YOLO系列目标检测数据集-----数据增强
boxes[:, 3:] = boxes[:, 3:].clamp(min=0.0, max=1.0)# 确保宽度和高度在 [0, 1] 范围内。boxes = boxes_xyxy.clamp(min=0, max=size).reshape([-1, 4])# 压缩到固定范围。boxes[:, 1:] = boxes[:, 1:].clamp(min=0.0, max=1.0)# 确保坐标在
-
现有已经实现了一个功能完善的目标检测数据增强工具,支持多种增强方式。
-
通过添加边界框合法性检查、增强方法组合、随机种子控制等改进,可以进一步提高代码的鲁棒性和可扩展性。
-
如果需要支持更多高级增强方法,可以参考 YOLOv5 或 YOLOv8 的实现。
1、目标检测 数据集
结合了网络资源做出了相应修改,与时俱进,同时修复Bug
1)借鉴博文:YOLO目标检测数据集增强(不需要重新打标签,标签增强后会自动生成)_yolo数据增强-CSDN博客
2)代码如下:https://github.com/REN-HT/Data-Augmentation/blob/main/data_augmentation.py
# -*- coding: utf-8 -*-
"""
Created on 2025-03-24
@author: Fan yi ming 修改:shangjun
Func: 对于目标检测的数据增强[YOLO](特点是数据增强后标签也要更改)
review:常用的数据增强方式;
1.翻转:左右和上下翻转,随机翻转
2.随机裁剪,图像缩放
3.改变色调
4.添加噪声(自己调整椒盐等)
注意: boxes的标签和坐标一个是int,一个是float,存放的时候要注意处理方式。
参考:https://github.com/REN-HT/Data-Augmentation/blob/main/data_augmentation.py
"""
import torch
from PIL import Image
from PIL import ImageDraw
from PIL import ImageFile
ImageFile.LOAD_TRUNCATED_IMAGES = True
from torchvision import transforms
import numpy as np
import matplotlib.pyplot as plt
import os
import random
class DataAugmentationOnDetection:
def __init__(self, seed=0):
super(DataAugmentationOnDetection, self).__init__()
np.random.seed(seed)
torch.manual_seed(seed)
random.seed(seed)
def check_boxes(self, boxes, img_size):
"""
检查边界框的合法性,确保坐标在 [0, 1] 范围内
"""
w, h = img_size
boxes[:, 1:] = boxes[:, 1:].clamp(min=0.0, max=1.0) # 确保坐标在 [0, 1] 范围内
boxes[:, 3:] = boxes[:, 3:].clamp(min=0.0, max=1.0) # 确保宽度和高度在 [0, 1] 范围内
return boxes
def resize_keep_ratio(self, image, boxes, target_size):
"""
将图像缩放到size尺寸,调整相应的boxes,同时保持长宽比(最长的边是target size)
"""
old_size = image.size[0:2] # 原始图像大小
ratio = min(float(target_size) / (old_size[i]) for i in range(len(old_size))) # 计算缩放比例
new_size = tuple([int(i * ratio) for i in old_size]) # 计算新尺寸
return image.resize(new_size, Image.BILINEAR), boxes
def resizeDown_keep_ratio(self, image, boxes, target_size):
"""
缩小图像尺寸,保持长宽比,但不放大图像
"""
old_size = image.size[0:2]
ratio = min(float(target_size) / (old_size[i]) for i in range(len(old_size)))
ratio = min(ratio, 1)
new_size = tuple([int(i * ratio) for i in old_size])
return image.resize(new_size, Image.BILINEAR), boxes
def resize(self, img, boxes, size):
"""
将图像长和宽缩放到指定值size,并且相应调整boxes
"""
return img.resize((size, size), Image.BILINEAR), boxes
def random_flip_horizon(self, img, boxes, h_rate=1):
"""
随机水平翻转
"""
if np.random.random() < h_rate:
transform = transforms.RandomHorizontalFlip(p=1)
img = transform(img)
if len(boxes) > 0:
boxes[:, 1] = 1 - boxes[:, 1] # 调整边界框的x坐标
return img, boxes
def random_flip_vertical(self, img, boxes, v_rate=1):
"""
随机垂直翻转
"""
if np.random.random() < v_rate:
transform = transforms.RandomVerticalFlip(p=1)
img = transform(img)
if len(boxes) > 0:
boxes[:, 2] = 1 - boxes[:, 2] # 调整边界框的y坐标
return img, boxes
def center_crop(self, img, boxes, target_size=None):
