YOLOV 训练

百度网盘链接：https://pan.baidu.com/s/12x2t-EYbRC6q2IqqiznZVQ?pwd=LIIB 
提取码：LIIB 

环境搭建网上搜一下，cuda cudnn nvidia驱动 anconda pytorch 

pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple/

pytorch 安装与cuda驱动一致的版本

图片放在my_data的images下

标注xml数据将文件放my_data的Annotations下运行xml_txt.py

标注txt文件直接将结果放在my_data的labels（xml文件最终是要转为txt的，所以上面在转换）

上述结束运行my_data下的data.py将数据集划分为3组   放置在VOCData文件下

回到yolov7下运行：

//训练用
python train.py --weights ./yolov7.pt --cfg ./cfg/training/yolov7.yaml --data ./myvoc.yaml --device 0 --batch-size 4 --epoch 300
//训练结束测试用
python detect.py --weigths ./runs/train/exp24/weights/best.pt --source ./VOCData/1/

1.数据集xml文件转txt格式  xml_data.py

import glob
import os
# xml文件的路径
xml_path = './Annotations/'

#定义从xml获取信息的函数
def _read_anno(filename):

    import xml.etree.ElementTree as ET
    tree = ET.parse(filename)
    #获取宽w和高h
    a = tree.find('size')
    w,h = [int(a.find('width').text),
           int(a.find('height').text)]
    #print(w,h)

    objects = []
    #这里是针对错误xml文件，图片的w和h都为0，这样的xml文件可以直接忽视，返回空列表
    if w == 0:
        return []
    #这里需要根据需要修改，因为我训练的目的是判断是否戴了头盔，因此从xml获取的name为none或者0的label都为0，其他的颜色或者1都为1
    for obj in tree.findall('object'):
    	#获取name，我上边的实例图片中的红色区域
        name = obj.find('name').text
        #修改label，这里是不同数据集大融合的关键
        if name == 'head':
            label = 0
        elif name == 'trail':
            label = 1
		#读取检测框的左上、右下角点的坐标
        bbox = obj.find('bndbox')
        x1, y1, x2, y2 = [int(bbox.find('xmin').text),
                          int(bbox.find('ymin').text),
                          int(bbox.find('xmax').text),
                          int(bbox.find('ymax').text)]
		#这里也很关键，yolov5需要中心点以及宽和高的标注信息，并且进行归一化，下边label后边的四个值即是归一化后保留4位有效数字的x，y，w，h
        obj_struct = [label,round((x1+x2)/(2.0*w),4), round((y1+y2)/(2.0*h),4), round((x2-x1)/(w),4),round((y2-y1)/(h),4)]
        print(obj_struct)

        objects.append(obj_struct)


    return objects
#接下来是写入txt文件中
if __name__ == '__main__':
    #定义一个空的字符串
    t = ''
    # txt文件存放的路径
    txt_path = "./labels/"
    #获取所有的xml文件路径
    allfilepath = []
    for file in os.listdir(xml_path):
        if file.endswith('.xml'):
            file = os.path.join(xml_path,file)
            allfilepath.append(file)
            # print(allfilepath)
        else:
            pass
     #生成需要的对应xml文件名的txt
    for file in allfilepath:
        print("file", file)
        #获取xml的文件名
        filename = file.split('/')[2]
        indexname = filename.split('.')[0]
        # print('indexname', indexname)
        result = _read_anno(file)
        # print('result', result)
        #跳过空列表
        if len(result)==0:
            continue
        #写入信息，注意每次循环结束都把t重新定义，result是一个二维列表（行数为目标个数，列对应label和位置信息），为了避免读取出错（还有一个原因是我菜），我们一个一个的写入。       
        txtfile = open(os.path.join(txt_path,indexname+'.txt'), 'w')
        for line in result:
            for a in line:
                # print('a', a)
                t = t + str(a) + ' '
                # print('t', t)
                txtfile.writelines(t)
                t = ''
            txtfile.writelines('\n')            

2.创建训练data.py

# 将图片和标注数据按比例切分为 训练集和测试集
import shutil
import random
import os
 
# 原始路径
image_original_path = "./images/"
label_original_path = "./labels/"
 
cur_path = os.getcwd()
 
