{"id":83408,"date":"2026-07-25T21:49:18","date_gmt":"2026-07-25T13:49:18","guid":{"rendered":"https:\/\/www.wsisp.com\/helps\/83408.html"},"modified":"2026-07-25T21:49:18","modified_gmt":"2026-07-25T13:49:18","slug":"%e5%bb%ba%e7%ad%91%e7%89%a9%e5%9b%be%e5%83%8f%e5%88%86%e5%89%b2%e8%af%86%e5%88%ab-%e8%87%aa%e5%8a%a8%e6%8f%90%e5%8f%96%e5%9f%8e%e5%b8%82%e5%bb%ba%e7%ad%91%e5%88%86%e5%b8%83%e5%9b%be-%e5%bb%ba","status":"publish","type":"post","link":"https:\/\/www.wsisp.com\/helps\/83408.html","title":{"rendered":"\u5efa\u7b51\u7269\u56fe\u50cf\u5206\u5272\u8bc6\u522b \u81ea\u52a8\u63d0\u53d6\u57ce\u5e02\u5efa\u7b51\u5206\u5e03\u56fe - \u5efa\u7b51\u5bc6\u5ea6\u5206\u6790\u3001\u57ce\u5e02\u6269\u5f20\u76d1\u6d4b \u65e0\u4eba\u673a \u5efa\u7b51\u7269\u623f\u5c4b\u68c0\u6d4b\u6570\u636e\u96c6 \u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u56fe\u50cf"},"content":{"rendered":"

yolo \u63a9\u819c\u53ca\u65e0\u63a9\u819c \u65e0\u4eba\u673a \u5efa\u7b51\u7269\u623f\u5c4b\u68c0\u6d4b\u6570\u636e\u96c6，\u5305\u542b\u5927\u91cf\u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u56fe\u50cf，\u56fe\u7247\u6e05\u6670\u6807\u6ce8，\u9002\u5408\u76ee\u6807\u68c0\u6d4b\u6a21\u578b\u8bad\u7ec3\u3002\u6570\u636e\u96c6\u5df2\u6807\u6ce8\u597d，\u53ef\u4ee5\u76f4\u63a5\u4f7f\u7528，\u652f\u6301\u591a\u79cd\u683c\u5f0f\u8f93\u51fa，\u4e00\u7c7b\u3002<\/h2>\n

\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n1
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n1
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n1
\n\"\u5728\u8fd9\u91cc\u63d2\u5165\u56fe\u7247\u63cf\u8ff0\"
\n\u60a8\u63d0\u4f9b\u7684\u6570\u636e\u96c6\u662f \u9ad8\u5206\u8fa8\u7387\u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u7684\u822a\u62cd\u56fe\u50cf，\u5305\u542b \u8fb9\u754c\u6846（Bounding Box）\u548c\u63a9\u819c（Mask）\u6807\u6ce8，\u9002\u7528\u4e8e \u76ee\u6807\u68c0\u6d4b\u4e0e\u5b9e\u4f8b\u5206\u5272\u4efb\u52a1\u3002=<\/p>\n

\n

📊 \u4e00\u3001\u6570\u636e\u96c6\u6838\u5fc3\u7279\u5f81<\/h3>\n\n\u7279\u6027\u63cf\u8ff0<\/tr>\n\n\n\n\n\n\n
\u56fe\u50cf\u6765\u6e90<\/td>\n\u822a\u62cd\/\u536b\u661f\u5f71\u50cf（\u5982\u65e0\u4eba\u673a\u3001\u536b\u661f\u9065\u611f）<\/td>\n<\/tr>\n
\u6807\u6ce8\u7c7b\u578b<\/td>\n\u2705 \u8fb9\u754c\u6846（YOLO \u683c\u5f0f）\u2705 \u5b9e\u4f8b\u63a9\u819c（Polygon \u6216 RLE）<\/td>\n<\/tr>\n
\u6807\u6ce8\u8d28\u91cf<\/td>\n\u7cbe\u786e\u5230\u5efa\u7b51\u8f6e\u5ed3，\u542b\u4e0d\u89c4\u5219\u5f62\u72b6\u3001\u906e\u6321\u3001\u591a\u5c42\u7ed3\u6784<\/td>\n<\/tr>\n
\u573a\u666f\u591a\u6837\u6027<\/td>\n\u57ce\u5e02\u5bc6\u96c6\u533a\u3001\u90ca\u533a\u3001\u5de5\u4e1a\u533a\u3001\u519c\u6751\u7b49<\/td>\n<\/tr>\n
\u652f\u6301\u683c\u5f0f<\/td>\nYOLOv5\/v8（.txt）\u3001COCO（JSON）\u3001Pascal VOC\u3001DeepLab \u7b49<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
\n

