search

inkyusa / weedNet

weedNet: Dense semantic weed classification using multispectral images and MAV for smart farming
GNU General Public License v3.0
53 stars 18 forks source link

Data used in the paper #3

Open daobilige-su opened 5 years ago

daobilige-su commented 5 years ago

Hi inkyusa,

Thanks for sharing the data and the code.

After I read the paper, if I understand correctly, 132 crop only images and 243 weed only images are used to train your model and 90 crop weed mixture images are used to test your model, as can be seen from TABEL I of your paper.

However, I can only find 90 train images and 85 test images in the repo. Am I missing something here? or the data used in you paper is not here?

Cheers, Su

seuriously commented 2 years ago

Hi, I've also checked the 4ch dataset. I can't seem to find the blue band for RGB purpose. I assumed suffix:

image