探索 Ultralytics YOLOv8标记图片

1、下载YOLOv8模型文件

下载地址：https://docs.ultralytics.com/zh/models/yolov8/#performance-metrics

2、编写python脚本 aaa.py

import cv2
import numpy as np
from ultralytics import YOLO
import matplotlib.pyplot as plt

def plot_detection(image, boxes, scores, class_ids, class_names, conf_threshold=0.5):
    """
    将检测结果绘制到图像上
    """
    # 创建图像的副本用于绘制
    img_with_detections = image.copy()
    
    # 定义颜色列表，为不同类别使用不同颜色
    colors = [(0, 255, 0), (255, 0, 0), (0, 0, 255), 
              (255, 255, 0), (255, 0, 255), (0, 255, 255)]
    
    # 遍历所有检测结果
    for i, (box, score, class_id) in enumerate(zip(boxes, scores, class_ids)):
        # 过滤低置信度的检测结果
        if score < conf_threshold:
            continue
            
        # 获取框的坐标 (x1, y1, x2, y2)
        x1, y1, x2, y2 = map(int, box)
        
        # 选择颜色 (循环使用颜色列表)
        color = colors[class_id % len(colors)]
        
        # 1. 绘制边界框
        cv2.rectangle(img_with_detections, (x1, y1), (x2, y2), color, 2)
        
        # 2. 绘制标签背景
        label = f"{class_names[class_id]}: {score:.2f}"
        (label_width, label_height), baseline = cv2.getTextSize(
            label, cv2.FONT_HERSHEY_SIMPLEX, 0.5, 1)
        
        cv2.rectangle(img_with_detections, 
                     (x1, y1 - label_height - 5), 
                     (x1 + label_width, y1), 
                     color, -1)  # -1 表示填充矩形
        
        # 3. 绘制标签文本
        cv2.putText(img_with_detections, label, 
                   (x1, y1 - 5), 
                   cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, 255, 255), 1)
    
    return img_with_detections

def main():
    # 1. 加载预训练的 YOLOv8n 模型
    print("正在加载 YOLOv8n 模型...")
    model = YOLO('yolov8n.pt')  # 会自动下载模型如果不存在
    
    # 2. 加载要检测的图像
    image_path = '123.jpg'  # 替换为你的图片路径
    print(f"正在加载图像: {image_path}")
    image = cv2.imread(image_path)
    if image is None:
        print(f"错误: 无法加载图像 {image_path}")
        return
    
    # 3. 进行推理
    print("正在进行推理...")
    results = model(image, verbose=False)  # verbose=False 减少输出信息
    
    # 4. 提取检测结果
    # 获取第一个结果（因为只输入了一张图片）
    result = results[0]
    
    # 提取边界框信息
    boxes = result.boxes.xyxy.cpu().numpy()      # 边界框坐标 [x1, y1, x2, y2]
    scores = result.boxes.conf.cpu().numpy()     # 置信度分数
    class_ids = result.boxes.cls.cpu().numpy().astype(int)  # 类别ID
    
    # 获取类别名称
    class_names = result.names
    print(class_names)
    
    print(f"检测到 {len(boxes)} 个对象")
    for i, (box, score, class_id) in enumerate(zip(boxes, scores, class_ids)):
        print(f"对象 {i+1}: {class_names[class_id]} (置信度: {score:.3f})")
    
    # 5. 将检测结果绘制到图像上
    print("正在绘制检测结果...")
    result_image = plot_detection(image, boxes, scores, class_ids, class_names)
    
    # 6. 保存和显示结果
    output_path = 'detection_result123-x.jpg'
    cv2.imwrite(output_path, result_image)
    print(f"结果已保存到: {output_path}")
    
    # 使用 matplotlib 显示结果（支持中文显示）
    plt.figure(figsize=(12, 8))
    
    # 将 BGR 转换为 RGB 用于 matplotlib 显示
    result_image_rgb = cv2.cvtColor(result_image, cv2.COLOR_BGR2RGB)
    
    plt.imshow(result_image_rgb)
    plt.title('YOLOv8 result')
    plt.axis('off')
    plt.tight_layout()
    #plt.savefig('detection_result_plt.jpg', dpi=300, bbox_inches='tight')
    plt.show()

def detect_specific_classes():
    """
    示例：只检测特定的类别（如可能类似于电表的物体）
    """
    # 可能类似于电表的 COCO 类别
    target_classes = {
        67: 'cell phone',    # 手机 - 最可能检测到数字屏幕
        62: 'tvmonitor',     # 显示器
        75: 'remote',        # 遥控器
        74: 'clock'          # 时钟
    }
    