🌐 \u4e8c\u3001\u4e3b\u8981\u5e94\u7528\u9886\u57df<\/h3>\n

1. 🏙\ufe0f \u667a\u6167\u57ce\u5e02\u4e0e\u57ce\u5e02\u89c4\u5212<\/h4>\n
    \n
  • \u81ea\u52a8\u63d0\u53d6\u57ce\u5e02\u5efa\u7b51\u5206\u5e03\u56fe<\/li>\n
  • \u5efa\u7b51\u5bc6\u5ea6\u5206\u6790\u3001\u57ce\u5e02\u6269\u5f20\u76d1\u6d4b<\/li>\n
  • \u571f\u5730\u5229\u7528\u5206\u7c7b（\u4f4f\u5b85\/\u5546\u4e1a\/\u5de5\u4e1a）<\/li>\n
  • \u653f\u5e9c\u90e8\u95e8\u7528\u4e8e\u4eba\u53e3\u666e\u67e5\u3001\u623f\u4ea7\u767b\u8bb0<\/li>\n<\/ul>\n

    \u2705 \u793a\u4f8b：\u901a\u8fc7 AI \u5feb\u901f\u751f\u6210\u201c\u57ce\u5e02\u4e09\u7ef4\u6a21\u578b\u201d\u57fa\u7840\u6570\u636e<\/p>\n

    \n

    2. 🔧 \u5730\u7406\u4fe1\u606f\u7cfb\u7edf（GIS）\u4e0e\u9065\u611f<\/h4>\n
      \n
    • \u9ad8\u7cbe\u5ea6\u5730\u56fe\u66f4\u65b0（\u5982 Google Earth\u3001\u767e\u5ea6\u5730\u56fe）<\/li>\n
    • \u6570\u5b57\u5b6a\u751f\u57ce\u5e02\u5efa\u8bbe<\/li>\n
    • \u5730\u5f62\u5efa\u6a21\u4e0e\u7a7a\u95f4\u5206\u6790<\/li>\n
    • \u7ed3\u5408 LiDAR \u6570\u636e\u8fdb\u884c\u4e09\u7ef4\u91cd\u5efa<\/li>\n<\/ul>\n

      \u2705 \u793a\u4f8b：\u81ea\u52a8\u8bc6\u522b\u65b0\u5efa\u5efa\u7b51，\u8f85\u52a9\u6d4b\u7ed8\u516c\u53f8\u66f4\u65b0\u6570\u636e\u5e93<\/p>\n

      \n

      3. 🚧 \u707e\u5bb3\u8bc4\u4f30\u4e0e\u5e94\u6025\u7ba1\u7406<\/h4>\n
        \n
      • \u5730\u9707\u3001\u6d2a\u6c34\u540e\u5feb\u901f\u8bc4\u4f30\u5efa\u7b51\u7269\u635f\u6bc1\u60c5\u51b5<\/li>\n
      • \u8bc6\u522b\u5012\u584c\u3001\u503e\u659c\u3001\u88c2\u7f1d\u5efa\u7b51<\/li>\n
      • \u5e94\u6025\u6551\u63f4\u8def\u5f84\u89c4\u5212<\/li>\n
      • \u635f\u5931\u9884\u4f30（\u57fa\u4e8e\u5efa\u7b51\u6570\u91cf\u4e0e\u9762\u79ef）<\/li>\n<\/ul>\n

        \u2705 \u793a\u4f8b：\u707e\u540e 1 \u5c0f\u65f6\u5185\u5b8c\u6210\u5168\u57ce\u5efa\u7b51\u72b6\u6001\u8bc4\u4f30<\/p>\n

        \n
        \n

        5. 🛠\ufe0f \u81ea\u52a8\u9a7e\u9a76\u4e0e\u673a\u5668\u4eba\u5bfc\u822a<\/h4>\n
          \n
        • \u9ad8\u7cbe\u5730\u56fe\u6784\u5efa（\u5efa\u7b51\u8fb9\u754c\u4f5c\u4e3a\u8def\u7f51\u53c2\u8003）<\/li>\n
        • \u5ba4\u5916\u73af\u5883\u7406\u89e3（\u5efa\u7b51\u4e3a\u9759\u6001\u969c\u788d\u7269）<\/li>\n
        • \u65e0\u4eba\u673a\u81ea\u4e3b\u98de\u884c\u8def\u5f84\u89c4\u5212（\u907f\u5f00\u5efa\u7b51\u7fa4）<\/li>\n<\/ul>\n

          \u2705 \u793a\u4f8b：\u81ea\u52a8\u9a7e\u9a76\u8f66\u8f86\u5728\u590d\u6742\u57ce\u533a\u4e2d\u8bc6\u522b\u5efa\u7b51\u8fb9\u7f18\u4ee5\u907f\u969c<\/p>\n

