import os
import numpy as np
import cv2, uuid
import PIL.Image as Image
import matplotlib.pyplot as plt
import xml.etree.ElementTree as ET
import random
def bboxes2xml(folder, img_name, width, height, gts, xml_save_to):
xml_file = open((xml_save_to + '/' + img_name + '.xml'), 'w')
xml_file.write('<annotation>\n')
xml_file.write(' <folder>' + folder + '</folder>\n')
xml_file.write(' <filename>' + str(img_name) + '.jpg' + '</filename>\n')
xml_file.write(' <size>\n')
xml_file.write(' <width>' + str(width) + '</width>\n')
xml_file.write(' <height>' + str(height) + '</height>\n')
xml_file.write(' <depth>3</depth>\n')
xml_file.write(' </size>\n')
for gt in gts:
xml_file.write(' <object>\n')
xml_file.write(' <name>' + str(gt[0]) + '</name>\n')
xml_file.write(' <pose>Unspecified</pose>\n')
xml_file.write(' <truncated>0</truncated>\n')
xml_file.write(' <difficult>0</difficult>\n')
xml_file.write(' <bndbox>\n')
xml_file.write(' <xmin>' + str(gt[1]) + '</xmin>\n')
xml_file.write(' <ymin>' + str(gt[2]) + '</ymin>\n')
xml_file.write(' <xmax>' + str(gt[3]) + '</xmax>\n')
xml_file.write(' <ymax>' + str(gt[4]) + '</ymax>\n')
xml_file.write(' </bndbox>\n')
xml_file.write(' </object>\n')
xml_file.write('</annotation>')
xml_file.close()
def get_new_img_xy(infile, image_size):
"""返回一个图片的宽、高像素"""
im = Image.open(infile)
(x, y) = im.size
lv = round(x / image_size, 2) + 0.01
x_s = x // lv
y_s = y // lv
# print("x_s", x_s, y_s)
# out = im.resize((x_s, y_s), Image.ANTIALIAS)
return x_s, y_s # 返回缩放到image_size的大小
def resize_by_width(infile, image_size):
"""按照宽度进行所需比例缩放"""
im = Image.open(infile)
(x, y) = im.size
lv = round(x / image_size, 2) + 0.01
x_s = int(x // lv)
y_s = int(y // lv)
# print("x_s", x_s, y_s)
out = im.resize((x_s, y_s), Image.ANTIALIAS)
# plt.imshow(out)
# plt.show()
return out, lv
def get_bboxes(xml_path):
tree = ET.parse(open(xml_path, 'rb'))
root = tree.getroot()
bboxes, cls = [], []
for obj in root.iter('object'):
obj_cls = obj.find('name').text
xmlbox = obj.find('bndbox')
xmin = float(xmlbox.find('xmin').text)
xmax = float(xmlbox.find('xmax').text)
ymin = float(xmlbox.find('ymin').text)
ymax = float(xmlbox.find('ymax').text)
bboxes.append([xmin, ymin, xmax, ymax])
cls.append(obj_cls)
bboxes = np.asarray(bboxes, np.int)
return bboxes, cls
# 得到图像缩放后，xml中对应坐标的位置
def get_new_point(bboxes1, lv, d_x, d_y):
get_list=[]
for points in bboxes1:
temp_list=[]
print('------------', points)
xmin, ymin, xmax, ymax = points
xmin = int(xmin // lv)+d_x
ymin = int(ymin // lv)+d_y
xmax = int(xmax // lv)+d_x
ymax = int(ymax // lv)+d_y
temp_list.append(xmin)
temp_list.append(ymin)
temp_list.append(xmax)
temp_list.append(ymax)
get_list.append(temp_list)
return get_list
# 定义图像拼接函数
def image_compose(image_colnum, image_size, image_rownum, image_names, image_save_path, x_new, y_new, xml_path,new_xml_path):
to_image = Image.new('RGB', (image_colnum * x_new, image_rownum * y_new)) # 创建一个新图
# plt.imshow(to_image)
# plt.show()
# 循环遍历，把每张图片按顺序粘贴到对应位置上
total_num = 0
bbox_list=[]
cls=[]
for y in range(1, image_rownum + 1):
for x in range(1, image_colnum + 1):
from_image, lv = resize_by_width(image_names[image_colnum * (y - 1) + x - 1], image_size)
