一. 数据类型转换
1.1 图片格式转换
bmp转jpg:
import os.path
from traceback import print_tb
from PIL import Image
from io import BytesIO
def scanDir(dirPath: str)->list:
"""扫描目录"""
fileList = []
for top, dirs, nondirs in os.walk(dirPath):
for item in nondirs:
if item.split('.')[-1] == 'bmp':
fileList.append(os.path.join(top, item))
return fileList
def bmp2jpg(filePath):
fileLst = scanDir(filePath)
# print(fileLst)
for oneFile in fileLst:
bmp = Image.open(oneFile)
output_buffer = BytesIO()
bmp.save(output_buffer,format ='JPEG')
byte_data = output_buffer.getvalue()
a = os.path.splitext(oneFile)[0]
# a= a.replace('dataset_1.11_jpg', 'dataset_1.11_jpg_jpg')
tmpUrl = f'{
a}.jpg'
with open(tmpUrl, 'wb') as f:
f.write(byte_data)
print(f"Write {
tmpUrl}!")
os.remove(tmpUrl.replace("jpg", "bmp"))
bmp2jpg("/media/geek/182CFAE12CFAB8B8/Dataset/go")
1.2 图片大小转换
import cv2
import os
path = "/home/geek/Project/cloth_check/Detection/dataset/JPEGImages/"
path_output = "/home/geek/Project/cloth_check/Detection/dataset/JPEGImages_5/"
filelist = os.listdir(path)
for filename in filelist:
filepath = path + filename
img = cv2.imread(filepath)
# image = cv2.resize(img, (int(img.shape[0]*0.3), int(img.shape[1]*0.3)))
image = cv2.resize(img, (640, 640))
print(path_output+filename)
cv2.imwrite(path_output+filename, image)
二. 数据标注格式转换
2.1 VOC转COCO格式
"""
将voc格式转json格式用于caffe2的detectron的训练
在detectron中voc_2007_train.json和voc_2007_val.json中categories的顺序必须保持一致
因此最好事先确定category的顺序，书写在category_set中
"""
import xml.etree.ElementTree as ET
import os
import json
import collections
coco = dict()
coco['images'] = []
coco['type'] = 'instances'
coco['annotations'] = []
coco['categories'] = []
#category_set = dict()
image_set = set()
image_id = 2019000001 #train:2018xxx; val:2019xxx; test:2020xxx
category_item_id = 1
annotation_id = 1
category_set = ['flaw']
'''
def addCatItem(name):
global category_item_id
category_item = dict()
category_item['supercategory'] = 'none'
category_item_id += 1
category_item['id'] = category_item_id
category_item['name'] = name
coco['categories'].append(category_item)
category_set[name] = category_item_id
return category_item_id
'''
def addCatItem(name):
'''
增加json格式中的categories部分
'''
global category_item_id
category_item = collections.OrderedDict()
category_item['supercategory'] = 'none'
category_item['id'] = category_item_id
category_item['name'] = name
coco['categories'].append(category_item)
category_item_id += 1
def addImgItem(file_name, size):
global image_id
if file_name is None:
raise Exception('Could not find filename tag in xml file.')
if size['width'] is None:
raise Exception('Could not find width tag in xml file.')
if size['height'] is None:
raise Exception('Could not find height tag in xml file.')