          \n

          6. 🌱 \u751f\u6001\u73af\u5883\u4e0e\u53ef\u6301\u7eed\u53d1\u5c55<\/h4>\n
            \n
          • \u5efa\u7b51\u5c4b\u9876\u7eff\u5316\u76d1\u6d4b（\u5149\u4f0f\u677f vs \u7eff\u690d）<\/li>\n
          • \u57ce\u5e02\u70ed\u5c9b\u6548\u5e94\u5206\u6790（\u5efa\u7b51\u5bc6\u5ea6\u5f71\u54cd\u6e29\u5ea6）<\/li>\n
          • \u592a\u9633\u80fd\u6f5c\u529b\u8bc4\u4f30（\u5c4b\u9876\u9762\u79ef + \u671d\u5411）<\/li>\n<\/ul>\n

            \u2705 \u793a\u4f8b：\u653f\u5e9c\u63a8\u52a8\u201c\u7eff\u8272\u5c4b\u9876\u201d\u653f\u7b56，AI \u76d1\u6d4b\u6267\u884c\u60c5\u51b5<\/p>\n

            \n

            \u5e26\u63a9\u819c（Mask）\u4e0e\u65e0\u63a9\u819c（\u4ec5\u8fb9\u754c\u6846）\u7684\u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u6570\u636e\u96c6 \u7684\u5b8c\u6574\u8bad\u7ec3\u4ee3\u7801，\u652f\u6301 \u76ee\u6807\u68c0\u6d4b（Bounding Box） \u548c \u5b9e\u4f8b\u5206\u5272（Instance Segmentation） \u4e24\u79cd\u4efb\u52a1\u3002\u4f7f\u7528 Ultralytics YOLOv8 \u6846\u67b6（\u622a\u81f32025\u5e74\u6700\u7a33\u5b9a\u3001\u529f\u80fd\u6700\u5168\u7684\u5b98\u65b9YOLO\u5b9e\u73b0），\u5e76\u63d0\u4f9b\u4ece\u6570\u636e\u51c6\u5907\u5230\u6a21\u578b\u5bfc\u51fa\u7684\u5168\u6d41\u7a0b\u3002<\/p>\n

            \n

            📁 \u4e00\u3001\u9879\u76ee\u7ed3\u6784<\/h3>\n

            Building_Dataset\/
            \n\u251c\u2500\u2500 dataset.yaml # \u6570\u636e\u96c6\u914d\u7f6e\u6587\u4ef6<\/span>
            \n\u251c\u2500\u2500 train.py # \u8bad\u7ec3\u811a\u672c<\/span>
            \n\u251c\u2500\u2500 predict_demo.py # \u63a8\u7406\u6f14\u793a<\/span>
            \n\u251c\u2500\u2500 dataset\/
            \n\u2502 \u251c\u2500\u2500 images\/
            \n\u2502 \u2502 \u251c\u2500\u2500 train\/
            \n\u2502 \u2502 \u251c\u2500\u2500 val\/
            \n\u2502 \u2502 \u2514\u2500\u2500 test\/
            \n\u2502 \u2514\u2500\u2500 labels\/
            \n\u2502 \u251c\u2500\u2500 train\/ # YOLO\u683c\u5f0f：*.txt（\u542bbox + \u53ef\u9009polygon）<\/span>
            \n\u2502 \u251c\u2500\u2500 val\/
            \n\u2502 \u2514\u2500\u2500 test\/
            \n\u2514\u2500\u2500 requirements.txt<\/p>\n

            \u2705 \u652f\u6301\u4e24\u79cd\u6a21\u5f0f：<\/p>\n

              \n
            • \u4ec5\u68c0\u6d4b：labels\/*.txt \u4e2d\u6bcf\u884c\u4e3a class_id x y w h<\/li>\n
            • \u5b9e\u4f8b\u5206\u5272：labels\/*.txt \u4e2d\u6bcf\u884c\u4e3a class_id x1 y1 x2 y2 … xn yn（\u5f52\u4e00\u5316\u5750\u6807）<\/li>\n<\/ul>\n
              \n

              📦 \u4e8c\u3001\u4f9d\u8d56\u5b89\u88c5（requirements.txt）<\/h3>\n

              ultralytics==8.2.0
              \ntorch>=1.10.0
              \ntorchvision>=0.11.0
              \nopencv-python>=4.6.0
              \nnumpy>=1.21.0
              \nmatplotlib>=3.7.0
              \nscipy>=1.10.0<\/p>\n

              \u5b89\u88c5\u547d\u4ee4：<\/p>\n

              pip install<\/span> -r<\/span> requirements.txt<\/p>\n

              \n

              📄 \u4e09\u3001\u6570\u636e\u96c6\u914d\u7f6e\u6587\u4ef6 dataset.yaml<\/h3>\n

              \u25b6 \u573a\u666f1：\u4ec5\u76ee\u6807\u68c0\u6d4b（\u65e0\u63a9\u819c）<\/h4>\n

              # dataset.yaml (Detection Only)<\/span>
              \npath<\/span>:<\/span> .\/dataset
              \ntrain<\/span>:<\/span> images\/train
              \nval<\/span>:<\/span> images\/val
              \ntest<\/span>:<\/span> images\/test<\/p>\n

              nc<\/span>:<\/span> 1<\/span>
              \nnames<\/span>:<\/span> [<\/span>'building'<\/span>]<\/span><\/p>\n

