Python图片与label统计

# coding: utf-8

# In[1]:

import os
import pandas as pd
import shutil
import numpy as np

# In[2]:

root_path = 'F:\\template\\handobj-mark\\data\\img'

# In[3]:

os.chdir(root_path)

# In[4]:

files = os.listdir(root_path)

# In[5]:

data = pd.DataFrame(np.zeros(5000),columns = ['pngname'])
data1 = pd.DataFrame(np.zeros(5000),columns = ['txtname'])
i = 0
j = 0

# In[6]:

for filename in files:
    (name1,extension1) = os.path.splitext(filename)
    if extension1 == '.png':
        data['pngname'][i] = name1
        i = i+1
    else:
        data1['txtname'][j] = name1
        j = j+1

# In[7]:

datapng = data[data.pngname > 0]

# In[8]:

datatxt = data1[data1.txtname > 0]

# In[10]:

if datapng.shape[0] <= datatxt.shape[0]:
    data2 = pd.merge(left=datatxt, right=datapng, how='left', left_on='txtname', right_on='pngname')
    print('OK!')
else:
    data2 = pd.merge(left=datapng, right=datatxt, how='left', left_on='pngname', right_on='txtname')
    print('Error!')

# In[11]:

data2["bz"] = data2["txtname"] - data2["pngname"]
data2[data2.bz != 0]

# In[12]:

root1 = 'F:\\02_Porject_GR_Data\\01_Original_Data\\QD_Data_0726\\Left_hand\\shixisheng\\irMap\\irGrayMap'

# In[13]:

num1 = len(os.listdir(root1))

# In[14]:

originaldata = pd.DataFrame(np.ones(num1),columns = ['original'])

# In[15]:

for i in range(1,num1):
    originaldata["original"][i] = originaldata["original"][i] + originaldata["original"][i-1]

# In[16]:

result = pd.merge(left=originaldata, right=datapng, how='left', left_on='original', right_on='pngname')

# In[17]:

result["bz"] = result["original"] - result["pngname"]
resultdata = result[result.bz != 0]

# In[18]:

result1 = resultdata["original"].astype(np.int64).astype(np.str)

# In[19]:

result1 = pd.DataFrame(result1)

# In[20]:

result1["original"] = result1["original"] + '.png'

# In[23]:

result1.to_excel("F:\\删除的图片.xlsx",index = False)