sampling_image_6channels.py

import os

import numpy as np
import scipy.misc as misc
from libtiff import TIFF

base_dir_train = "../ISPRS_semantic_labeling_Vaihingen/train"
base_dir_validate = "../ISPRS_semantic_labeling_Vaihingen/validate"
base_dir_annotations = "../ISPRS_semantic_labeling_Vaihingen/annotations"
base_dir_top = "../ISPRS_semantic_labeling_Vaihingen/top"
base_dir_ndsm = "../ISPRS_semantic_labeling_Vaihingen/ndsm"
base_dir_dsm = "../ISPRS_semantic_labeling_Vaihingen/dsm"
base_dir_ndvi= "../ISPRS_semantic_labeling_Vaihingen/ndvi"
base_dir_train_validate_gt = "../ISPRS_semantic_labeling_Vaihingen/train_validate_gt"
image_size = 224
num_cropping_per_image = 3333
validate_image=[]

def create_training_dataset():
    for filename in os.listdir(base_dir_annotations):
        if filename in validate_image:
            continue
        top_image = misc.imread(os.path.join(base_dir_top,os.path.splitext(filename)[0]+".tif"))
        annotation_image = misc.imread(os.path.join(base_dir_annotations, filename))
        dsm_image_name= filename.replace('top_mosaic','dsm').replace('png','tif').replace('area','matching_area')

        # dsm_image= misc.imread(base_dir_dsm+"/"+dsm_image_name)

        dsm_image = TIFF.open(base_dir_dsm+"/"+dsm_image_name, 'r')
        dsm_image = dsm_image.read_image()

        ndsm_image_name= dsm_image_name.replace('.tif','')+"_normalized.jpg"
        ndsm_image= misc.imread(base_dir_ndsm+"/"+ndsm_image_name)
        ndvi_image_name = "ndvi"+ndsm_image_name.replace('dsm_09cm_matching_area','').replace('_normalized.jpg','.tif')
        ndvi_image = misc.imread(base_dir_ndvi + "/" + ndvi_image_name)

        width= np.shape(top_image)[1]
        height= np.shape(top_image)[0]
        for i in range(num_cropping_per_image):
            x = int(np.random.uniform(0, height - image_size + 1))
            y = int(np.random.uniform(0, width - image_size + 1))
            print((x,y))
            top_image_cropped= top_image[x:x + image_size, y:y + image_size, :]
            ndsm_image_cropped= ndsm_image[x:x + image_size, y:y + image_size]
            ndsm_image_cropped= np.expand_dims(ndsm_image_cropped,axis=2)
            dsm_image_cropped= dsm_image[x:x + image_size, y:y + image_size]
            dsm_image_cropped= np.expand_dims(dsm_image_cropped,axis=2)
            ndvi_image_cropped = ndvi_image[x:x + image_size, y:y + image_size]
            ndvi_image_cropped = np.expand_dims(ndvi_image_cropped, axis=2)
            array_to_save= np.concatenate((top_image_cropped, ndvi_image_cropped,ndsm_image_cropped,dsm_image_cropped),axis=2).astype(dtype=np.float16)
            np.save(os.path.join(base_dir_train, os.path.splitext(filename)[0] + "_" + str(i)+".npy"),array_to_save)
            annotation_image_cropped= annotation_image[x:x + image_size, y:y + image_size]
            misc.imsave(os.path.join(base_dir_train_validate_gt, os.path.splitext(filename)[0] + "_" + str(i) + ".png"), annotation_image_cropped)
    return None

if __name__=="__main__":
    create_training_dataset()