基于YOLO26的南瓜叶片病害检测系统（Python源码+数据集+Pyside6界面）

重复模块m (n次迭代):
┌─────────────────────────────────────────────────────────┐
│                                                         │
│  如果 attn=True:                                        │
│    Sequential(                                          │
│        Bottleneck(self.c, self.c),                     │  ←─ 先特征提取
│        PSABlock(self.c, attn_ratio=0.5, num_heads=...) │  ←─ 后注意力增强
│    )                                                    │
│                                                         │
│  否则如果 c3k=True:                                     │
│    C3k(self.c, self.c, 2)                             │  ←─ 同YOLOv11
│                                                         │
│  否则:                                                  │
│    Bottleneck(self.c, self.c)                         │  ←─ 同YOLOv11
│                                                         │
└─────────────────────────────────────────────────────────┘

class BboxLoss(nn.Module):
    """Criterion class for computing training losses for bounding boxes."""

    def __init__(self, reg_max: int = 16):
        """Initialize the BboxLoss module with regularization maximum and DFL settings."""
        super().__init__()
        self.dfl_loss = DFLoss(reg_max) if reg_max > 1 else None

# pumpkin_disease_detection.py
# 基于 YOLO26 的南瓜叶片病害检测系统

from ultralytics import YOLO  # 假设 YOLO26 遵循 Ultralytics API
import os
import cv2
import glob
import matplotlib.pyplot as plt

# ========================== 1. 配置参数 ==========================
DATA_YAML = "pumpkin_disease.yaml"   # 数据集配置文件 (包含 paths, nc, names)
MODEL_NAME = "yolo26n.pt"           # 预训练模型 (n/s/m/l/x)
EPOCHS = 100
IMG_SIZE = 640
BATCH_SIZE = 16
DEVICE = 0                          # GPU 设备号，-1 表示 CPU

# 数据集配置文件内容示例 (pumpkin_disease.yaml):
"""
train: ./datasets/train/images
val: ./datasets/val/images
test: ./datasets/test/images
nc: 5
names: ['Bacterial Leaf Spot', 'Downy Mildew', 'Healthy Leaf', 'Mosaic disease', 'Powdery_Mildew']
"""

# ========================== 2. 加载模型 ==========================
def load_model(weights_path=MODEL_NAME):
    """加载预训练 YOLO26 模型"""
    model = YOLO(weights_path)
    return model

# ========================== 3. 训练模型 ==========================
def train_model(model, data_yaml):
    """训练模型"""
    results = model.train(
        data=data_yaml,
        epochs=EPOCHS,
        imgsz=IMG_SIZE,
        batch=BATCH_SIZE,
        device=DEVICE,
        project="pumpkin_disease_detection",
        name="exp1",
        exist_ok=True,
        patience=10,
        lr0=0.01,
        augment=True,
        mosaic=0.5,
        mixup=0.2,
        copy_paste=0.2
    )
    return results

# ========================== 4. 评估模型 ==========================
def evaluate_model(model, data_yaml):
    """在验证集上评估模型"""
    metrics = model.val(data=data_yaml, device=DEVICE)
    print(f"mAP@0.5: {metrics.box.map50:.4f}")
    print(f"mAP@0.5:0.95: {metrics.box.map:.4f}")
    print(f"Precision: {metrics.box.mp:.4f}")
    print(f"Recall: {metrics.box.mr:.4f}")
    return metrics

# ========================== 5. 单张图像推理 ==========================
def predict_single(model, image_path, conf_thres=0.25, iou_thres=0.45):
    """对单张图像进行病害检测并显示结果"""
    results = model(image_path, conf=conf_thres, iou=iou_thres)[0]
    # 加载图像并绘制检测框
    img = cv2.imread(image_path)
    img_rgb = cv2.cvtColor(img, cv2.COLOR_BGR2RGB)
    for box in results.boxes:
        x1, y1, x2, y2 = map(int, box.xyxy[0].tolist())
        conf = box.conf[0].item()
        cls_id = int(box.cls[0].item())
        label = f"{results.names[cls_id]} {conf:.2f}"
        cv2.rectangle(img, (x1, y1), (x2, y2), (0, 255, 0), 2)
        cv2.putText(img, label, (x1, y1-5), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0,255,0), 2)
    # 显示
    plt.imshow(cv2.cvtColor(img, cv2.COLOR_BGR2RGB))
    plt.axis('off')
    plt.title("Detection Result")
    plt.show()
    return results

