{"id":20442,"date":"2024-11-29T16:32:21","date_gmt":"2024-11-29T08:32:21","guid":{"rendered":"http:\/\/192.168.10.115\/?p=20442"},"modified":"2024-11-29T16:32:21","modified_gmt":"2024-11-29T08:32:21","slug":"2024-11-29-%e5%95%86%e6%b1%a4%e5%92%8c%e6%b5%99%e6%b1%9f%e5%a4%a7%e5%ad%a6%e5%bc%ba%e5%bc%ba%e8%81%94%e5%90%88%ef%bc%9a%e5%a6%82%e4%bd%95%e9%92%88%e5%af%b9%e5%9f%ba%e4%ba%8e%e6%b7%b1%e5%ba%a6%e5%ad%a6","status":"publish","type":"post","link":"http:\/\/222.128.65.89:18086\/index.php\/2024\/11\/29\/20442\/","title":{"rendered":"2024-11-29 \u5546\u6c64\u548c\u6d59\u6c5f\u5927\u5b66\u5f3a\u5f3a\u8054\u5408\uff1a\u5982\u4f55\u9488\u5bf9\u57fa\u4e8e\u6df1\u5ea6\u5b66\u4e60\u7684SLAM\u6784\u5efa\u7075\u6d3b\u4e14\u6a21\u5757\u5316\u6846\u67b6"},"content":{"rendered":"\n

\u25cf \u8bba\u6587\u94fe\u63a5\uff1a<\/em>https:\/\/arxiv.org\/pdf\/2410.23690<\/a><\/p>\n\n\n\n

\u25cf \u4ee3\u7801\u94fe\u63a5\uff1a<\/em>GitHub – openxrlab\/xrdslam: Platform for Deep Learning based SLAM<\/a><\/p>\n\n\n\n

1.\u6458\u8981<\/p>\n\n\n\n

\u672c\u6587\u63d0\u51fa\u4e86\u4e00\u79cd\u7075\u6d3b\u7684SLAM\u6846\u67b6\u2014\u2014XRDSLAM\u3002\u5b83\u91c7\u7528\u6a21\u5757\u5316\u4ee3\u7801\u8bbe\u8ba1\u548c\u591a\u8fdb\u7a0b\u8fd0\u884c\u673a\u5236\uff0c\u4ece\u800c\u63d0\u4f9b\u4e86\u9ad8\u5ea6\u53ef\u590d\u7528\u7684\u57fa\u7840\u6a21\u5757\uff0c\u4f8b\u5982\u7edf\u4e00\u7684\u6570\u636e\u96c6\u7ba1\u7406\u3001\u4e09\u7ef4\u53ef\u89c6\u5316\u3001\u7b97\u6cd5\u914d\u7f6e\u548c\u5ea6\u91cf\u8bc4\u4f30\u3002\u8fd9\u53ef\u4ee5\u5e2e\u52a9\u5f00\u53d1\u8005\u5feb\u901f\u6784\u5efa\u4e00\u4e2a\u5b8c\u6574\u7684SLAM\u7cfb\u7edf\uff0c\u7075\u6d3b\u7ec4\u5408\u4e0d\u540c\u7684\u7b97\u6cd5\u6a21\u5757\u5e76\u4e14\u8fdb\u884c\u6807\u51c6\u5316\u7684\u57fa\u51c6\u6d4b\u8bd5\uff0c\u4ee5\u6bd4\u8f83\u7cbe\u5ea6\u548c\u6548\u7387\u3002\u5728\u8be5\u6846\u67b6\u5185\uff0c\u96c6\u6210\u4e86\u82e5\u5e72\u79cd\u4e0d\u540c\u7c7b\u578b\u7684\u6700\u5148\u8fdbSLAM\u7b97\u6cd5\uff0c\u5305\u62ec\u57fa\u4e8eNeRF\u548c3DGS\u7684SLAM\uff08\u751a\u81f3\u91cc\u7a0b\u8ba1\u6216\u8005\u91cd\u5efa\u7b97\u6cd5\uff09\uff0c\u5176\u5c55\u73b0\u4e86\u7075\u6d3b\u6027\u548c\u53ef\u6269\u5c55\u6027\u3002\u672c\u6587\u8fd8\u5bf9\u8fd9\u4e9b\u96c6\u6210\u7684\u7b97\u6cd5\u8fdb\u884c\u7efc\u5408\u6bd4\u8f83\u548c\u8bc4\u4f30\uff0c\u5206\u6790\u4e86\u6bcf\u79cd\u7b97\u6cd5\u7684\u7279\u70b9\u3002\u6700\u540e\uff0c\u672c\u6587\u5c06\u6240\u6709\u4ee3\u7801\u3001\u914d\u7f6e\u548c\u6570\u636e\u5f00\u6e90\uff0c\u65e8\u5728\u4fc3\u8fdb\u5f00\u6e90\u751f\u6001\u7cfb\u7edf\u4e2dSLAM\u6280\u672f\u7684\u5e7f\u6cdb\u7814\u7a76\u548c\u53d1\u5c55\u3002<\/p>\n\n\n\n