"""
中心裁剪,裁剪成 (size, size) 的正方形
"""
w, h = img.size
size = min(w, h)
if len(boxes) > 0:
# 转换到xyxy格式
label = boxes[:, 0].reshape([-1, 1])
x_, y_, w_, h_ = boxes[:, 1], boxes[:, 2], boxes[:, 3], boxes[:, 4]
x1 = (w * x_ - 0.5 * w * w_).reshape([-1, 1])
y1 = (h * y_ - 0.5 * h * h_).reshape([-1, 1])
x2 = (w * x_ + 0.5 * w * w_).reshape([-1, 1])
y2 = (h * y_ + 0.5 * h * h_).reshape([-1, 1])
boxes_xyxy = torch.cat([x1, y1, x2, y2], dim=1)
# 边框转换
if w > h:
boxes_xyxy[:, [0, 2]] = boxes_xyxy[:, [0, 2]] - (w - h) / 2
else:
boxes_xyxy[:, [1, 3]] = boxes_xyxy[:, [1, 3]] - (h - w) / 2
in_boundary = [i for i in range(boxes_xyxy.shape[0])]
for i in range(boxes_xyxy.shape[0]):
# 判断x是否超出界限
if (boxes_xyxy[i, 0] < 0 and boxes_xyxy[i, 2] < 0) or (boxes_xyxy[i, 0] > size and boxes_xyxy[i, 2] > size):
in_boundary.remove(i)
# 判断y是否超出界限
elif (boxes_xyxy[i, 1] < 0 and boxes_xyxy[i, 3] < 0) or (boxes_xyxy[i, 1] > size and boxes_xyxy[i, 3] > size):
in_boundary.append(i)
boxes_xyxy = boxes_xyxy[in_boundary]
boxes = boxes_xyxy.clamp(min=0, max=size).reshape([-1, 4]) # 压缩到固定范围
label = label[in_boundary]
# 转换到YOLO格式
x1, y1, x2, y2 = boxes[:, 0], boxes[:, 1], boxes[:, 2], boxes[:, 3]
xc = ((x1 + x2) / (2 * size)).reshape([-1, 1])
yc = ((y1 + y2) / (2 * size)).reshape([-1, 1])
wc = ((x2 - x1) / size).reshape([-1, 1])
hc = ((y2 - y1) / size).reshape([-1, 1])
boxes = torch.cat([xc, yc, wc, hc], dim=1)
# 图像转换
transform = transforms.CenterCrop(size)
img = transform(img)
if target_size:
img = img.resize((target_size, target_size), Image.BILINEAR)
if len(boxes) > 0:
return img, torch.cat([label.reshape([-1, 1]), boxes], dim=1)
else:
return img, boxes
def random_bright(self, img, u=120, p=1):
"""
随机亮度变换
"""
if np.random.random() < p:
alpha = np.random.uniform(-u, u) / 255
img += alpha
img = img.clamp(min=0.0, max=1.0)
return img
def random_contrast(self, img, lower=0.5, upper=1.5, p=1):
"""
随机增强对比度
"""
if np.random.random() < p:
alpha = np.random.uniform(lower, upper)
img *= alpha
img = img.clamp(min=0, max=1.0)
return img
def random_saturation(self, img, lower=0.5, upper=1.5, p=1):
"""
随机饱和度变换
"""
if np.random.random() < p:
alpha = np.random.uniform(lower, upper)
img[1] = img[1] * alpha
img[1] = img[1].clamp(min=0, max=1.0)
return img
def add_gasuss_noise(self, img, mean=0, std=0.1):
"""
添加高斯噪声
"""
noise = torch.normal(mean, std, img.shape)
img += noise
img = img.clamp(min=0, max=1.0)
return img
def add_salt_noise(self, img):
"""
添加盐噪声
"""
noise = torch.rand(img.shape)
alpha = np.random.random() / 5 + 0.7
img[noise[:, :, :] > alpha] = 1.0
return img
def add_pepper_noise(self, img):
"""
添加胡椒噪声
"""
noise = torch.rand(img.shape)
alpha = np.random.random() / 5 + 0.7
img[noise[:, :, :] > alpha] = 0
return img
def plot_pics(img, boxes):
"""
显示图像和候选框
"""
plt.imshow(img)
label_colors = [(213, 110, 89)]
w, h = img.size