# from_image = Image.open(image_names[image_colnum * (y - 1) + x - 1]).resize((image_size,image_size ), Image.ANTIALIAS)
to_image.paste(from_image, ((x - 1) * x_new, (y - 1) * y_new))
# plt.imshow(to_image)
# plt.show()
# 处理对应的xml文件
xml_name = image_names[image_colnum * (y - 1) + x - 1].split('\\')[-1]
xml_path_ = os.path.join(xml_path, xml_name.replace('.jpg', '.xml')) # 拼接后得到对应的xml路径
bboxes1, cls1 = get_bboxes(xml_path_)
print(bboxes1, cls1)
d_x = (x - 1) * x_new # 间隔距离
d_y = (y - 1) * y_new # 间隔距离
if len(bboxes1)!=0:
point = get_new_point(bboxes1, lv, d_x, d_y)
bbox_list.append(point)
cls = cls + cls1
total_num += 1
if total_num == len(image_names):
break
gts=[]
if len(bbox_list)!=0:
bboxes = np.vstack(tuple(bbox_list))
gts = [[c] + b.tolist() for c, b in zip(cls, bboxes)]
else:
pass
bboxes2xml(folder=image_save_path,
img_name=image_save_path.split('\\')[-1].replace('.jpg',''),
width=image_colnum * x_new, height=image_rownum * y_new,
gts=gts, xml_save_to=new_xml_path)
return to_image.save(image_save_path) # 保存新图
def merge_images(image_dir_path, image_size, image_colnum, new_images_dir, xml_path,new_xml_path):
# 获取图片集地址下的所有图片名称
image_fullpath_list = image_dir_path # 获取得到当前路径下所有的图片，得到列表
print("image_fullpath_list", len(image_fullpath_list), image_fullpath_list)
uuid_path = uuid.uuid4().hex
new_img_path = os.path.join(new_images_dir, uuid_path)
image_save_path = r'{}.jpg'.format(new_img_path) # 图片转换后的地址
print(image_save_path)
# image_rownum = 4 # 图片间隔，也就是合并成一张图后，一共有几行
image_rownum_yu = len(image_fullpath_list) % image_colnum
if image_rownum_yu == 0:
image_rownum = len(image_fullpath_list) // image_colnum
else:
image_rownum = len(image_fullpath_list) // image_colnum + 1
x_list = []
y_list = []
for img_file in image_fullpath_list:
img_x, img_y = get_new_img_xy(img_file, image_size)
x_list.append(img_x)
y_list.append(img_y)
# print("x_list", sorted(x_list))
# print("y_list", sorted(y_list))
x_new = int(x_list[len(x_list) // 5 * 4])
y_new = int(y_list[len(y_list) // 5 * 4])
# print(" x_new, y_new", x_new, y_new)
image_compose(image_colnum, image_size, image_rownum, image_fullpath_list, image_save_path, x_new, y_new,
xml_path,new_xml_path) # 调用函数
# for img_file in image_fullpath_list:
# resize_by_width(img_file,image_size)
if __name__ == '__main__':
image_size = 320 # 每张小图片的大小
num_in_img = 4 # 拼接的图片数量
image_colnum = int((num_in_img) / 2) # 合并成一张图后，一行有几个小图
image_dir_path = r'C:\Users\GuoQingru\Desktop\printer1k\images' # 图片集地址
xml_path = r'C:\Users\GuoQingru\Desktop\printer1k\xml' # 原先的xml目录
new_images_dir = r'C:\Users\GuoQingru\Desktop\printer1k\new_images' # 合成图像的保存目录
new_xml_path=r'C:\Users\GuoQingru\Desktop\printer1k\new_xml' # 生成的xml保存位置
listdir = os.listdir(image_dir_path)
print('^^^^^^^^^^^^^^^^',listdir)
random.shuffle(listdir)
print('^^^^^^^^^^^^^^^^', listdir)
img_sum = []
while listdir:
tem_sum = []
for index in range(num_in_img):
pop = listdir.pop()
file_path = os.path.join(image_dir_path, pop)
tem_sum.append(file_path)
if len(listdir) < num_in_img:
break
img_sum.append(tem_sum)
if len(listdir) < num_in_img:
break
# print(img_sum)
# merge_images(image_dir_path, image_size, image_colnum)
for images in img_sum:
if len(images) != num_in_img:
continue
merge_images(images, image_size, image_colnum, new_images_dir, xml_path,new_xml_path)
import os