2.\u4ecb\u7ecd<\/p>\n\n\n\n

SLAM\uff08\u540c\u65f6\u5b9a\u4f4d\u548c\u5efa\u56fe\uff09\u6280\u672f\u5728\u81ea\u52a8\u9a7e\u9a76\u3001\u673a\u5668\u4eba\u548cAR\/VR\u7b49\u9886\u57df\u4e2d\u53d1\u6325\u7740\u81f3\u5173\u91cd\u8981\u7684\u4f5c\u7528\u3002\u901a\u8fc7\u5206\u6790\u548c\u878d\u5408\u4f20\u611f\u5668\u6570\u636e\uff0c\u5b9e\u73b0\u4e86\u8bbe\u5907\u7684\u7cbe\u786e6DoF\uff08\u81ea\u7531\u5ea6\uff09\u5b9a\u4f4d\uff0c\u5e76\u4e14\u5b9e\u65f6\u6784\u5efa\u73af\u5883\u5730\u56fe\uff0c\u4e3a\u81ea\u52a8\u9a7e\u9a76\u3001\u673a\u5668\u4eba\u5bfc\u822a\u548cAR\/VR\u7b49\u5e94\u7528\u63d0\u4f9b\u4e86\u57fa\u7840\u529f\u80fd\u3002\u5730\u56fe\u8868\u793a\u7684\u5f62\u5f0f\u5bf9\u4e8eSLAM\u7b97\u6cd5\u7684\u8bbe\u8ba1\u662f\u81f3\u5173\u91cd\u8981\u7684\u3002\u5c3d\u7ba1\u4f20\u7edf\u7684SLAM\u7b97\u6cd5<\/a>\u901a\u5e38\u57fa\u4e8e\u7a00\u758f\u6216\u8005\u534a\u7a20\u5bc6\u7684\u70b9\u4e91\u6765\u8868\u8fbe\u573a\u666f\u5e76\u4e14\u901a\u8fc7\u624b\u5199\u89c4\u5219\u6765\u7ba1\u7406\u5730\u56fe\uff0c\u4f46\u662f\u8fd9\u9650\u5236\u4e86SLAM\u6280\u672f\u7684\u53d1\u5c55\u3002<\/p>\n\n\n\n

