前期准备工作

需要使用到的库，需要训练的素材一份图片或者视频

import ultralytics # Yolo V8 本体
import lableimg # 图片标注工具

接着新建一份工作目录如下

--- data
  --- Annotations # 存放标记数据
  --- images # 存放需要训练的图片素材
  --- imageSets # 存放训练集 测试集 验证集的路径数据
  --- labels # 存放 Yolo 格式的数据集

视频转图片

可以使用 OpenCV 按帧进行转换，使用如下代码就得到一份图片素材

import os
import cv2

def video_to_frame(video_path):
    videCapture = cv2.VideoCapture()
    videCapture.open(video_path)

    vide_name = os.path.basename(video_path).split('.mp4')[0]

    n = 1
    frametime = 30

    frames = videCapture.get(cv2.CAP_PROP_FRAME_COUNT)
    print(frames)
    frame_path = os.path.join(os.path.dirname(video_path), vide_name)
    if not os.path.exists(frame_path): os.mkdir(frame_path)

    for i in range(int(frames)):
        ret, frame = videCapture.read()
        if ret: 
            if i % frametime == 0:
                pic_path = os.path.join(frame_path, f'{vide_name}_{n}.jpg')
                cv2.imencode('.jpg', frame)[1].tofile(pic_path)
                print(pic_path)
                n += 1

if __name__ == "__main__":
    video_folder_path = r'C:\Users\Downloads\Video'
    vide_names = os.listdir(video_folder_path)
    for video_name in vide_names:
        video_path = os.path.join(video_folder_path, video_name)
        try:
            video_to_frame(video_path)
        except Exception as e:
            print(f'ERROR : {e}')

进行图片标注

这里就不细讲了，具体可百度 labelimg 的使用

按比例划分训练集和验证集

import os
import random
path_folder_data = os.path.dirname(os.path.realpath(__file__))
os.chdir('./data')

# 训练集和验证集的比例分配
trainval_percent = 0.1
train_percent = 0.9

# 标注文件的路径
xmlfilepath = 'Annotations'

# 生成的txt文件存放路径
txtsavepath = 'ImageSets\Main'
total_xml = os.listdir(xmlfilepath)
if not os.path.exists(txtsavepath):
    os.makedirs(txtsavepath)

num = len(total_xml)
list = range(num)
tv = int(num * trainval_percent)
tr = int(tv * train_percent)
trainval = random.sample(list, tv)
train = random.sample(trainval, tr)

ftrainval = open('ImageSets/Main/trainval.txt', 'w')
ftest = open('ImageSets/Main/test.txt', 'w')
ftrain = open('ImageSets/Main/train.txt', 'w')
fval = open('ImageSets/Main/val.txt', 'w')

for i in list:
    name = total_xml[i][:-4] + '\n'
    if i in trainval:
        ftrainval.write(name)
        if i in train:
            ftest.write(name)
        else:
            fval.write(name)
    else:
        ftrain.write(name)

ftrainval.close()
ftrain.close()
fval.close()
ftest.close()

将标注数据转化为 Yolo 格式

import xml.etree.ElementTree as ET
import pickle
import os
from os import listdir, getcwd
from os.path import join
path_folder_data = os.path.dirname(os.path.realpath(__file__)) + "\data"

# 原始脚本中包含了VOC2012，这里，我们把它删除
# sets=[('2012', 'train'), ('2012', 'val'), ('2007', 'train'), ('2007', 'val'), ('2007', 'test')]
sets=['train', 'test', 'val']

# classes也需要根据自己的实际情况修改
# classes = ["aeroplane", "bicycle", "bird", "boat", "bottle", "bus", "car", "cat", "chair", "cow", "diningtable", "dog", "horse", "motorbike", "person", "pottedplant", "sheep", "sofa", "train", "tvmonitor"]
classes = ['person', 'bicycle', 'car', 'motorcycle']


def convert(size, box):
    dw = 1./size[0]
    dh = 1./size[1]
    x = (box[0] + box[1])/2.0
    y = (box[2] + box[3])/2.0
    w = box[1] - box[0]
    h = box[3] - box[2]
    x = x*dw
    w = w*dw
    y = y*dh
    h = h*dh
    return (x,y,w,h)

def convert_annotation(image_id):
    in_file = open(f'{path_folder_data}//Annotations//{image_id}.xml', encoding='utf-8')
    out_file = open(f'{path_folder_data}//labels//{image_id}.txt', 'w')
    tree=ET.parse(in_file)
    root = tree.getroot()
    size = root.find('size')
    w = int(size.find('width').text)
    h = int(size.find('height').text)

    for obj in root.iter('object'):
        difficult = obj.find('difficult').text
        cls = obj.find('name').text
        if cls not in classes or int(difficult) == 1:
            continue
        cls_id = classes.index(cls)
        xmlbox = obj.find('bndbox')
        b = (float(xmlbox.find('xmin').text), float(xmlbox.find('xmax').text), float(xmlbox.find('ymin').text), float(xmlbox.find('ymax').text))
        bb = convert((w,h), b)
        out_file.write(str(cls_id) + " " + " ".join([str(a) for a in bb]) + '\n')

wd = getcwd()

for image_set in sets:
    if not os.path.exists(f'{path_folder_data}//labels/'):
        os.makedirs(f'{path_folder_data}//labels/')
    image_ids = open(f'{path_folder_data}//imageSets//Main//{image_set}.txt').read().splitlines()
    list_file = open(f'{path_folder_data}//{image_set}.txt', 'w')
    for image_id in image_ids:
        list_file.write(f'{path_folder_data}//images//{image_id}.jpg\n')
        convert_annotation(image_id)
    list_file.close()

生成配置文件

# train/val/test sets
train: C:\User\WorkSpace\Python\Yolo\data\train.txt
val: C:\User\WorkSpace\Python\Yolo\data\val.txt
test: C:\User\WorkSpace\Python\Yolo\data\test.txt

# class names
names:
  0: person
  1: bicycle
  2: car
  3: motorcycle

# nc: number of classes
nc: 4

开始训练

至此需要训练的数据已准备完毕，可以开始训练，具体参数可以查看官网

yolo detect train data=coco128.yaml model=yolov8n.pt epochs=100 imgsz=640
# data 为配置文件，epoch 为训练次数，model 为原始模型文件

训练完成后会在 data 下生成 run 文件夹，一路点进去就可以找到训练的结果以及生成的模型文件

训练模型的使用

yolo detect predict model=yolov8n.pt source="https://ultralytics.com/images/bus.jpg" save=True # predict with official model
yolo detect predict model=path/to/best.pt source="https://ultralytics.com/images/bus.jpg" save=True # predict with custom model

参考

Yolo V8 官方文档