1. 对图片进行预测

from ultralytics import YOLO

model = YOLO(r'D:\Code\Graduation project\ultralytics-main\ultralytics-main\runs\detect\train\weights\last.pt')
model.predict(r'D:\Code\Graduation project\ultralytics-main\ultralytics-main\datasets\det_fire_smoke_15000\ceshi\}1]}(EEWQK6G7I[)ZN9HX@0.png', save=True, conf=0.5)

2. 对视频进行预测

from ultralytics import YOLO

model = YOLO(r'D:\Code\Graduation project\ultralytics-main\ultralytics-main\runs\detect\train\weights\last.pt')
model.predict(r'D:\Code\Graduation project\ultralytics-main\ultralytics-main\datasets\det_fire_smoke_15000\ceshi\fire.mp4', save=True, conf=0.5)

3. 摄像头实时检测

import cv2
from ultralytics import YOLO

# 加载YOLOv8模型
model = YOLO(r'D:\\Code\Graduation project\\ultralytics-main\\ultralytics-main\\runs\\detect\\train2\\weights\\best.pt')  # 这里要什么就换什么yolov8i默认目标检测（i= n,m,l,x）尾缀有-pose,-seg

# 打开视频文件
# video_path = "path/to/your/video/file.mp4"
cap = cv2.VideoCapture(0)   # 0表示默认摄像头，如果有多个摄像头，可以尝试使用1, 2, 等

# 遍历视频帧
while cap.isOpened():
    # 从视频中读取一帧
    success, frame = cap.read()

    if success:
        # 在该帧上运行YOLOv8推理
        results = model(frame)

        # 在帧上可视化结果
        annotated_frame = results[0].plot()

        # 显示带注释的帧
        cv2.imshow("YOLOv8推理", annotated_frame)

        # 如果按下'q'则中断循环
        if cv2.waitKey(1) & 0xFF == ord("q"):
            break
    else:
        # 如果视频结束则中断循环
        break

# 释放视频捕获对象并关闭显示窗口
cap.release()
cv2.destroyAllWindows()