\u8fd1\u5e74\u6765\uff0c\u795e\u7ecf\u8f90\u5c04\u573a<\/a>\uff08NeRF\uff09\u548c3D Gaussian Splatting\uff083DGS\uff09\u7b49\u6280\u672f\u5728\u65b0\u578b\u89c6\u56fe\u5408\u6210\u65b9\u9762\u53d6\u5f97\u4e86\u91cd\u5927\u8fdb\u5c55\u3002\u57fa\u4e8e\u5b66\u4e60\u7684\u573a\u666f\u8868\u793a\u5229\u7528\u4e86\u53ef\u5fae\u5206\u6805\u683c\u5316\u3001\u8fde\u7eed\u5efa\u6a21\u3001\u566a\u58f0\u5904\u7406\u548c\u7a7a\u6d1e\u586b\u5145\u7b49\u4f18\u52bf\uff0c\u4e3aSLAM\u9886\u57df\u6ce8\u5165\u4e86\u65b0\u7684\u80fd\u91cf\u3002\u4e00\u4e9b\u6700\u8fd1\u7684\u5de5\u4f5c\u8bd5\u56fe\u5c06NeRF\u548c3DGS\u52a0\u5165\u5230SLAM\u7cfb\u7edf\u4e2d\u3002\u4e00\u4e9b\u65b9\u6cd5\u76f4\u63a5\u4f7f\u7528RGB-D\u6570\u636e\u4f5c\u4e3a\u989c\u8272\u548c\u6df1\u5ea6\u76d1\u7763\u6765\u8bad\u7ec3\u7f51\u7edc\u53c2\u6570\u5e76\u4e14\u4f18\u5316\u4f4d\u59ff\u3002\u5176\u5b83\u65b9\u6cd5\u4ec5\u4f9d\u8d56\u4e8eRGB\u6570\u636e\uff0c\u901a\u8fc7\u52a0\u5165\u6b63\u5219\u5316\u7ea6\u675f\u6765\u4f18\u5316\u7f51\u7edc\u548c\u4f4d\u59ff\u3002\u4e00\u4e9b\u5de5\u4f5c\u8fd8\u901a\u8fc7\u5f15\u5165ICP\u548cORB-SLAM\u7b49\u6280\u672f\uff0c\u5c06\u65b0\u6280\u672f\u4e0e\u8fd9\u4e9b\u4f20\u7edf\u7684SLAM\u6280\u672f\u76f8\u7ed3\u5408\u3002\u6b64\u5916\uff0c\u8fd8\u6709\u4e00\u4e9b\u5de5\u4f5c\u8bd5\u56fe\u96c6\u6210\u989d\u5916\u7684\u7f51\u7edc\u6765\u63d0\u4f9b\u5148\u9a8c\u4fe1\u606f\uff08\u4f8b\u5982\u6df1\u5ea6\u5148\u9a8c<\/a>\u3001\u5149\u6d41\u548c\u6cd5\u5411\u91cf\uff09\uff0c\u65e8\u5728\u63d0\u9ad8\u8bad\u7ec3\u901f\u5ea6\u5e76\u4e14\u6539\u8fdb\u91cd\u5efa\u8d28\u91cf\u3002<\/p>\n\n\n\n