import numpy as np
import cv2, uuid
import PIL.Image as Image
import matplotlib.pyplot as plt
import xml.etree.ElementTree as ET
import random
def bboxes2xml(folder, img_name, width, height, gts, xml_save_to):
xml_file = open((xml_save_to + '/' + img_name + '.xml'), 'w')
xml_file.write('<annotation>\n')
xml_file.write(' <folder>' + folder + '</folder>\n')
xml_file.write(' <filename>' + str(img_name) + '.jpg' + '</filename>\n')
xml_file.write(' <size>\n')
xml_file.write(' <width>' + str(width) + '</width>\n')
xml_file.write(' <height>' + str(height) + '</height>\n')
xml_file.write(' <depth>3</depth>\n')
xml_file.write(' </size>\n')
for gt in gts:
xml_file.write(' <object>\n')
xml_file.write(' <name>' + str(gt[0]) + '</name>\n')
xml_file.write(' <pose>Unspecified</pose>\n')
xml_file.write(' <truncated>0</truncated>\n')
xml_file.write(' <difficult>0</difficult>\n')
xml_file.write(' <bndbox>\n')
xml_file.write(' <xmin>' + str(gt[1]) + '</xmin>\n')
xml_file.write(' <ymin>' + str(gt[2]) + '</ymin>\n')
xml_file.write(' <xmax>' + str(gt[3]) + '</xmax>\n')
xml_file.write(' <ymax>' + str(gt[4]) + '</ymax>\n')
xml_file.write(' </bndbox>\n')
xml_file.write(' </object>\n')
xml_file.write('</annotation>')
xml_file.close()
def get_new_img_xy(infile, image_size):
"""返回一个图片的宽、高像素"""
im = Image.open(infile)
(x, y) = im.size
lv = round(x / image_size, 2) + 0.01
x_s = x // lv
y_s = y // lv
# print("x_s", x_s, y_s)
# out = im.resize((x_s, y_s), Image.ANTIALIAS)
return x_s, y_s # 返回缩放到image_size的大小
def resize_by_width(infile, image_size):
"""按照宽度进行所需比例缩放"""
im = Image.open(infile)
(x, y) = im.size
lv = round(x / image_size, 2) + 0.01
x_s = int(x // lv)
y_s = int(y // lv)
# print("x_s", x_s, y_s)
out = im.resize((x_s, y_s), Image.ANTIALIAS)
# plt.imshow(out)
# plt.show()
return out, lv
def get_bboxes(xml_path):
tree = ET.parse(open(xml_path, 'rb'))
root = tree.getroot()
bboxes, cls = [], []
for obj in root.iter('object'):
obj_cls = obj.find('name').text
xmlbox = obj.find('bndbox')
xmin = float(xmlbox.find('xmin').text)
xmax = float(xmlbox.find('xmax').text)
ymin = float(xmlbox.find('ymin').text)
ymax = float(xmlbox.find('ymax').text)
bboxes.append([xmin, ymin, xmax, ymax])
cls.append(obj_cls)
bboxes = np.asarray(bboxes, np.int)
return bboxes, cls
# 得到图像缩放后，xml中对应坐标的位置
def get_new_point(bboxes1, lv, d_x, d_y):
get_list=[]
for points in bboxes1:
temp_list=[]
print('------------', points)
xmin, ymin, xmax, ymax = points
xmin = int(xmin // lv)+d_x
ymin = int(ymin // lv)+d_y
xmax = int(xmax // lv)+d_x
ymax = int(ymax // lv)+d_y
temp_list.append(xmin)
temp_list.append(ymin)
temp_list.append(xmax)
temp_list.append(ymax)
get_list.append(temp_list)
return get_list
# 定义图像拼接函数
def image_compose(image_colnum, image_size, image_rownum, image_names, image_save_path, x_new, y_new, xml_path,new_xml_path):
to_image = Image.new('RGB', (image_colnum * x_new, image_rownum * y_new)) # 创建一个新图
# plt.imshow(to_image)
# plt.show()
# 循环遍历，把每张图片按顺序粘贴到对应位置上
total_num = 0
bbox_list=[]
cls=[]
for y in range(1, image_rownum + 1):
for x in range(1, image_colnum + 1):
from_image, lv = resize_by_width(image_names[image_colnum * (y - 1) + x - 1], image_size)