    # 加载模型和图像
    model = YOLO('yolov8n.pt')
    image = cv2.imread('your_meter_image.jpg')
    
    # 推理
    results = model(image, verbose=False)
    result = results[0]
    
    # 提取结果
    boxes = result.boxes.xyxy.cpu().numpy()
    scores = result.boxes.conf.cpu().numpy()
    class_ids = result.boxes.cls.cpu().numpy().astype(int)
    class_names = result.names
    
    # 过滤出目标类别
    filtered_boxes = []
    filtered_scores = []
    filtered_class_ids = []
    
    for box, score, class_id in zip(boxes, scores, class_ids):
        if class_id in target_classes and score > 0.3:  # 降低阈值
            filtered_boxes.append(box)
            filtered_scores.append(score)
            filtered_class_ids.append(class_id)
            print(f"检测到可能的目标: {target_classes[class_id]} (置信度: {score:.3f})")
    
    # 绘制结果
    if filtered_boxes:
        result_image = plot_detection(image, filtered_boxes, filtered_scores, 
                                    filtered_class_ids, class_names, conf_threshold=0.3)
        cv2.imwrite('meter_detection_result.jpg', result_image)
        print("电表检测结果已保存")
    else:
        print("未检测到可能的目标物体")

if __name__ == "__main__":
    # 运行主函数进行通用物体检测
    main()
    
    # 或者运行特定类别检测（用于电表检测尝试）
    # detect_specific_classes()

3、运行python脚本

python aaa.py

正在加载 YOLOv8n 模型...
正在加载图像: 10.jpg
正在进行推理...
检测品类：{0: 'person', 1: 'bicycle', 2: 'car', 3: 'motorcycle', 4: 'airplane', 5: 'bus', 6: 'train', 
7: 'truck', 8: 'boat', 9: 'traffic light', 10: 'fire hydrant', 11: 'stop sign', 12: 
'parking meter', 13: 'bench', 14: 'bird', 15: 'cat', 16: 'dog', 17: 'horse', 18: 'sheep', 
19: 'cow', 20: 'elephant', 21: 'bear', 22: 'zebra', 23: 'giraffe', 24: 'backpack', 25: 
'umbrella', 26: 'handbag', 27: 'tie', 28: 'suitcase', 29: 'frisbee', 30: 'skis', 31: 
'snowboard', 32: 'sports ball', 33: 'kite', 34: 'baseball bat', 35: 'baseball glove', 36:
 'skateboard', 37: 'surfboard', 38: 'tennis racket', 39: 'bottle', 40: 'wine glass', 41: 
'cup', 42: 'fork', 43: 'knife', 44: 'spoon', 45: 'bowl', 46: 'banana', 47: 'apple', 48: 
'sandwich', 49: 'orange', 50: 'broccoli', 51: 'carrot', 52: 'hot dog', 53: 'pizza', 54: 
'donut', 55: 'cake', 56: 'chair', 57: 'couch', 58: 'potted plant', 59: 'bed', 60: 'dining 
table', 61: 'toilet', 62: 'tv', 63: 'laptop', 64: 'mouse', 65: 'remote', 66: 'keyboard',
 67: 'cell phone', 68: 'microwave', 69: 'oven', 70: 'toaster', 71: 'sink', 72: 
'refrigerator', 73: 'book', 74: 'clock', 75: 'vase', 76: 'scissors', 77: 'teddy bear', 78:
 'hair drier', 79: 'toothbrush'}

检测到 2 个对象
对象 1: chair (置信度: 0.968)
对象 2: chair (置信度: 0.946)
正在绘制检测结果...
结果已保存到: detection_result10.jpg

python脚本、模型文件、图片文件在一个目录中

参考官网内容：https://docs.ultralytics.com/zh/models/yolov8/#performance-metrics