\u7136\u800c\uff0c\u968f\u7740\u65b0\u8bba\u6587\u548c\u6280\u672f\u7684\u4e0d\u65ad\u6d8c\u73b0\uff0c\u8ddf\u8e2a\u8fdb\u5c55\u548c\u96c6\u6210\u4ee3\u7801\u53d8\u5f97\u8d8a\u6765\u8d8a\u5177\u6709\u6311\u6218\u6027\u3002\u5f88\u591a\u7814\u7a76\u5b9e\u73b0\u5206\u6563\u4e8e\u72ec\u7acb\u7684\u4ee3\u7801\u4ed3\u5e93\u4e2d\uff0c\u7f3a\u4e4f\u7edf\u4e00\u7684\u5f00\u53d1\u548c\u8bc4\u4f30\u8fc7\u7a0b\u3002\u8fd9\u4f7f\u5f97\u516c\u6b63\u4e14\u5168\u9762\u7684\u6bd4\u8f83\u53d8\u5f97\u975e\u5e38\u56f0\u96be\u3002\u6b64\u5916\uff0cSLAM\u7cfb\u7edf\u7684\u590d\u6742\u5ea6\u4e3a\u4ece\u5934\u5f00\u59cb\u6784\u5efa\u5b8c\u6574\u7684SLAM\u7cfb\u7edf\u5e26\u6765\u4e86\u5f88\u9ad8\u7684\u6280\u672f\u969c\u788d\u548c\u5f00\u53d1\u6210\u672c\u3002\u53d7\u5230NeRFStudio\u5728NeRF\u9886\u57df\u4e2d\u7684\u6210\u529f\u542f\u53d1\uff0c\u672c\u6587\u5f15\u5165\u4e86XRDSLAM\uff0c\u8fd9\u662f\u4e00\u79cd\u7528\u4e8e\u57fa\u4e8e\u6df1\u5ea6\u5b66\u4e60SLAM\u7684\u53ef\u6269\u5c55\u4e14\u591a\u529f\u80fd\u7684\u6846\u67b6\u3002\u672c\u9879\u5de5\u4f5c\u7684\u4e3b\u8981\u8d21\u732e\u5982\u4e0b\uff1a<\/p>\n\n\n\n

\u6a21\u5757\u5316\u8bbe\u8ba1<\/a><\/strong><\/p>\n\n\n\n

\uff1a<\/strong>XRDSLAM\u91c7\u7528\u591a\u8fdb\u7a0b\u673a\u5236\u548c\u6a21\u5757\u5316\u4ee3\u7801\u8bbe\u8ba1\uff0c\u4ece\u800c\u5c06\u8ddf\u8e2a\u3001\u5efa\u56fe\u548c\u53ef\u89c6\u5316\u8fc7\u7a0b\u89e3\u8026\u3002\u8be5\u6846\u67b6\u63d0\u4f9b\u4e86\u6a21\u5757\u5316\u6570\u636e\u8f93\u5165\/\u8f93\u51fa\u3001\u914d\u7f6e\u89e3\u6790\u3001\u53ef\u89c6\u5316\u548c\u7ed3\u679c\u5bfc\u51fa\u529f\u80fd\u6a21\u5757\uff0c\u5141\u8bb8\u7528\u4e8e\u7075\u6d3b\u7ec4\u5408\u548c\u66ff\u6362\u4e0d\u540c\u7684\u7b97\u6cd5\u6a21\u5757\uff0c\u4ee5\u5b9e\u73b0\u5feb\u901f\u8fed\u4ee3\u548c\u4f18\u5316\u3002<\/p>\n\n\n\n

\u7edf\u4e00\u7684\u6d41\u7a0b\uff1a<\/strong>\u8be5\u6846\u67b6\u63d0\u4f9b\u4e86\u4e00\u4e2a\u5177\u6709\u53ef\u590d\u7528\u7ec4\u4ef6\u7684\u7edf\u4e00SLAM\u5f00\u53d1\u8fc7\u7a0b\uff0c\u4f7f\u5f00\u53d1\u8005\u80fd\u591f\u8f7b\u677e\u5730\u6784\u5efa\u5b8c\u6574\u7684SLAM\u7b97\u6cd5\u3002\u5b83\u8fd8\u6807\u51c6\u5316\u4e86\u6570\u636e\u5904\u7406\u548c\u8bc4\u4f30\uff0c\u5b9e\u73b0\u4e86\u4e0d\u540cSLAM\u7b97\u6cd5\u4e4b\u95f4\u516c\u6b63\u4e14\u6709\u6548\u7684\u8bc4\u4f30\u3002\u5177\u4f53\u800c\u8a00\uff0c\u672c\u6587\u8bbe\u8ba1\u4e86\u4e00\u79cd\u4e13\u95e8\u7684Model\u7c7b\u6765\u7ba1\u7406\u6df1\u5ea6\u5b66\u4e60\u6a21\u578b\u7684\u6240\u6709\u64cd\u4f5c\uff0c\u4f7f\u5f00\u53d1\u8005\u66f4\u5bb9\u6613\u5730\u6784\u5efa\u57fa\u4e8e\u5b66\u4e60\u7684\u7b97\u6cd5\u3002<\/p>\n\n\n\n

