企业官网建设流程全解析

YOLO扑克牌检测数据集(极高精度)

1

数据集详情

• 检测类别：
  10 Diamonds
  10 Hearts
  10 Spades
  10 Trefoils
  2 Diamonds
  2 Hearts
  2 Spades
  2 Trefoils
  3 Diamonds
  3 Hearts
  3 Spades
  3 Trefoils
  4 Diamonds
  4 Hearts
  4 Spades
  4 Trefoils
  5 Diamonds
  5 Hearts
  5 Spades
  5 Trefoils
  59
  6 Diamonds
  6 Hearts
  6 Spades
  6 Trefoils
  7 Diamonds
  7 Hearts
  7 Spades
  7 Trefoils
  8 Diamonds
  8 Hearts
  8 Spades
  8 Trefoils
  9 Diamonds
  9 Hearts
  9 Spades
  9 Trefoils
  A Diamonds
  A Hearts
  A Spades
  A Trefoils
  J Diamonds
  J Hearts
  J Spades
  J Trefoils
  K Diamonds
  K Hearts
  K Spades
  K Trefoils
  Q Diamonds
  Q Hearts
  Q Spades
  Q Trefoils
  

• 图片总量：1285张

• 训练集：964张

• 验证集：193张

• 测试集：128张

• 标注状态：全部已标注，即拿即用

• 数据格式：支持YOLO格式及其他格式（可按需提供）

YOLO扑克牌检测数据集

一、数据集信息表

二、核心标签

#扑克牌检测 #YOLO数据集 #棋牌识别 #目标检测 #AI视觉训练数据集

三、数据集目录结构

poker_dataset/ ├── images/ │ ├── train/ │ ├── val/ │ └── test/ ├── labels/ │ ├── train/ │ ├── val/ │ └── test/ └── poker.yaml

四、数据集配置文件poker.yaml

path:./poker_datasettrain:images/trainval:images/valtest:images/testnc:53names:0:10 Diamonds1:10 Hearts2:10 Spades3:10 Trefoils4:2 Diamonds5:2 Hearts6:2 Spades7:2 Trefoils8:3 Diamonds9:3 Hearts10:3 Spades11:3 Trefoils12:4 Diamonds13:4 Hearts14:4 Spades15:4 Trefoils16:5 Diamonds17:5 Hearts18:5 Spades19:5 Trefoils20:5921:6 Diamonds22:6 Hearts23:6 Spades24:6 Trefoils25:7 Diamonds26:7 Hearts27:7 Spades28:7 Trefoils29:8 Diamonds30:8 Hearts31:8 Spades32:8 Trefoils33:9 Diamonds34:9 Hearts35:9 Spades36:9 Trefoils37:A Diamonds38:A Hearts39:A Spades40:A Trefoils41:J Diamonds42:J Hearts43:J Spades44:J Trefoils45:K Diamonds46:K Hearts47:K Spades48:K Trefoils49:Q Diamonds50:Q Hearts51:Q Spades52:Q Trefoils

五、环境依赖安装

pipinstallultralytics torch opencv-python

六、模型训练代码train_poker.py

fromultralyticsimportYOLOdefmain():# 加载预训练模型，支持 YOLOv8 / YOLOv11 等版本model=YOLO("yolov8s.pt")# 启动训练train_results=model.train(data="./poker_dataset/poker.yaml",epochs=120,imgsz=640,batch=16,device=0,# 无GPU则改为 device="cpu"workers=4,patience=20,project="runs/train",name="poker_detect",exist_ok=True,pretrained=True)print("训练完成，权重保存路径：runs/train/poker_detect/weights")# 测试集评估metrics=model.val(data="./poker_dataset/poker.yaml",split="test")print(f"测试集 mAP@50:{metrics.box.map50:.4f}")if__name__=="__main__":main()

七、推理测试代码predict_poker.py

fromultralyticsimportYOLOimportcv2# 加载训练完成的最优权重model=YOLO("runs/train/poker_detect/weights/best.pt")if__name__=="__main__":img_path="test_poker.jpg"# 执行检测并保存结果results=model.predict(source=img_path,conf=0.25,iou=0.45,save=True)# 可视化展示result_img=results[0].plot()cv2.imshow("Poker Detection",result_img)cv2.waitKey(0)cv2.destroyAllWindows()