Sorghum Semantic Segmentation using machine learning and hyperspectral imaging

sorghum semantic segmentation project related code and data

code

• code for training and testing 7 machine learning methods in R

  • Analysis_7methods.R

• code for training Artificial neural networks (ANNs) in python

  • Analysis_ANNs.py (training)
  • Analysis_ANNs_predict.py (prediction)

• feature selection results

  • important_wavebands.xlsx

hyperspectral_cubes

• Sorhgum and maize hyperspectral cubes in numpy array for testing
  • maize_hyperspectral_cube.npy (495*320*243)
  • sorghum_hyperspectral_cube1.npy (561*320*243)
  • sorghum_hyperspectral_cube2.npy (560*320*243)

phenotypes

• traits extracted from segmented sorghum images
  • pheno_height.csv (phenotypes for plant height)
  • pheno_panicle.csv (phenotypes for panicle size)
  • pheno_stalk.csv (phenotypes for stalk size)
  • pheno_leaf.csv (phenotpes for leaf size)
  • pheno_panicleleaf.csv (phenotypes of the ratio of panicle and leaf size)
  • pheno_stalkleaf.csv (phenotypes of the ratio of stalk and leaf size)
  • pheno_paniclestalk.csv (phenotypes for the ratio of panicle stalk size)

pixel_data

• sorghum pixel classification dataset for background, leaf, stalk, and panicle

  • sorghum_features.csv (7560*243)
  • sorghum_labels.csv (7560*1)

• maize pixel classification dataset for background, leaf, stalk, and tassel

  • maize_features.csv (4000*243)
  • maize_labels.csv (4000*1)