SOTA\u7b97\u6cd5<\/a><\/strong>\u7684\u96c6\u6210\uff1a<\/strong>\u672c\u6587\u96c6\u6210\u4e86\u6700\u5148\u8fdb\uff08SOTA\uff09\u7684SLAM\u7b97\u6cd5\uff0c\u8fd9\u9a8c\u8bc1\u4e86\u6846\u67b6\u7684\u7075\u6d3b\u6027\u548c\u53ef\u6269\u5c55\u6027\uff0c\u4e5f\u4fc3\u8fdb\u4e86\u5f00\u6e90\u793e\u533a<\/a><\/p>\n\n\n\n

\u5bf9\u4e0d\u540c\u7b97\u6cd5\u7684\u5b66\u4e60\u548c\u6bd4\u8f83\u3002<\/p>\n\n\n\n

\u5f00\u6e90\u4ee3\u7801<\/a><\/strong>\uff1a<\/strong>\u6240\u6709\u4ee3\u7801\u3001\u914d\u7f6e\u548c\u6570\u636e\u5747\u662f\u5f00\u6e90\u7684\uff0c\u65e8\u5728\u4fc3\u8fdb\u793e\u533a\u9a71\u52a8\u7684\u53d1\u5c55\u548c\u8fd9\u4e9b\u6280\u672f\u7684\u5e7f\u6cdb\u5e94\u7528\u3002<\/p>\n\n\n\n

3.\u7cfb\u7edf<\/p>\n\n\n\n

3.1 \u6846\u67b6\u6982\u8ff0<\/p>\n\n\n\n

XRDSLAM\u63d0\u4f9b\u4e86\u4e00\u79cd\u9ad8\u5ea6\u53ef\u914d\u7f6e\u3001\u9ad8\u6548\u4e14\u6613\u4e8e\u6269\u5c55\u7684\u6846\u67b6\u3002\u53d7\u5230NeRFStudio\u7684\u542f\u53d1\uff0cXRDSLAM\u91c7\u7528\u6a21\u5757\u5316\u4ee3\u7801\u8bbe\u8ba1\uff0c\u901a\u8fc7\u7ee7\u627f\u548c\u591a\u6001\u6765\u7ec4\u7ec7SLAM\u7ec4\u4ef6\u3002\u5b83\u901a\u8fc7\u96c6\u4e2d\u7ba1\u7406\u7684\u914d\u7f6e\u7c7b\u6765\u5904\u7406\u53c2\u6570\uff0c\u5e76\u4e14\u4f7f\u7528\u591a\u8fdb\u7a0b\u5e76\u884c\u5904\u7406\u6765\u52a0\u901f\u8ddf\u8e2a\u3001\u5efa\u56fe\u548c\u53ef\u89c6\u5316\u4efb\u52a1\uff0c\u4ece\u800c\u663e\u8457\u63d0\u5347\u7cfb\u7edf\u6027\u80fd\u3002XRDSLAM\u7ed3\u5408\u4e86\u7075\u6d3b\u7684\u914d\u7f6e\u3001\u8d44\u6e90\u5171\u4eab\u548c\u591a\u8fdb\u7a0b\u673a\u5236\uff0c\u4ee5\u6784\u5efa\u4e00\u4e2a\u591a\u529f\u80fd\u3001\u9ad8\u6548\u7684\u5f00\u53d1\u5e73\u53f0\u3002<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

\u25b2\u56fe1\uff5cXRDSLAM\u7684\u7cfb\u7edf\u6982\u8ff0<\/p>\n\n\n\n

XRDSLAM\u6846\u67b6\u5982\u56fe1\u6240\u793a\uff0c\u4e3b\u8981\u5904\u7406\u56db\u4e2a\u90e8\u5206\u7684\u4ea4\u4e92\uff1a<\/p>\n\n\n\n