# 训练集路径
train_image_path = os.path.join(cur_path, "../VOCData/images/train/")
train_label_path = os.path.join(cur_path, "../VOCData/labels/train/")
 
# 验证集路径
val_image_path = os.path.join(cur_path, "../VOCData/images/val/")
val_label_path = os.path.join(cur_path, "../VOCData/labels/val/")
 
# 测试集路径
test_image_path = os.path.join(cur_path, "../VOCData/images/test/")
test_label_path = os.path.join(cur_path, "../VOCData/labels/test/")
 
# 训练集目录
list_train = os.path.join(cur_path, "../VOCData/train.txt")
list_val = os.path.join(cur_path, "../VOCData/val.txt")
list_test = os.path.join(cur_path, "../VOCData/test.txt")
 
train_percent = 0.6
val_percent = 0.2
test_percent = 0.2
 
 
def del_file(path):
    for i in os.listdir(path):
        file_data = path + "\\" + i
        os.remove(file_data)
 
 
def mkdir():
    if not os.path.exists(train_image_path):
        os.makedirs(train_image_path)
    else:
        del_file(train_image_path)
    if not os.path.exists(train_label_path):
        os.makedirs(train_label_path)
    else:
        del_file(train_label_path)
 
    if not os.path.exists(val_image_path):
        os.makedirs(val_image_path)
    else:
        del_file(val_image_path)
    if not os.path.exists(val_label_path):
        os.makedirs(val_label_path)
    else:
        del_file(val_label_path)
 
    if not os.path.exists(test_image_path):
        os.makedirs(test_image_path)
    else:
        del_file(test_image_path)
    if not os.path.exists(test_label_path):
        os.makedirs(test_label_path)
    else:
        del_file(test_label_path)
 
 
def clearfile():
    if os.path.exists(list_train):
        os.remove(list_train)
    if os.path.exists(list_val):
        os.remove(list_val)
    if os.path.exists(list_test):
        os.remove(list_test)
 
 
def main():
    mkdir()
    clearfile()
 
    file_train = open(list_train, 'w')
    file_val = open(list_val, 'w')
    file_test = open(list_test, 'w')
 
    total_txt = os.listdir(label_original_path)
    num_txt = len(total_txt)
    list_all_txt = range(num_txt)
 
    num_train = int(num_txt * train_percent)
    num_val = int(num_txt * val_percent)
    num_test = num_txt - num_train - num_val
 
    train = random.sample(list_all_txt, num_train)
    # train从list_all_txt取出num_train个元素
    # 所以list_all_txt列表只剩下了这些元素
    val_test = [i for i in list_all_txt if not i in train]
    # 再从val_test取出num_val个元素，val_test剩下的元素就是test
    val = random.sample(val_test, num_val)
 
    print("训练集数目：{}, 验证集数目：{}, 测试集数目：{}".format(len(train), len(val), len(val_test) - len(val)))
    for i in list_all_txt:
        name = total_txt[i][:-4]
 
        srcImage = image_original_path + name + '.jpg'
        srcLabel = label_original_path + name + ".txt"
 
        if i in train:
            dst_train_Image = train_image_path + name + '.jpg'
            dst_train_Label = train_label_path + name + '.txt'
            shutil.copyfile(srcImage, dst_train_Image)
            shutil.copyfile(srcLabel, dst_train_Label)
            file_train.write(dst_train_Image + '\n')
        elif i in val:
            dst_val_Image = val_image_path + name + '.jpg'
            dst_val_Label = val_label_path + name + '.txt'
            shutil.copyfile(srcImage, dst_val_Image)
            shutil.copyfile(srcLabel, dst_val_Label)
            file_val.write(dst_val_Image + '\n')
        else:
            dst_test_Image = test_image_path + name + '.jpg'
            dst_test_Label = test_label_path + name + '.txt'
            shutil.copyfile(srcImage, dst_test_Image)
            shutil.copyfile(srcLabel, dst_test_Label)
            file_test.write(dst_test_Image + '\n')
 
    file_train.close()
    file_val.close()
    file_test.close()
 
 
if __name__ == "__main__":
    main()