              \u25b6 \u573a\u666f2：\u5b9e\u4f8b\u5206\u5272（\u5e26\u63a9\u819c）<\/h4>\n

              # dataset.yaml (Segmentation)<\/span>
              \npath<\/span>:<\/span> .\/dataset
              \ntrain<\/span>:<\/span> images\/train
              \nval<\/span>:<\/span> images\/val
              \ntest<\/span>:<\/span> images\/test<\/p>\n

              nc<\/span>:<\/span> 1<\/span>
              \nnames<\/span>:<\/span> [<\/span>'building'<\/span>]<\/span><\/p>\n

              # \u6ce8\u610f：YOLOv8-seg \u81ea\u52a8\u8bc6\u522b labels \u4e2d\u7684\u591a\u70b9\u5750\u6807\u4f5c\u4e3a mask<\/span><\/p>\n

              💡 \u6807\u6ce8\u793a\u4f8b（labels\/train\/001.txt）：<\/p>\n

              0 0.5 0.5 0.2 0.3 # \u68c0\u6d4b\u6a21\u5f0f
              \n0 0.4 0.4 0.6 0.4 0.6 0.6 0.4 0.6 # \u5206\u5272\u6a21\u5f0f（4\u4e2a\u70b9\u7684\u77e9\u5f62）<\/p>\n

              \n

              🚀 \u56db\u3001\u5b8c\u6574\u8bad\u7ec3\u4ee3\u7801 train.py<\/h3>\n

              # train.py<\/span>
              \nfrom<\/span> ultralytics import<\/span> YOLO
              \nimport<\/span> torch
              \nimport<\/span> os<\/p>\n

              def<\/span> main<\/span>(<\/span>)<\/span>:<\/span>
              \n device =<\/span> 'cuda'<\/span> if<\/span> torch.<\/span>cuda.<\/span>is_available(<\/span>)<\/span> else<\/span> 'cpu'<\/span>
              \n print<\/span>(<\/span>f"🚀 \u4f7f\u7528\u8bbe\u5907: <\/span>{<\/span>device}<\/span><\/span>"<\/span><\/span>)<\/span><\/p>\n

              # ==============================<\/span>
              \n # \u9009\u62e9\u4efb\u52a1\u7c7b\u578b<\/span>
              \n # ==============================<\/span>
              \n TASK =<\/span> "segment"<\/span> # \u53ef\u9009: "detect" \u6216 "segment"<\/span><\/p>\n

              if<\/span> TASK ==<\/span> "detect"<\/span>:<\/span>
              \n model_name =<\/span> "yolov8s.pt"<\/span> # \u76ee\u6807\u68c0\u6d4b\u6a21\u578b<\/span>
              \n project_name =<\/span> "building_detection"<\/span>
              \n elif<\/span> TASK ==<\/span> "segment"<\/span>:<\/span>
              \n model_name =<\/span> "yolov8s-seg.pt"<\/span> # \u5b9e\u4f8b\u5206\u5272\u6a21\u578b<\/span>
              \n project_name =<\/span> "building_segmentation"<\/span>
              \n else<\/span>:<\/span>
              \n raise<\/span> ValueError(<\/span>"TASK \u5fc5\u987b\u662f 'detect' \u6216 'segment'"<\/span>)<\/span><\/p>\n

              # \u52a0\u8f7d\u9884\u8bad\u7ec3\u6a21\u578b<\/span>
              \n model =<\/span> YOLO(<\/span>model_name)<\/span><\/p>\n

              # \u8bad\u7ec3\u53c2\u6570<\/span>
              \n results =<\/span> model.<\/span>train(<\/span>
              \n data=<\/span>'dataset.yaml'<\/span>,<\/span>
              \n epochs=<\/span>100<\/span>,<\/span>
              \n imgsz=<\/span>640<\/span>,<\/span> # \u5efa\u7b51\u7ec6\u8282\u4e30\u5bcc，\u5efa\u8bae \u2265640<\/span>
              \n batch=<\/span>16<\/span>,<\/span> # \u6839\u636e GPU \u8c03\u6574（32G \u663e\u5b58\u53ef\u8bbe 32）<\/span>
              \n name=<\/span>project_name,<\/span>
              \n device=<\/span>device,<\/span>
              \n patience=<\/span>25<\/span>,<\/span> # \u65e9\u505c<\/span>
              \n save=<\/span>True<\/span>,<\/span>
              \n plots=<\/span>True<\/span>,<\/span> # \u751f\u6210 PR \u66f2\u7ebf\u3001\u6df7\u6dc6\u77e9\u9635\u7b49<\/span>
              \n exist_ok=<\/span>False<\/span>,<\/span>
              \n workers=<\/span>8<\/span>,<\/span>
              \n optimizer=<\/span>'AdamW'<\/span>,<\/span>
              \n lr0=<\/span>0.01<\/span>,<\/span>
              \n lrf=<\/span>0.01<\/span>,<\/span>
              \n augment=<\/span>True<\/span>,<\/span> # \u542f\u7528 Mosaic\u3001HSV\u3001\u65cb\u8f6c\u7b49\u589e\u5f3a<\/span>
              \n degrees=<\/span>10.0<\/span>,<\/span> # \u00b110\u00b0 \u65cb\u8f6c（\u9002\u5408\u822a\u62cd）<\/span>
              \n translate=<\/span>0.1<\/span>,<\/span>
              \n scale=<\/span>0.5<\/span>,<\/span>
              \n flipud=<\/span>0.0<\/span>,<\/span> # \u822a\u62cd\u56fe\u901a\u5e38\u4e0d\u4e0a\u4e0b\u7ffb\u8f6c<\/span>
              \n fliplr=<\/span>0.5<\/span>,<\/span> # \u5de6\u53f3\u7ffb\u8f6c\u5b89\u5168<\/span>
              \n mosaic=<\/span>1.0<\/span>,<\/span>
              \n mixup=<\/span>0.1<\/span>
              \n )<\/span><\/p>\n