1\uff09\u6570\u636e\u96c6\u8f93\u5165\uff1a<\/strong>\u8be5\u6846\u67b6\u9700\u8981\u4ece\u4e00\u4e2a\u6216\u8005\u591a\u4e2a\u4f20\u611f\u5668\u63a5\u6536\u6570\u636e\uff0c\u4f8b\u5982RGB\u56fe\u50cf\u3001\u6df1\u5ea6\u56fe\u50cf\u3001IMU\u6570\u636e\u7b49\u3002XRDSLAM\u901a\u8fc7\u591a\u4e2aDataset\u7c7b<\/a>\u6765\u5904\u7406\u4e0d\u540c\u683c\u5f0f\u7684\u6570\u636e\u96c6\uff0c\u5e76\u4e14\u63d0\u4f9b\u4e86\u7edf\u4e00\u7684\u6570\u636e\u52a0\u8f7d\u63a5\u53e3\u548c\u9884\u5904\u7406\u529f\u80fd\uff1b<\/p>\n\n\n\n

2\uff09\u6838\u5fc3SLAM\u7b97\u6cd5\uff1a<\/strong>XRDSLAM\u901a\u8fc7Algorithm\u548cModel\u7c7b\u6765\u7ec4\u7ec7\u6838\u5fc3SLAM\u7b97\u6cd5\u3002\u7528\u6237\u53ef\u4ee5\u901a\u8fc7\u7ee7\u627f\u548c\u66ff\u6362\u7ec4\u4ef6\u6765\u5feb\u901f\u5b9e\u73b0\u548c\u4f18\u5316SLAM\u7b97\u6cd5\u3002Model\u7c7b\u7531Algorithm\u7c7b\u8c03\u7528\u3002\u5f00\u53d1\u8005\u53ef\u4ee5\u901a\u8fc7Model\u7c7b\u6765\u8bbe\u8ba1\u548c\u7ba1\u7406\u6df1\u5ea6\u5b66\u4e60\u7b97\u6cd5\uff1b<\/p>\n\n\n\n

3\uff09\u53ef\u89c6\u5316\uff1a<\/strong>XRDSLAM\u63d0\u4f9b\u4e86\u5728\u7ebf\u548c\u79bb\u7ebf\u53ef\u89c6\u5316\u5de5\u5177\u3002\u5728\u7ebf\u53ef\u89c6\u5316\u5de5\u5177\u5b9e\u65f6\u663e\u793aSLAM\u7cfb\u7edf\u7684\u72b6\u6001\uff0c\u5305\u62ec2D\u56fe\u50cf\u6e32\u67d3\u548c3D\u573a\u666f\u8868\u793a\uff0c\u5e2e\u52a9\u7528\u6237\u8fdb\u884c\u8c03\u8bd5\u548c\u4f18\u5316\u3002\u79bb\u7ebf\u53ef\u89c6\u5316\u5de5\u5177\u7528\u4e8e\u6e32\u67d3\u548c\u663e\u793aSLAM\u7b97\u6cd5\u7684\u7ed3\u679c\uff0c\u652f\u6301\u4fdd\u5b58\u4e3a\u89c6\u9891\u6587\u4ef6\u4ee5\u4f9b\u540e\u7eed\u5206\u6790\uff1b<\/p>\n\n\n\n

4\uff09\u8bc4\u4f30\uff1a<\/strong>XRDSLAM\u81ea\u52a8\u4fdd\u5b58SLAM\u7cfb\u7edf\u7684\u7ed3\u679c\uff0c\u5e76\u4e14\u4f7f\u7528\u8bc4\u4f30\u6a21\u5757\u6765\u8bc4\u4f30\u8f68\u8ff9\u548c\u91cd\u5efa\u7ed3\u679c\u3002<\/p>\n\n\n\n