# from_image = Image.open(image_names[image_colnum * (y - 1) + x - 1]).resize((image_size,image_size ), Image.ANTIALIAS)
to_image.paste(from_image, ((x - 1) * x_new, (y - 1) * y_new))
# plt.imshow(to_image)
# plt.show()
# 处理对应的xml文件
xml_name = image_names[image_colnum * (y - 1) + x - 1].split('\\')[-1]
xml_path_ = os.path.join(xml_path, xml_name.replace('.jpg', '.xml')) # 拼接后得到对应的xml路径
bboxes1, cls1 = get_bboxes(xml_path_)
print(bboxes1, cls1)
d_x = (x - 1) * x_new # 间隔距离
d_y = (y - 1) * y_new # 间隔距离
if len(bboxes1)!=0:
point = get_new_point(bboxes1, lv, d_x, d_y)
bbox_list.append(point)
cls = cls + cls1
total_num += 1
if total_num == len(image_names):
break
gts=[]
if len(bbox_list)!=0:
bboxes = np.vstack(tuple(bbox_list))
gts = [[c] + b.tolist() for c, b in zip(cls, bboxes)]
else:
pass
bboxes2xml(folder=image_save_path,
img_name=image_save_path.split('\\')[-1].replace('.jpg',''),
width=image_colnum * x_new, height=image_rownum * y_new,
gts=gts, xml_save_to=new_xml_path)
return to_image.save(image_save_path) # 保存新图
def merge_images(image_dir_path, image_size, image_colnum, new_images_dir, xml_path,new_xml_path):
# 获取图片集地址下的所有图片名称
image_fullpath_list = image_dir_path # 获取得到当前路径下所有的图片，得到列表
print("image_fullpath_list", len(image_fullpath_list), image_fullpath_list)
uuid_path = uuid.uuid4().hex
new_img_path = os.path.join(new_images_dir, uuid_path)
image_save_path = r'{}.jpg'.format(new_img_path) # 图片转换后的地址
print(image_save_path)
# image_rownum = 4 # 图片间隔，也就是合并成一张图后，一共有几行
image_rownum_yu = len(image_fullpath_list) % image_colnum
if image_rownum_yu == 0:
image_rownum = len(image_fullpath_list) // image_colnum
else:
image_rownum = len(image_fullpath_list) // image_colnum + 1
x_list = []
y_list = []
for img_file in image_fullpath_list:
img_x, img_y = get_new_img_xy(img_file, image_size)
x_list.append(img_x)
y_list.append(img_y)
# print("x_list", sorted(x_list))
# print("y_list", sorted(y_list))
x_new = int(x_list[len(x_list) // 5 * 4])
y_new = int(y_list[len(y_list) // 5 * 4])
# print(" x_new, y_new", x_new, y_new)
image_compose(image_colnum, image_size, image_rownum, image_fullpath_list, image_save_path, x_new, y_new,
xml_path,new_xml_path) # 调用函数
# for img_file in image_fullpath_list:
# resize_by_width(img_file,image_size)
if __name__ == '__main__':
image_size = 320 # 每张小图片的大小
num_in_img = 9 # 拼接的图片数量
image_colnum = int((num_in_img) / 3) # 合并成一张图后，一行有几个小图
image_dir_path = r'C:\Users\GuoQingru\Desktop\printer1k\images' # 图片集地址
xml_path = r'C:\Users\GuoQingru\Desktop\printer1k\xml' # 原先的xml目录
new_images_dir = r'C:\Users\GuoQingru\Desktop\printer1k\new_images' # 合成图像的保存目录
new_xml_path=r'C:\Users\GuoQingru\Desktop\printer1k\new_xml' # 生成的xml保存位置
listdir = os.listdir(image_dir_path)
print('^^^^^^^^^^^^^^^^',listdir)
random.shuffle(listdir)
print('^^^^^^^^^^^^^^^^', listdir)
img_sum = []
while listdir:
tem_sum = []
for index in range(num_in_img):
pop = listdir.pop()
file_path = os.path.join(image_dir_path, pop)
tem_sum.append(file_path)
if len(listdir) < num_in_img:
break
img_sum.append(tem_sum)
if len(listdir) < num_in_img:
break
# print(img_sum)
# merge_images(image_dir_path, image_size, image_colnum)
for images in img_sum:
if len(images) != num_in_img:
continue
merge_images(images, image_size, image_colnum, new_images_dir, xml_path,new_xml_path)
文章出处登录后可见！
已经登录？立即刷新