              # \u5728\u6d4b\u8bd5\u96c6\u4e0a\u8bc4\u4f30<\/span>
              \n metrics =<\/span> model.<\/span>val(<\/span>data=<\/span>'dataset.yaml'<\/span>,<\/span> split=<\/span>'test'<\/span>)<\/span>
              \n print<\/span>(<\/span>"\\\\n\u2705 \u6d4b\u8bd5\u96c6\u7ed3\u679c:"<\/span>)<\/span>
              \n if<\/span> TASK ==<\/span> "detect"<\/span>:<\/span>
              \n print<\/span>(<\/span>f" mAP@0.5: <\/span>{<\/span>metrics.<\/span>box.<\/span>map50:<\/span>.5f<\/span>}<\/span><\/span>"<\/span><\/span>)<\/span>
              \n print<\/span>(<\/span>f" mAP@0.5:0.95: <\/span>{<\/span>metrics.<\/span>box.<\/span>map<\/span>:<\/span>.5f<\/span>}<\/span><\/span>"<\/span><\/span>)<\/span>
              \n else<\/span>:<\/span>
              \n print<\/span>(<\/span>f" mAP@0.5 (Box): <\/span>{<\/span>metrics.<\/span>box.<\/span>map50:<\/span>.5f<\/span>}<\/span><\/span>"<\/span><\/span>)<\/span>
              \n print<\/span>(<\/span>f" mAP@0.5 (Mask): <\/span>{<\/span>metrics.<\/span>seg.<\/span>map50:<\/span>.5f<\/span>}<\/span><\/span>"<\/span><\/span>)<\/span><\/p>\n

              if<\/span> __name__ ==<\/span> '__main__'<\/span>:<\/span>
              \n main(<\/span>)<\/span><\/p>\n

              \u8fd0\u884c\u8bad\u7ec3：<\/p>\n

              python train.py<\/p>\n

              \n

              🔍 \u4e94\u3001\u63a8\u7406\u4e0e\u53ef\u89c6\u5316 predict_demo.py<\/h3>\n

              # predict_demo.py<\/span>
              \nfrom<\/span> ultralytics import<\/span> YOLO
              \nimport<\/span> cv2<\/p>\n

              # \u52a0\u8f7d\u8bad\u7ec3\u597d\u7684\u6a21\u578b<\/span>
              \nmodel =<\/span> YOLO(<\/span>'runs\/segment\/building_segmentation\/weights\/best.pt'<\/span>)<\/span> # \u6216 detect \u8def\u5f84<\/span><\/p>\n

              # \u5355\u5f20\u56fe\u50cf\u63a8\u7406<\/span>
              \nresults =<\/span> model.<\/span>predict(<\/span>
              \n source=<\/span>'test_image.jpg'<\/span>,<\/span>
              \n conf=<\/span>0.25<\/span>,<\/span>
              \n iou=<\/span>0.45<\/span>,<\/span>
              \n save=<\/span>True<\/span>,<\/span> # \u4fdd\u5b58\u5e26\u6807\u6ce8\u7684\u56fe\u50cf<\/span>
              \n show=<\/span>True<\/span> # \u5b9e\u65f6\u663e\u793a<\/span>
              \n)<\/span><\/p>\n

              # \u83b7\u53d6\u7ed3\u679c<\/span>
              \nfor<\/span> result in<\/span> results:<\/span>
              \n boxes =<\/span> result.<\/span>boxes # \u8fb9\u754c\u6846<\/span>
              \n masks =<\/span> result.<\/span>masks # \u63a9\u819c（\u82e5\u4e3a segmentation）<\/span>
              \n print<\/span>(<\/span>f"\u68c0\u6d4b\u5230 <\/span>{<\/span>len<\/span>(<\/span>boxes)<\/span>}<\/span><\/span> \u680b\u5efa\u7b51"<\/span><\/span>)<\/span>
              \n if<\/span> masks is<\/span> not<\/span> None<\/span>:<\/span>
              \n print<\/span>(<\/span>f"\u751f\u6210 <\/span>{<\/span>len<\/span>(<\/span>masks)<\/span>}<\/span><\/span> \u4e2a\u63a9\u819c"<\/span><\/span>)<\/span><\/p>\n