3.2 \u591a\u8fdb\u7a0b<\/h3>\n\n\n\n

SLAM\u7cfb\u7edf\u901a\u5e38\u7531\u591a\u4e2a\u6a21\u5757\u7ec4\u6210\uff0c\u4f8b\u5982\u8ddf\u8e2a\u3001\u5efa\u56fe\u548c\u95ed\u73af\u68c0\u6d4b<\/a>\u3002\u4e3a\u4e86\u5b9e\u73b0\u6a21\u5757\u89e3\u8026\u548c\u72ec\u7acb\u8fd0\u884c\uff0cXRDSLAM\u91c7\u7528\u4e86\u591a\u8fdb\u7a0b\u673a\u5236\u3002\u8be5\u6846\u67b6\u5305\u542b\u4e09\u4e2a\u5173\u952e\u7684\u5b50\u8fdb\u7a0b\uff1aTracker\u3001Mapper\u548cVisualizer\u3002<\/p>\n\n\n\n

1\uff09Tracker\uff1a<\/strong>\u5b83\u8d1f\u8d23\u5904\u7406\u8f93\u5165\u6570\u636e\u5e76\u4e14\u8f93\u51fa\u76f8\u673a\u4f4d\u59ff\u4fe1\u606f\u3002\u5b83\u4eceDataset\u7c7b\u4e2d\u63a5\u6536\u6570\u636e\uff0c\u8ba1\u7b97\u76f8\u673a\u7684\u4f4d\u59ff\u4fe1\u606f\u5e76\u4e14\u5c06\u7ed3\u679c\u4f20\u7ed9Mapper\u8fdb\u7a0b\uff1b<\/p>\n\n\n\n

2\uff09Mapper\uff1a<\/strong>\u5b83\u63d0\u53d6\u5173\u952e\u5e27\u5e76\u4e14\u4f18\u5316\u4f4d\u59ff\u548c\u5730\u56fe\u4fe1\u606f\u3002Mapper\u8fdb\u7a0b\u4eceTracker\u8fdb\u7a0b\u63a5\u6536\u5e27\u6570\u636e\u548c\u4f4d\u59ff\u53c2\u6570\uff0c\u4ee5\u4e0d\u65ad\u6269\u5c55\u5730\u56fe\uff0c\u8fdb\u4e00\u6b65\u4f18\u5316\u76f8\u673a\u4f4d\u59ff\u548c\u5730\u56fe\u53c2\u6570\uff0c\u5e76\u4e14\u5904\u7406\u5730\u56fe\u878d\u5408\u548c\u66f4\u65b0\u7b49\u4efb\u52a1\uff1b<\/p>\n\n\n\n

3\uff09Visualizer\uff1a<\/strong>\u5b83\u8d1f\u8d23\u53ef\u89c6\u5316\u8f68\u8ff9\u3001\u6e32\u67d3\u56fe\u50cf\u548c\u91cd\u5efa\u7ed3\u679c\u3002\u5b83\u5b9e\u65f6\u5730\u4eceTracker\u8fdb\u7a0b\u63a5\u6536\u6570\u636e\u5e76\u4e14\u53ef\u89c6\u5316\u5b83\u4eec\u3002<\/p>\n\n\n\n

