当前位置:网站首页>YOLO物体识别,生成数据用到的工具
YOLO物体识别,生成数据用到的工具
2022-07-02 06:34:00 【懵懂的梦花火】
YOLO物体识别,生成数据用到的工具
1 remove_spaces(去掉文件名称中的空格,将“ ”替换为“_”)
运行前:
import os
import string
def listdir(path, list_name):
for file in os.listdir(path):
file_path = os.path.join(path, file)
if os.path.isdir(file_path):
listdir(file_path, list_name)
elif os.path.splitext(file_path)[1] == '.xml' or os.path.splitext(file_path)[1] == '.jpg':
list_name.append(file_path)
if (1):
# 1. 读取xml文件
file_path = "C:/31/darknet-master/build/darknet/x64/data/voc/VOC2020/JPEGImages"
list_name = []
listdir(file_path, list_name)
# print(list_name)
size = 0
str_ = ""
for file_name in list_name:
# print(tree)
# 2. 属性修改
# A. 找到父节点
str_ = ((file_name).replace(' ','_'))
os.rename(file_name, str_)
size+=1
print( file_name + " " + str(size) +' succeful! ')
运行后:
2 change_xml (修改xml文件中出现的中文和路径错误)
生成数据集时出现如下错误:
运行前:
from xml.etree.ElementTree import ElementTree, Element
def read_xml(in_path):
'''''读取并解析xml文件
in_path: xml路径
return: ElementTree'''
tree = ElementTree()
tree.parse(in_path)
return tree
def write_xml(tree, out_path):
'''''将xml文件写出
tree: xml树
out_path: 写出路径'''
tree.write(out_path, encoding="utf-8", xml_declaration=True)
def if_match(node, kv_map):
'''''判断某个节点是否包含所有传入参数属性
node: 节点
kv_map: 属性及属性值组成的map'''
for key in kv_map:
if node.get(key) != kv_map.get(key):
return False
return True
# ---------------search -----
def find_nodes(tree, path):
'''''查找某个路径匹配的所有节点
tree: xml树
path: 节点路径'''
return tree.findall(path)
def get_node_by_keyvalue(nodelist, kv_map):
'''''根据属性及属性值定位符合的节点,返回节点
nodelist: 节点列表
kv_map: 匹配属性及属性值map'''
result_nodes = []
for node in nodelist:
if if_match(node, kv_map):
result_nodes.append(node)
return result_nodes
# ---------------change -----
def change_node_properties(nodelist, kv_map, is_delete=False):
'''''修改/增加 /删除 节点的属性及属性值
nodelist: 节点列表
kv_map:属性及属性值map'''
for node in nodelist:
for key in kv_map:
if is_delete:
if key in node.attrib:
del node.attrib[key]
else:
node.set(key, kv_map.get(key))
def change_node_text(nodelist, text, is_add=False, is_delete=False):
'''''改变/增加/删除一个节点的文本
nodelist:节点列表
text : 更新后的文本'''
for node in nodelist:
if is_add:
node.text += text
elif is_delete:
node.text = ""
else:
node.text = text
def create_node(tag, property_map, content):
'''''新造一个节点
tag:节点标签
property_map:属性及属性值map
content: 节点闭合标签里的文本内容
return 新节点'''
element = Element(tag, property_map)
element.text = content
return element
def add_child_node(nodelist, element):
'''''给一个节点添加子节点
nodelist: 节点列表
element: 子节点'''
for node in nodelist:
node.append(element)
def del_node_by_tagkeyvalue(nodelist, tag, kv_map):
'''''同过属性及属性值定位一个节点,并删除之
nodelist: 父节点列表
tag:子节点标签
kv_map: 属性及属性值列表'''
for parent_node in nodelist:
children = parent_node.getchildren()
for child in children:
if child.tag == tag and if_match(child, kv_map):
parent_node.remove(child)
import os
import string
def listdir(path, list_name):
for file in os.listdir(path):
file_path = os.path.join(path, file)
if os.path.isdir(file_path):
listdir(file_path, list_name)
elif os.path.splitext(file_path)[1] == '.xml':
list_name.append(file_path)
if (1):
# 1. 读取xml文件
file_path = "C:/Users/zhumengbo/Desktop/new/original"
list_name = []
listdir(file_path, list_name)
# print(list_name)
size = 0
str_ = ""
for strPath in list_name:
tree = read_xml(strPath)