              \n

              📤 \u516d\u3001\u6a21\u578b\u5bfc\u51fa（\u7528\u4e8e\u90e8\u7f72）<\/h3>\n

              # export_model.py<\/span>
              \nfrom<\/span> ultralytics import<\/span> YOLO<\/p>\n

              model =<\/span> YOLO(<\/span>'runs\/segment\/building_segmentation\/weights\/best.pt'<\/span>)<\/span><\/p>\n

              # \u5bfc\u51fa\u4e3a ONNX（\u901a\u7528）<\/span>
              \nmodel.<\/span>export(<\/span>format<\/span>=<\/span>'onnx'<\/span>,<\/span> imgsz=<\/span>640<\/span>,<\/span> simplify=<\/span>True<\/span>)<\/span><\/p>\n

              # \u5bfc\u51fa\u4e3a TensorRT（Jetson \/ NVIDIA GPU）<\/span>
              \nmodel.<\/span>export(<\/span>format<\/span>=<\/span>'engine'<\/span>,<\/span> imgsz=<\/span>640<\/span>,<\/span> half=<\/span>True<\/span>,<\/span> device=<\/span>0<\/span>)<\/span><\/p>\n

              # \u5bfc\u51fa\u4e3a OpenVINO（Intel CPU）<\/span>
              \nmodel.<\/span>export(<\/span>format<\/span>=<\/span>'openvino'<\/span>,<\/span> imgsz=<\/span>640<\/span>)<\/span><\/p>\n

              \n

              📊 \u4e03\u3001\u9884\u671f\u6027\u80fd（\u57fa\u4e8e\u7c7b\u4f3c\u6570\u636e\u96c6\u7ecf\u9a8c）<\/h3>\n\n\u4efb\u52a1\u6a21\u578b\u9884\u671f mAP@0.5\u63a8\u7406\u901f\u5ea6（RTX 3090）<\/tr>\n\n\n\n
              \u76ee\u6807\u68c0\u6d4b<\/td>\nYOLOv8s<\/td>\n0.92\u20130.96<\/td>\n120 FPS<\/td>\n<\/tr>\n
              \u5b9e\u4f8b\u5206\u5272<\/td>\nYOLOv8s-seg<\/td>\n0.88\u20130.93<\/td>\n45 FPS<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

              \u2705 \u5efa\u7b51\u8f6e\u5ed3\u6e05\u6670\u3001\u80cc\u666f\u7b80\u5355，mAP > 0.9 \u662f\u5b8c\u5168\u53ef\u5b9e\u73b0\u7684<\/p>\n

              \n
              \n","protected":false},"excerpt":{"rendered":"