for i in range(boxes.shape[0]):
box = boxes[i, 1:]
xc, yc, wc, hc = box
x = w * xc - 0.5 * w * wc
y = h * yc - 0.5 * h * hc
box_w, box_h = w * wc, h * hc
plt.gca().add_patch(plt.Rectangle(xy=(x, y), width=box_w, height=box_h,
edgecolor=[c / 255 for c in label_colors[0]],
fill=False, linewidth=2))
plt.show()
def get_image_list(image_path):
"""
获取图像文件列表
"""
files_list = []
for root, sub_dirs, files in os.walk(image_path):
for special_file in files:
special_file = special_file[0: len(special_file)]
files_list.append(special_file)
return files_list
def get_label_file(label_path, image_name):
"""
获取标签文件
"""
fname = os.path.join(label_path, image_name[0: len(image_name) - 4] + ".txt")
data2 = []
if not os.path.exists(fname):
return data2
if os.path.getsize(fname) == 0:
return data2
else:
with open(fname, 'r', encoding='utf-8') as infile:
for line in infile:
data_line = line.strip("\n").split()
data2.append([float(i) for i in data_line])
return data2
def save_Yolo(img, boxes, save_path, prefix, image_name):
"""
保存增强后的图像和标签
"""
if not os.path.exists(save_path) or \
not os.path.exists(os.path.join(save_path, "images")):
os.makedirs(os.path.join(save_path, "images"))
os.makedirs(os.path.join(save_path, "labels"))
try:
img.save(os.path.join(save_path, "images", prefix + image_name))
with open(os.path.join(save_path, "labels", prefix + image_name[0:len(image_name) - 4] + ".txt"), 'w', encoding="utf-8") as f:
if len(boxes) > 0: # 判断是否为空
for data in boxes:
str_in = ""
for i, a in enumerate(data):
if i == 0:
str_in += str(int(a))
else:
str_in += " " + str(float(a))
f.write(str_in + '\n')
except:
print("ERROR: ", image_name, " is bad.")
def runAugumentation(image_path, label_path, save_path):
"""
运行数据增强
"""
image_list = get_image_list(image_path)
for image_name in image_list:
print("dealing: " + image_name)
img = Image.open(os.path.join(image_path, image_name))
boxes = get_label_file(label_path, image_name)
boxes = torch.tensor(boxes)
# 下面是执行的数据增强功能,可自行选择
DAD = DataAugmentationOnDetection()
""" 尺寸变换 """
# 缩小尺寸
# t_img, t_boxes = DAD.resizeDown_keep_ratio(img, boxes, 1024)
# save_Yolo(t_img, boxes, save_path, prefix="rs_", image_name=image_name)
# 水平旋转
t_img, t_boxes = DAD.random_flip_horizon(img, boxes.clone())
save_Yolo(t_img, t_boxes, save_path, prefix="fh_", image_name=image_name)
# 竖直旋转
t_img, t_boxes = DAD.random_flip_vertical(img, boxes.clone())
save_Yolo(t_img, t_boxes, save_path, prefix="fv_", image_name=image_name)
# center_crop
t_img, t_boxes = DAD.center_crop(img, boxes.clone(), 1024)
save_Yolo(t_img, t_boxes, save_path, prefix="cc_", image_name=image_name)
""" 图像变换,用tensor类型"""
to_tensor = transforms.ToTensor()
to_image = transforms.ToPILImage()
img = to_tensor(img)
# random_bright
t_img, t_boxes = DAD.random_bright(img.clone()), boxes
save_Yolo(to_image(t_img), boxes, save_path, prefix="rb_", image_name=image_name)