#image_item = dict() #按照一定的顺序，这里采用collections.OrderedDict()
image_item = collections.OrderedDict()
jpg_name = os.path.splitext(file_name)[0]+os.path.splitext(file_name)[1]
image_item['file_name'] = jpg_name
image_item['width'] = size['width']
image_item['height'] = size['height']
image_item['id'] = image_id
coco['images'].append(image_item)
image_set.add(jpg_name)
image_id = image_id+1
return image_id
def addAnnoItem(object_name, image_id, category_id, bbox):
global annotation_id
#annotation_item = dict()
annotation_item = collections.OrderedDict()
annotation_item['segmentation'] = []
seg = []
# bbox[] is x,y,w,h
# left_top
seg.append(bbox[0])
seg.append(bbox[1])
# left_bottom
seg.append(bbox[0])
seg.append(bbox[1] + bbox[3])
# right_bottom
seg.append(bbox[0] + bbox[2])
seg.append(bbox[1] + bbox[3])
# right_top
seg.append(bbox[0] + bbox[2])
seg.append(bbox[1])
annotation_item['segmentation'].append(seg)
annotation_item['area'] = bbox[2] * bbox[3]
annotation_item['iscrowd'] = 0
annotation_item['image_id'] = image_id
annotation_item['bbox'] = bbox
annotation_item['category_id'] = category_id
annotation_item['id'] = annotation_id
annotation_item['ignore'] = 0
annotation_id += 1
coco['annotations'].append(annotation_item)
def parseXmlFiles(xml_path):
xmllist = os.listdir(xml_path)
xmllist.sort()
for f in xmllist:
if not f.endswith('.xml'):
continue
bndbox = dict()
size = dict()
current_image_id = None
current_category_id = None
file_name = None
size['width'] = None
size['height'] = None
size['depth'] = None
xml_file = os.path.join(xml_path, f)
print(xml_file)
tree = ET.parse(xml_file)
root = tree.getroot() #抓根结点元素
if root.tag != 'annotation': #根节点标签
raise Exception('pascal voc xml root element should be annotation, rather than {}'.format(root.tag))
# elem is <folder>, <filename>, <size>, <object>
for elem in root:
current_parent = elem.tag
current_sub = None
object_name = None
#elem.tag, elem.attrib，elem.text
if elem.tag == 'folder':
continue
if elem.tag == 'filename':
file_name = elem.text
if file_name in category_set:
raise Exception('file_name duplicated')
# add img item only after parse <size> tag
elif current_image_id is None and file_name is not None and size['width'] is not None:
if file_name not in image_set:
current_image_id = addImgItem(file_name, size)#图片信息
print('add image with {} and {}'.format(file_name, size))
else:
raise Exception('duplicated image: {}'.format(file_name))
# subelem is <width>, <height>, <depth>, <name>, <bndbox>
for subelem in elem:
bndbox['xmin'] = None
bndbox['xmax'] = None
bndbox['ymin'] = None
bndbox['ymax'] = None
current_sub = subelem.tag
if current_parent == 'object' and subelem.tag == 'name':
object_name = subelem.text
#if object_name not in category_set:
# current_category_id = addCatItem(object_name)
#else:
#current_category_id = category_set[object_name]
# print(category_set)
current_category_id = category_set.index(object_name)+1 #index默认从0开始,但是json文件是从1开始，所以+1
elif current_parent == 'size':
if size[subelem.tag] is not None:
raise Exception('xml structure broken at size tag.')
size[subelem.tag] = int(subelem.text)
# option is <xmin>, <ymin>, <xmax>, <ymax>, when subelem is <bndbox>
for option in subelem:
if current_sub == 'bndbox':
if bndbox[option.tag] is not None:
raise Exception('xml structure corrupted at bndbox tag.')
bndbox[option.tag] = int(option.text)
# only after parse the <object> tag
if bndbox['xmin'] is not None:
if object_name is None:
raise Exception('xml structure broken at bndbox tag')
if current_image_id is None:
raise Exception('xml structure broken at bndbox tag')
if current_category_id is None:
raise Exception('xml structure broken at bndbox tag')
bbox = []
# x
bbox.append(bndbox['xmin'])
# y
bbox.append(bndbox['ymin'])
# w
bbox.append(bndbox['xmax'] - bndbox['xmin'])
# h
bbox.append(bndbox['ymax'] - bndbox['ymin'])
print(
'add annotation with {},{},{},{}'.format(object_name, current_image_id-1, current_category_id, bbox))
addAnnoItem(object_name, current_image_id-1, current_category_id, bbox)
#categories部分
for categoryname in category_set:
addCatItem(categoryname)
if __name__ == '__main__':
xml_path = '/media/geek/182CFAE12CFAB8B8/Dataset/Test/Annotations'
json_file = '/media/geek/182CFAE12CFAB8B8/Dataset/Test/annotations.json'
parseXmlFiles(xml_path)
json.dump(coco, open(json_file, 'w'))
三. 数据清洗
3.1 VOC标注文件清洗
经常会出现Annotations文件夹中的xml数量小于JPEGImages文件夹中对应原图的数量,因此对JPEGImages中的多余图片进行清除:
import os
import shutil
image_path = "/media/geek/182CFAE12CFAB8B8/Dataset/Test/JPEGImages/"
Annota_path = "/media/geek/182CFAE12CFAB8B8/Dataset/Test/Annotations/"
Annota_list = os.listdir(Annota_path)
image_list = os.listdir(image_path)
for image_name in image_list:
fg = 0
for xml_name in Annota_list:
if xml_name[:-4]==image_name[:-4]:
fg = 1
if fg==0:
os.remove(image_path + image_name)
# print(num)