Skip to content

Rice Detect

Import package

import rasterio
from Godream.plotimg import overlay_map
from Godream.model import riceByOptical, riceBySar

For Optical image

Your satellite image bands have to set the order of bands like this:

band_1 as red band

band_2 as green band

band_3 as blue band

band_4 as nir band

#set input
trainset = 'data/trainset_DN.geojson'
raster_img = "data/S2_image3.tif"
output_tiff = 'classified_S2.tiff'

# visualize input
filev = [trainset]
filer = [raster_img]
overlay_map( filev, filer, with_draw_tools=True,zoom=None )

123

Optical rice model

Raster images should have 4 bands that compose of red, green, blue and nir bands respectively.

riceByOptical(raster_img, output_tiff, figsize = (12, 7), num_estimator = 100)

123

For SAR image

# set input parameter
raster_stack = "D:\DGEO\myDATA\VH_stack.tif" # SAR_VH stack 19 images
output_tiff = 'classified.tiff'

SAR rice model

This model uses time series data of SAR VH polarization to run the model. Thus, input data should be stack 19 images of SAR VH polarization images.

riceBySar(raster_stack, output_tiff, figsize = (12, 7), num_estimator = 100 )

123