# random_contrast 对比度变化
t_img, t_boxes = DAD.random_contrast(img.clone()), boxes
save_Yolo(to_image(t_img), boxes, save_path, prefix="rc_", image_name=image_name)
# random_saturation 饱和度变化
t_img, t_boxes = DAD.random_saturation(img.clone()), boxes
save_Yolo(to_image(t_img), boxes, save_path, prefix="rs_", image_name=image_name)
# 高斯噪声
t_img, t_boxes = DAD.add_gasuss_noise(img.clone()), boxes
save_Yolo(to_image(t_img), boxes, save_path, prefix="gn_", image_name=image_name)
# add_salt_noise
t_img, t_boxes = DAD.add_salt_noise(img.clone()), boxes
save_Yolo(to_image(t_img), boxes, save_path, prefix="sn_", image_name=image_name)
# add_pepper_noise
t_img, t_boxes = DAD.add_pepper_noise(img.clone()), boxes
save_Yolo(to_image(t_img), boxes, save_path, prefix="pn_", image_name=image_name)
print("end: " + image_name)
if __name__ == '__main__':
# 图像和标签文件夹
image_path = r"images" # 图片位置路径
label_path = r"labels" # 标签位置路径
save_path = r"Augumentation" # 结果保存位置路径,可以是一个不存在的文件夹
# 运行
runAugumentation(image_path, label_path, save_path)
3)改进点总结:
-
边界框合法性检查:
-
在每次增强后,检查边界框的坐标是否在图像范围内。
-
-
随机种子控制:
-
在
DataAugmentationOnDetection的__init__方法中设置随机种子,确保增强过程可复现。
-
-
增强方法组合:
-
支持多种增强方法的组合,例如同时进行翻转和裁剪。
-
-
文件名唯一性处理:
-
在保存增强后的图像和标签时,使用增强方法的名称作为前缀,确保文件名唯一。
-
-
代码结构优化:
-
将增强方法封装在类中,便于扩展和维护
-
数据集划分
imgpath = 'VOCdevkit/JPEGImages' # 将上面的images路径丢这里 txtpath = 'VOCdevkit/txt' # 将上面的labels路径丢这里
## 数据集划分
import os, shutil, random
random.seed(0)
import numpy as np
from sklearn.model_selection import train_test_split
val_size = 0.1
test_size = 0.1
postfix = 'jpg'
imgpath = 'VOCdevkit/JPEGImages'
txtpath = 'VOCdevkit/txt'
os.makedirs('images/train', exist_ok=True)
os.makedirs('images/val', exist_ok=True)
os.makedirs('images/test', exist_ok=True)
os.makedirs('labels/train', exist_ok=True)
os.makedirs('labels/val', exist_ok=True)
os.makedirs('labels/test', exist_ok=True)
listdir = np.array([i for i in os.listdir(txtpath) if 'txt' in i])
random.shuffle(listdir)
train, val, test = listdir[:int(len(listdir) * (1 - val_size - test_size))], listdir[int(len(listdir) * (1 - val_size - test_size)):int(len(listdir) * (1 - test_size))], listdir[int(len(listdir) * (1 - test_size)):]
print(f'train set size:{len(train)} val set size:{len(val)} test set size:{len(test)}')
for i in train:
shutil.copy('{}/{}.{}'.format(imgpath, i[:-4], postfix), 'images/train/{}.{}'.format(i[:-4], postfix))
shutil.copy('{}/{}'.format(txtpath, i), 'labels/train/{}'.format(i))
for i in val:
shutil.copy('{}/{}.{}'.format(imgpath, i[:-4], postfix), 'images/val/{}.{}'.format(i[:-4], postfix))
shutil.copy('{}/{}'.format(txtpath, i), 'labels/val/{}'.format(i))
for i in test:
shutil.copy('{}/{}.{}'.format(imgpath, i[:-4], postfix), 'images/test/{}.{}'.format(i[:-4], postfix))
shutil.copy('{}/{}'.format(txtpath, i), 'labels/test/{}'.format(i))更多推荐



所有评论(0)