# ========================== 6. 批量推理并保存结果 ==========================
def predict_batch(model, image_dir, output_dir, conf_thres=0.25):
    """批量处理图像并保存标注结果"""
    os.makedirs(output_dir, exist_ok=True)
    image_paths = glob.glob(os.path.join(image_dir, "*.jpg")) + \
                  glob.glob(os.path.join(image_dir, "*.png"))
    for img_path in image_paths:
        results = model(img_path, conf=conf_thres)[0]
        # 保存带框的图像
        annotated = results.plot()
        out_path = os.path.join(output_dir, os.path.basename(img_path))
        cv2.imwrite(out_path, annotated)
        # 可选：保存检测标签到 txt 文件
        label_path = out_path.replace('.jpg', '.txt').replace('.png', '.txt')
        with open(label_path, 'w') as f:
            for box in results.boxes:
                x1, y1, x2, y2 = box.xyxy[0].tolist()
                cls_id = int(box.cls[0].item())
                conf = box.conf[0].item()
                f.write(f"{cls_id} {x1} {y1} {x2} {y2} {conf}\n")
    print(f"Detection results saved to {output_dir}")

# ========================== 7. 导出模型（用于边缘部署） ==========================
def export_model(model, format='onnx'):
    """导出模型为 ONNX/TensorRT 等格式"""
    model.export(format=format, imgsz=IMG_SIZE)
    print(f"Model exported to {format} format.")

# ========================== 8. 主程序 ==========================
if __name__ == "__main__":
    print("=== 南瓜叶片病害检测系统 (YOLO26) ===")

    # 1. 加载预训练模型
    model = load_model()

    # 2. 训练模型（如果有标注数据）
    print("\n开始训练...")
    train_results = train_model(model, DATA_YAML)

    # 3. 评估模型
    print("\n评估模型...")
    eval_metrics = evaluate_model(model, DATA_YAML)

    # 4. 单张图像测试
    test_image = "./sample/pumpkin_leaf.jpg"   # 替换为实际测试图像路径
    if os.path.exists(test_image):
        print("\n单张图像检测演示...")
        predict_single(model, test_image)

    # 5. 批量处理测试集
    test_image_dir = "./datasets/test/images"
    output_dir = "./results"
    if os.path.exists(test_image_dir):
        print(f"\n批量处理测试集，结果保存至 {output_dir}")
        predict_batch(model, test_image_dir, output_dir)

    # 6. 导出模型（可选）
    # export_model(model, format='onnx')

    print("\n系统运行完成。")

Bacterial Leaf Spot: 细菌性叶斑病

Downy Mildew: 霜霉病

Healthy Leaf: 健康叶片

Mosaic disease: 花叶病

Powdery_Mildew: 白粉病

path: D:/YOLO26/data/Pumpkin  # dataset root dir

names:
- Bacterial Leaf Spot
- Downy Mildew
- Healthy Leaf
- Mosaic disease
- Powdery_Mildew
nc: 5

test: ./test/images
train: ./train/images
val: ./valid/images

import warnings
warnings.filterwarnings('ignore')
from ultralytics import YOLO

if __name__ == '__main__':
    model = YOLO('ultralytics/cfg/models/26/yolo26.yaml')
    model.train(data='data/Pumpkin.yaml',
                cache=False,
                imgsz=640,
                epochs=200,
                batch=8,
                close_mosaic=10,
                device='0',
                optimizer='SGD', # using SGD
                project='runs/train',
                name='exp',
                )

YOLO26 summary (fused): 122 layers, 2,375,811 parameters, 0 gradients, 5.2 GFLOPs
                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100% ━━━━━━━━━━━━ 20/20 3.0it/s 6.6s
                   all        612       1115      0.801       0.63      0.729      0.478
   Bacterial Leaf Spot        121        252      0.727      0.595      0.682      0.409
          Downy Mildew        111        196      0.831      0.587      0.698      0.461
          Healthy Leaf        120        223      0.783      0.677      0.756      0.513
        Mosaic disease        121        208      0.855      0.601      0.722        0.5
        Powdery_Mildew        139        236      0.811       0.69      0.787      0.504

pip install --upgrade pip
pip install pyside6 -i https://mirror.baidu.com/pypi/simple