# print(tree)
# 2. 属性修改
# A. 找到父节点
nodes = find_nodes(tree, "path")
str_ = ((strPath).replace('\\','/'))
change_node_text(nodes, str_)
write_xml(tree, str_)
size+=1
print( str_ + " " + str(size) +' succeful! ')
修改文件名:将 新建文件夹 修改 为fruit_plates,程序自动根据当前新的文件夹名称来修改原本的路径
运行后:
3 去掉多余jpg和xml文件
在数据集中,jpg文件和xml需要成对出现
import os
import string
def listdir(path, list_name, format):
for file in os.listdir(path):
file_path = os.path.join(path, file)
if os.path.isdir(file_path):
listdir(file_path, list_name)
elif os.path.splitext(file_path)[1] == format:
list_name.append(os.path.splitext(file_path)[0].split('\\')[-1])
def aligning(path):
list_xml_name = []
list_jpg_name = []
listdir(path, list_xml_name, '.xml')
listdir(path, list_jpg_name, '.jpg')
# 删除多余的xml文件
for file_name_xml in list_xml_name:
is_exit = 0
for file_name_jpg in list_jpg_name:
if file_name_xml == file_name_jpg:
is_exit = 1
break
if is_exit == 0:
current_file_path = path + '/' + file_name_xml + '.xml'
if os.path.exists(current_file_path): # 如果文件存在
# 删除文件,可使用以下两种方法。
os.remove(current_file_path)
# 删除多余的jpg文件
for file_name_jpg in list_jpg_name:
is_exit = 0
for file_name_xml in list_xml_name:
if file_name_jpg == file_name_xml:
is_exit = 1
break
if is_exit == 0:
current_file_path = path + '/' + file_name_jpg + '.xml'
if os.path.exists(current_file_path): # 如果文件存在
# 删除文件,可使用以下两种方法。
os.remove(current_file_path)
if (1):
# 1. 读取xml文件
aligning('C:/Users//Desktop/img/label20201009/label/yellow_belt/yellow_belt2_xml')
边栏推荐
- Double non undergraduate students enter the factory, while I am still quietly climbing trees at the bottom (Part 1)
- 道阻且长,行则将至
- 每天睡觉前30分钟阅读_day4_Files
- [go practical basis] how to bind and use URL parameters in gin
- 大学生四六级作文模板(自创版,成功跨过六级)
- idea查看字节码配置
- 图像识别-数据增广
- 深入剖析JVM是如何执行Hello World的
- 双非本科生进大厂,而我还在底层默默地爬树(上)
- Chrome用户脚本管理器-Tampermonkey 油猴
猜你喜欢
[go practical basis] how to bind and use URL parameters in gin
Number structure (C language -- code with comments) -- Chapter 2, linear table (updated version)
微服务实战|手把手教你开发负载均衡组件
记录一下初次使用Xray的有趣过程
Don't look for it. All the necessary plug-ins for Chrome browser are here
Matplotlib剑客行——布局指南与多图实现(更新)
Pool de connexion redis personnalisé
Machine learning practice: is Mermaid a love movie or an action movie? KNN announces the answer
微服务实战|熔断器Hystrix初体验
[go practical basis] how to set the route in gin
随机推荐
Chrome browser plug-in fatkun installation and introduction
Typeerror: X () got multiple values for argument 'y‘
Who is better for Beijing software development? How to find someone to develop system software
BugkuCTF-web21(详细解题思路及步骤)
zk配置中心---Config Toolkit配置与使用
MySQL multi column in operation
Read Day5 30 minutes before going to bed every day_ All key values in the map, how to obtain all value values
每天睡觉前30分钟阅读_day3_Files
Pdf document of distributed service architecture: principle + Design + practice, (collect and see again)
AMQ 4043 solution for errors when using IBM MQ remote connection
C语言之二进制与十进制
破茧|一文说透什么是真正的云原生
图像识别-数据标注
C语言之分草莓
Oracle modifies tablespace names and data files
Knife4j 2. Solution to the problem of file control without selection when uploading x version files
Solution to amq4036 error in remote connection to IBM MQ
洞见云原生|微服务及微服务架构浅析
Probability is not yet. Look at statistical learning methods -- Chapter 4, naive Bayesian method
别找了,Chrome浏览器必装插件都在这了