048 pixels).

03.04.2023

Note:main.cpp should be recompiled.

Requirements

  • C/C++
  • OpenCV Library(3.4.8)

Operating environments and dependencies

  • Windows ≥10
  • Ubuntu 18.04.6

Content

  • create_file.h: creates files that are used to save images.

  • create_img: creates the paths that use to save images.

  • External_matrix.h: locates those individual soybean seed on the original soybean image and then crops out them from the original soybean image. Finally, it resizes those individual soybean seed images to 227*227 pixels.

  • MSRCR.h,MSRCR_Function.h,MSRetinexCR.h : they are three parts of Multi-scale Retinex with Color Restoration (MSRCR). MSRCR is used to enhance the images' contrast and address the effect of uneven illumination.

  • Otsu.h: creates a threshold that is used to .

  • Read_txt.h: reads the image's name in the path.txt.

  • Watershed.h: is used to divide the physical touching seeds on the masking soybean images.

  • main.cpp: is the overall flow of the image segmentation algorithm where other modules are called.

./src: there are 2 images of the physical touching soybean seeds for testing this algorithm.

path.txt: is used to save images' name. In path.txt, those images would be addressed by this algorithm.

Example path.txt: Image_20220426214755980 Image_20220430064718171 . .

About

License:MIT License


Languages

Language:C 59.9%Language:C++ 40.1%