mirror of
https://github.com/azaion/ai-training.git
synced 2026-04-22 10:26:36 +00:00
Oleksandr Bezdieniezhnykh
78 lines
2.7 KiB
Python
78 lines
2.7 KiB
Python
import os
|
|
import shutil
|
|
import xml.etree.cElementTree as et
|
|
from pathlib import Path
|
|
|
|
labels_dir = 'labels'
|
|
images_dir = 'images'
|
|
|
|
tag_size = 'size'
|
|
tag_object = 'object'
|
|
tag_name = 'name'
|
|
tag_bndbox = 'bndbox'
|
|
name_class_map = {'Truck': 1, 'Car': 2, 'Taxi': 2} # 1 Вантажівка, 2 Машина легкова
|
|
forbidden_classes = ['Motorcycle']
|
|
default_class = 1
|
|
|
|
|
|
def convert_xml(folder):
|
|
os.makedirs(images_dir, exist_ok=True)
|
|
os.makedirs(labels_dir, exist_ok=True)
|
|
|
|
for f in os.listdir(folder):
|
|
if not f.endswith('.jpg'):
|
|
continue
|
|
label = f'{Path(f).stem}.xml'
|
|
lines = read_xml(folder, label)
|
|
if not lines:
|
|
print(f'Image {f} has only forbidden classes in annotations')
|
|
continue
|
|
|
|
shutil.copy(os.path.join(folder, f), os.path.join(images_dir, f))
|
|
with open(os.path.join(labels_dir, f'{Path(label).stem}.txt'), 'w') as label_file:
|
|
label_file.writelines(lines)
|
|
label_file.close()
|
|
print(f'Image {f} has been processed successfully')
|
|
|
|
|
|
def read_xml(folder, label):
|
|
tree = et.parse(os.path.join(folder, label))
|
|
root = tree.getroot()
|
|
lines = []
|
|
size_dict = {size_ch.tag: size_ch.text for size_ch in root.findall(f'{tag_size}/*')}
|
|
width = int(size_dict['width'])
|
|
height = int(size_dict['height'])
|
|
for node_object in tree.findall(tag_object):
|
|
class_num = default_class
|
|
c_x = c_y = c_w = c_h = 0
|
|
for node_object_ch in node_object:
|
|
if node_object_ch.tag == tag_name:
|
|
key = node_object_ch.text
|
|
if key in name_class_map:
|
|
class_num = name_class_map[key]
|
|
else:
|
|
if key in forbidden_classes:
|
|
class_num = -1
|
|
continue
|
|
else:
|
|
class_num = default_class
|
|
if node_object_ch.tag == tag_bndbox:
|
|
bbox_dict = {bbox_ch.tag: bbox_ch.text for bbox_ch in node_object_ch}
|
|
xmin = int(bbox_dict['xmin'])
|
|
xmax = int(bbox_dict['xmax'])
|
|
ymin = int(bbox_dict['ymin'])
|
|
ymax = int(bbox_dict['ymax'])
|
|
c_w = (xmax - xmin) / width
|
|
c_h = (ymax - ymin) / height
|
|
c_x = xmin / width + c_w / 2
|
|
c_y = ymin / height + c_h / 2
|
|
if class_num == -1:
|
|
continue
|
|
if c_x != 0 and c_y != 0 and c_w != 0 and c_h != 0:
|
|
lines.append(f'{class_num} {round(c_x, 5)} {round(c_y, 5)} {round(c_w, 5)} {round(c_h, 5)}\n')
|
|
return lines
|
|
|
|
|
|
if __name__ == '__main__':
|
|
convert_xml('datasets/others/UAVimages')
|