公开数据集VOC中有20类,有时并不是所有的类都是我们所需要的,可以根据需要从中提取出特定类并重新制作成LMDB数据源。
第一步,从VOC数据集中提取特定的类,在VOC2007,VOC2012数据集下分别运行该代码即可提取出特定的类,注意修改代码中的地址,代码如下(Ubuntu环境):
# -*- coding: utf-8 -*-
# @Function:There are 20 classes in VOC data set. If you need to extract specific classes, you can use this program to extract them.
import os
import shutil
ann_filepath='/home/abc/data/VOCdevkit/VOC2012/Annotations/' #注意修改成自己的地址
img_filepath='/home/abc/data/VOCdevkit/VOC2012/JPEGImages/' #注意修改成自己的地址
img_savepath='/home/abc/data/VOCdevkit/VOC2012/JPEGImages_ssd/' #注意修改成自己的地址
ann_savepath='/home/abc/data/VOCdevkit/VOC2012/Annotations_ssd/' #注意修改成自己的地址
if not os.path.exists(img_savepath):
os.mkdir(img_savepath)
if not os.path.exists(ann_savepath):
os.mkdir(ann_savepath)
names = locals()
classes = ['aeroplane','bicycle','bird', 'boat', 'bottle',
'bus', 'car', 'cat', 'chair', 'cow','diningtable',
'dog', 'horse', 'motorbike', 'pottedplant',
'sheep', 'sofa', 'train', 'tvmonitor', 'person']
for file in os.listdir(ann_filepath):
print(file)
fp = open(ann_filepath + '//' + file)
ann_savefile=ann_savepath+file
fp_w = open(ann_savefile, 'w')
lines = fp.readlines()
ind_start = []
ind_end = []
lines_id_start = lines[:]
lines_id_end = lines[:]
classes1 = '\t\t<name>person</name>\n' #自己需要几类就写几类,我这只需要person这一类
#classes2 = '\t\t<name>motorbike</name>\n'
#classes3 = '\t\t<name>bus</name>\n'
#classes4 = '\t\t<name>car</name>\n'
#classes5 = '\t\t<name>bicycle</name>\n'
#在xml中找到object块,并将其记录下来
while "\t<object>\n" in lines_id_start:
a = lines_id_start.index("\t<object>\n")
ind_start.append(a)
lines_id_start[a] = "delete"
while "\t</object>\n" in lines_id_end:
b = lines_id_end.index("\t</object>\n")
ind_end.append(b)
lines_id_end[b] = "delete"
#names中存放所有的object块
i =