对⼩样本数据进⾏数据增强
⼀、前情介绍
在之前对yolov3的学习中，有时候发现⼩样本数据集容易出现过拟合或者泛化能⼒不强的问题，在对这⼀问题提出的不同解决⽅法进⾏了摸索和尝试，发现提⾼数据集样本容量是⼀个⽐较直接和简单粗暴的⽅法，以下纪录这⼀实验⽅法。
⼆、环境
直接交代环境，都是相对较简单，在这⾥博主没遇到过坑
os
numpy
PIL
imgaug
三、代码
ElementTree as ET
import os
import numpy as np
from PIL import Image
import shutil
import imgaug as ia
from imgaug import augmenters as iaa
ia.seed(1)
def read_xml_annotation(root, image_id):
in_file =open(os.path.join(root, image_id))
tree = ET.parse(in_file)
root = t()
bndboxlist =[]
for object in root.findall('object'):# 到root节点下的所有country节点
bndbox =object.find('bndbox')# ⼦节点下节点rank的值
xmin =int(bndbox.find('xmin').text)
xmax =int(bndbox.find('xmax').text)
ymin =int(bndbox.find('ymin').text)
ymax =int(bndbox.find('ymax').text)
# print(xmin,ymin,xmax,ymax)
bndboxlist.append([xmin, ymin, xmax, ymax])
# print(bndboxlist)
bndbox = root.find('object').find('bndbox')
return bndboxlist
# (506.0000, 330.0000, 528.0000, 348.0000) -> (520.4747, 381.5080, 540.5596, 398.6603)
def change_xml_annotation(root, image_id, new_target):
new_xmin = new_target[0]
new_ymin = new_target[1]
new_xmax = new_target[2]
new_ymax = new_target[3]
in_file =open(os.path.join(root,str(image_id)+'.xml'))# 这⾥root分别由两个意思
tree = ET.parse(in_file)
xmlroot = t()
xmlroot = t()
object= xmlroot.find('object')
bndbox =object.find('bndbox')
xmin = bndbox.find('xmin')
< =str(new_xmin)
ymin = bndbox.find('ymin')
< =str(new_ymin)
xmax = bndbox.find('xmax')
< =str(new_xmax)
ymax = bndbox.find('ymax')
< =str(new_ymax)
tree.write(os.path.join(root,str("%06d"%(str(id)+'.xml'))))
def change_xml_list_annotation(root, image_id, new_target, saveroot,id):
in_file =open(os.path.join(root,str(image_id)+'.xml'))# 这⾥root分别由两个意思    tree = ET.parse(in_file)
elem = tree.find('filename')
< =(str("%06d"%int(id))+'.jpg')
xmlroot = t()
index =0
for object in xmlroot.findall('object'):# 到root节点下的所有country节点
bndbox =object.find('bndbox')# ⼦节点下节点rank的值
# xmin = int(bndbox.find('xmin').text)
# xmax = int(bndbox.find('xmax').text)
# ymin = int(bndbox.find('ymin').text)
# ymax = int(bndbox.find('ymax').text)
new_xmin = new_target[index][0]
new_ymin = new_target[index][1]
new_xmax = new_target[index][2]
new_ymax = new_target[index][3]
xmin = bndbox.find('xmin')
< =str(new_xmin)
ymin = bndbox.find('ymin')
< =str(new_ymin)
xmax = bndbox.find('xmax')
< =str(new_xmax)
ymax = bndbox.find('ymax')
< =str(new_ymax)
index = index +1
tree.write(os.path.join(saveroot,str("%06d"%int(id))+'.xml'))
def mkdir(path):
# 去除⾸位空格
path = path.strip()
# 去除尾部 \ 符号
path = path.rstrip("/")
# 判断路径是否存在
# 存在    True
# 不存在  False
isExists = ists()
# 判断结果
if not isExists:
# 如果不存在则创建⽬录
# 创建⽬录操作函数
os.makedirs(path)
print(path +' 创建成功')
return True
else:
else:
# 如果⽬录存在则不创建，并提⽰⽬录已存在
print(path +' ⽬录已存在')
return False
if __name__ =="__main__":
IMG_DIR ="⾃⼰的⽂件路径/image"
XML_DIR ="⾃⼰的⽂件路径/Anotations"
AUG_XML_DIR ="F:/crop/Annotations"# 存储增强后的XML⽂件夹路径
try:
<(AUG_XML_DIR)
except FileNotFoundError as e:
a =1
mkdir(AUG_XML_DIR)
AUG_IMG_DIR ="F:/crop/images"# 存储增强后的影像⽂件夹路径
try:
<(AUG_IMG_DIR)
except FileNotFoundError as e:
a =1
mkdir(AUG_IMG_DIR)
AUGLOOP =10# 每张影像增强的数量
error parse newboxes_img_aug_list =[]
new_bndbox =[]
new_bndbox_list =[]
# 影像增强
seq = iaa.Sequential([
iaa.Flipud(0.5),# vertically flip 20% of all images
iaa.Fliplr(0.5),# 镜像
iaa.Multiply((1.2,1.5)),# change brightness, doesn't affect BBs
iaa.GaussianBlur(sigma=(0,2.0)),# iaa.GaussianBlur(0.5),
iaa.Affine(
translate_px={"x":15,"y":15},
scale=(0.8,0.95),
rotate=(-30,30)
)
# translate by 40/60px on x/y axis, and scale to 50-70%, affects BBs
])
for root, sub_folders, files in os.walk(XML_DIR):
for name in files:
bndbox = read_xml_annotation(XML_DIR, name)
for epoch in range(AUGLOOP):
seq_det = _deterministic()# 保持坐标和图像同步改变，⽽不是随机
# 读取图⽚
img = Image.open(os.path.join(IMG_DIR, name[:-4]+'.jpg'))
# sp = img.size
img = np.asarray(img)
# bndbox 坐标增强
for i in range(len(bndbox)):
bbs = ia.BoundingBoxesOnImage([
ia.BoundingBox(x1=bndbox[i][0], y1=bndbox[i][1], x2=bndbox[i][2], y2=bndbox[i][3]), ], shape=img.shape)
bbs_aug = seq_det.augment_bounding_boxes([bbs])[0]
boxes_img_aug_list.append(bbs_aug)
这是增强了11倍的结果：
对增强后的数据集进⾏训练，没有任何问题：

版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。