              yolo \u63a9\u819c\u53ca\u65e0\u63a9\u819c \u65e0\u4eba\u673a \u5efa\u7b51\u7269\u623f\u5c4b\u68c0\u6d4b\u6570\u636e\u96c6，\u5305\u542b\u5927\u91cf\u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u56fe\u50cf，\u56fe\u7247\u6e05\u6670\u6807\u6ce8，\u9002\u5408\u76ee\u6807\u68c0\u6d4b\u6a21\u578b\u8bad\u7ec3\u3002\u6570\u636e\u96c6\u5df2\u6807\u6ce8\u597d，\u53ef\u4ee5\u76f4\u63a5\u4f7f\u7528，\u652f\u6301\u591a\u79cd\u683c\u5f0f\u8f93\u51fa，\u4e00\u7c7b\u3002111\u60a8\u63d0\u4f9b\u7684\u6570\u636e\u96c6\u662f \u9ad8\u5206\u8fa8\u7387\u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u7684\u822a\u62cd\u56fe\u50cf，\u5305\u542b \u8fb9\u754c\u6846（Bounding Box）\u548c\u63a9\u819c（Mask）\u6807\u6ce8，\u9002\u7528\u4e8e \u76ee\u6807\u68c0\u6d4b\u4e0e\u5b9e\u4f8b<\/p>\n","protected":false},"author":2,"featured_media":83404,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[156,50,3267,86,523],"topic":[],"class_list":["post-83408","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-server","tag-yolo","tag-50","tag-3267","tag-86","tag-523"],"yoast_head":"\n\u5efa\u7b51\u7269\u56fe\u50cf\u5206\u5272\u8bc6\u522b \u81ea\u52a8\u63d0\u53d6\u57ce\u5e02\u5efa\u7b51\u5206\u5e03\u56fe - \u5efa\u7b51\u5bc6\u5ea6\u5206\u6790\u3001\u57ce\u5e02\u6269\u5f20\u76d1\u6d4b \u65e0\u4eba\u673a \u5efa\u7b51\u7269\u623f\u5c4b\u68c0\u6d4b\u6570\u636e\u96c6 \u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u56fe\u50cf - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.wsisp.com\/helps\/83408.html\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"\u5efa\u7b51\u7269\u56fe\u50cf\u5206\u5272\u8bc6\u522b \u81ea\u52a8\u63d0\u53d6\u57ce\u5e02\u5efa\u7b51\u5206\u5e03\u56fe - \u5efa\u7b51\u5bc6\u5ea6\u5206\u6790\u3001\u57ce\u5e02\u6269\u5f20\u76d1\u6d4b \u65e0\u4eba\u673a \u5efa\u7b51\u7269\u623f\u5c4b\u68c0\u6d4b\u6570\u636e\u96c6 \u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u56fe\u50cf - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"og:description\" content=\"yolo \u63a9\u819c\u53ca\u65e0\u63a9\u819c \u65e0\u4eba\u673a \u5efa\u7b51\u7269\u623f\u5c4b\u68c0\u6d4b\u6570\u636e\u96c6，\u5305\u542b\u5927\u91cf\u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u56fe\u50cf，\u56fe\u7247\u6e05\u6670\u6807\u6ce8，\u9002\u5408\u76ee\u6807\u68c0\u6d4b\u6a21\u578b\u8bad\u7ec3\u3002\u6570\u636e\u96c6\u5df2\u6807\u6ce8\u597d，\u53ef\u4ee5\u76f4\u63a5\u4f7f\u7528，\u652f\u6301\u591a\u79cd\u683c\u5f0f\u8f93\u51fa，\u4e00\u7c7b\u3002111\u60a8\u63d0\u4f9b\u7684\u6570\u636e\u96c6\u662f \u9ad8\u5206\u8fa8\u7387\u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u7684\u822a\u62cd\u56fe\u50cf，\u5305\u542b \u8fb9\u754c\u6846（Bounding Box）\u548c\u63a9\u819c（Mask）\u6807\u6ce8，\u9002\u7528\u4e8e \u76ee\u6807\u68c0\u6d4b\u4e0e\u5b9e\u4f8b\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.wsisp.com\/helps\/83408.html\" \/>\n<meta property=\"og:site_name\" content=\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"article:published_time\" content=\"2026-07-25T13:49:18+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/07\/20260725134916-6a64bedc4759b.png\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"3 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/83408.html\",\"url\":\"https:\/\/www.wsisp.com\/helps\/83408.html\",\"name\":\"\u5efa\u7b51\u7269\u56fe\u50cf\u5206\u5272\u8bc6\u522b \u81ea\u52a8\u63d0\u53d6\u57ce\u5e02\u5efa\u7b51\u5206\u5e03\u56fe - \u5efa\u7b51\u5bc6\u5ea6\u5206\u6790\u3001\u57ce\u5e02\u6269\u5f20\u76d1\u6d4b \u65e0\u4eba\u673a \u5efa\u7b51\u7269\u623f\u5c4b\u68c0\u6d4b\u6570\u636e\u96c6 \u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u56fe\u50cf - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"isPartOf\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\"},\"datePublished\":\"2026-07-25T13:49:18+00:00\",\"dateModified\":\"2026-07-25T13:49:18+00:00\",\"author\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/83408.html#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.wsisp.com\/helps\/83408.html\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/83408.html#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/www.wsisp.com\/helps\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"\u5efa\u7b51\u7269\u56fe\u50cf\u5206\u5272\u8bc6\u522b \u81ea\u52a8\u63d0\u53d6\u57ce\u5e02\u5efa\u7b51\u5206\u5e03\u56fe - \u5efa\u7b51\u5bc6\u5ea6\u5206\u6790\u3001\u57ce\u5e02\u6269\u5f20\u76d1\u6d4b \u65e0\u4eba\u673a \u5efa\u7b51\u7269\u623f\u5c4b\u68c0\u6d4b\u6570\u636e\u96c6 \u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u56fe\u50cf\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\",\"url\":\"https:\/\/www.wsisp.com\/helps\/\",\"name\":\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"description\":\"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"contentUrl\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"caption\":\"admin\"},\"sameAs\":[\"http:\/\/wp.wsisp.com\"],\"url\":\"https:\/\/www.wsisp.com\/helps\/author\/admin\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"\u5efa\u7b51\u7269\u56fe\u50cf\u5206\u5272\u8bc6\u522b \u81ea\u52a8\u63d0\u53d6\u57ce\u5e02\u5efa\u7b51\u5206\u5e03\u56fe - \u5efa\u7b51\u5bc6\u5ea6\u5206\u6790\u3001\u57ce\u5e02\u6269\u5f20\u76d1\u6d4b \u65e0\u4eba\u673a \u5efa\u7b51\u7269\u623f\u5c4b\u68c0\u6d4b\u6570\u636e\u96c6 \u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u56fe\u50cf - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.wsisp.com\/helps\/83408.html","og_locale":"zh_CN","og_type":"article","og_title":"\u5efa\u7b51\u7269\u56fe\u50cf\u5206\u5272\u8bc6\u522b \u81ea\u52a8\u63d0\u53d6\u57ce\u5e02\u5efa\u7b51\u5206\u5e03\u56fe - \u5efa\u7b51\u5bc6\u5ea6\u5206\u6790\u3001\u57ce\u5e02\u6269\u5f20\u76d1\u6d4b \u65e0\u4eba\u673a \u5efa\u7b51\u7269\u623f\u5c4b\u68c0\u6d4b\u6570\u636e\u96c6 \u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u56fe\u50cf - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","og_description":"yolo \u63a9\u819c\u53ca\u65e0\u63a9\u819c \u65e0\u4eba\u673a \u5efa\u7b51\u7269\u623f\u5c4b\u68c0\u6d4b\u6570\u636e\u96c6，\u5305\u542b\u5927\u91cf\u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u56fe\u50cf，\u56fe\u7247\u6e05\u6670\u6807\u6ce8，\u9002\u5408\u76ee\u6807\u68c0\u6d4b\u6a21\u578b\u8bad\u7ec3\u3002\u6570\u636e\u96c6\u5df2\u6807\u6ce8\u597d，\u53ef\u4ee5\u76f4\u63a5\u4f7f\u7528，\u652f\u6301\u591a\u79cd\u683c\u5f0f\u8f93\u51fa，\u4e00\u7c7b\u3002111\u60a8\u63d0\u4f9b\u7684\u6570\u636e\u96c6\u662f \u9ad8\u5206\u8fa8\u7387\u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u7684\u822a\u62cd\u56fe\u50cf，\u5305\u542b \u8fb9\u754c\u6846（Bounding Box）\u548c\u63a9\u819c（Mask）\u6807\u6ce8，\u9002\u7528\u4e8e \u76ee\u6807\u68c0\u6d4b\u4e0e\u5b9e\u4f8b","og_url":"https:\/\/www.wsisp.com\/helps\/83408.html","og_site_name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","article_published_time":"2026-07-25T13:49:18+00:00","og_image":[{"url":"https:\/\/www.wsisp.com\/helps\/wp-content\/uploads\/2026\/07\/20260725134916-6a64bedc4759b.png"}],"author":"admin","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"3 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.wsisp.com\/helps\/83408.html","url":"https:\/\/www.wsisp.com\/helps\/83408.html","name":"\u5efa\u7b51\u7269\u56fe\u50cf\u5206\u5272\u8bc6\u522b \u81ea\u52a8\u63d0\u53d6\u57ce\u5e02\u5efa\u7b51\u5206\u5e03\u56fe - \u5efa\u7b51\u5bc6\u5ea6\u5206\u6790\u3001\u57ce\u5e02\u6269\u5f20\u76d1\u6d4b \u65e0\u4eba\u673a \u5efa\u7b51\u7269\u623f\u5c4b\u68c0\u6d4b\u6570\u636e\u96c6 \u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u56fe\u50cf - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","isPartOf":{"@id":"https:\/\/www.wsisp.com\/helps\/#website"},"datePublished":"2026-07-25T13:49:18+00:00","dateModified":"2026-07-25T13:49:18+00:00","author":{"@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41"},"breadcrumb":{"@id":"https:\/\/www.wsisp.com\/helps\/83408.html#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.wsisp.com\/helps\/83408.html"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.wsisp.com\/helps\/83408.html#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/www.wsisp.com\/helps"},{"@type":"ListItem","position":2,"name":"\u5efa\u7b51\u7269\u56fe\u50cf\u5206\u5272\u8bc6\u522b \u81ea\u52a8\u63d0\u53d6\u57ce\u5e02\u5efa\u7b51\u5206\u5e03\u56fe - \u5efa\u7b51\u5bc6\u5ea6\u5206\u6790\u3001\u57ce\u5e02\u6269\u5f20\u76d1\u6d4b \u65e0\u4eba\u673a \u5efa\u7b51\u7269\u623f\u5c4b\u68c0\u6d4b\u6570\u636e\u96c6 \u57ce\u5e02\u5efa\u7b51\u5c4b\u9876\u56fe\u50cf"}]},{"@type":"WebSite","@id":"https:\/\/www.wsisp.com\/helps\/#website","url":"https:\/\/www.wsisp.com\/helps\/","name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","description":"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41","name":"admin","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/","url":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","contentUrl":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","caption":"admin"},"sameAs":["http:\/\/wp.wsisp.com"],"url":"https:\/\/www.wsisp.com\/helps\/author\/admin"}]}},"_links":{"self":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/83408","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/comments?post=83408"}],"version-history":[{"count":0,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/83408\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media\/83404"}],"wp:attachment":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media?parent=83408"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/categories?post=83408"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/tags?post=83408"},{"taxonomy":"topic","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/topic?post=83408"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}