\u901a\u8fc7Pytorch\u7684\u591a\u8fdb\u7a0b\u6a21\u5757\uff0cXRDSLAM\u5b9e\u73b0\u4e86\u8fd9\u4e9b\u8fdb\u7a0b\u7684\u521b\u5efa\u548c\u521d\u59cb\u5316\uff0c\u5b9e\u73b0\u4e86\u5e76\u884c\u5904\u7406\u3002\u8fd9\u79cd\u591a\u8fdb\u7a0b\u67b6\u6784<\/a>\u4e0d\u4ec5\u63d0\u9ad8\u4e86\u7cfb\u7edf\u7684\u6548\u7387\u548c\u5b9e\u65f6\u6027\uff0c\u8fd8\u5141\u8bb8\u6bcf\u4e2a\u6a21\u5757\u72ec\u7acb\u7ba1\u7406\u8d44\u6e90\uff0c\u51cf\u5c11\u4e86\u8d44\u6e90\u7ade\u4e89\u548c\u51b2\u7a81\uff0c\u4ece\u800c\u63d0\u9ad8\u4e86\u7cfb\u7edf\u7684\u7a33\u5b9a\u6027\u548c\u53ef\u9760\u6027\u3002\u6b64\u5916\uff0c\u5b83\u8fd8\u652f\u6301\u6a21\u5757\u7684\u72ec\u7acb\u6269\u5c55\u548c\u4f18\u5316\uff0c\u8fdb\u4e00\u6b65\u589e\u5f3a\u4e86\u6846\u67b6\u7684\u7075\u6d3b\u6027\u3002<\/p>\n\n\n\n

3.3 \u6570\u636e\u5171\u4eab\u673a\u5236<\/h3>\n\n\n\n

XRDSLAM\u6846\u67b6\u901a\u8fc7\u4ee5\u4e0b\u65b9\u6cd5\u5b9e\u73b0\u8fdb\u7a0b\u95f4\u6570\u636e\u5171\u4eab\u548c\u534f\u8c03\uff1a<\/p>\n\n\n\n

1\uff09BaseManager\uff1a<\/strong>\u5b83\u7528\u4e8e\u521b\u5efaAlgorithm\u7c7b\u7684\u5171\u4eab\u5b9e\u4f8b\uff0c\u5141\u8bb8Tracker\u548cMapper\u8fdb\u7a0b\u65b9\u4fbf\u5730\u5171\u4eab\u6570\u636e\u548c\u529f\u80fd\u3002\u901a\u8fc7\u96c6\u4e2d\u7ba1\u7406\u6838\u5fc3\u8ddf\u8e2a\u548c\u5efa\u56fe\u529f\u80fd\uff0c\u5b83\u4fc3\u8fdb\u4e86\u9ad8\u6548\u7684\u6570\u636e\u4ea4\u6362\u548c\u529f\u80fd\u534f\u4f5c\uff1b<\/p>\n\n\n\n

2\uff09Queue\uff1a<\/strong>\u5b83\u7528\u4e8e\u5728\u8fdb\u7a0b\u4e4b\u95f4\u4f20\u9012\u6570\u636e\u6d41\u3002Tracker\u8fdb\u7a0b\u5c06\u5904\u7406\u7684\u4f4d\u59ff\u4fe1\u606f\u548c\u8f93\u5165\u5e27\u4fe1\u606f\u6253\u5305\u6210Frame\uff0c\u5e76\u4e14\u901a\u8fc7map_buffer\u5c06\u5176\u4f20\u5165Mapper\u8fdb\u7a0b\u3002\u5b83\u8fd8\u901a\u8fc7viz_buffer\u5c06\u7cfb\u7edf\u72b6\u6001\u4fe1\u606f\u4f20\u5165Visualizer\u8fdb\u7a0b\uff0c\u4ece\u800c\u5b9e\u73b0\u9ad8\u6548\u7684\u6570\u636e\u6d41\u4f20\u8f93\u3002<\/p>\n\n\n\n

3\uff09Event\uff1a<\/strong>\u5b83\u7528\u4e8e\u8fdb\u7a0b\u4e4b\u95f4\u7684\u8fd0\u884c\u548c\u72b6\u6001\u540c\u6b65\u3002\u4e8b\u4ef6event_ready\u548cevent_processed\u7528\u4e8e\u8868\u793a\u6570\u636e\u5904\u7406\u7684\u51c6\u5907\u548c\u5b8c\u6210\uff0c\u786e\u4fddTracker\u548cMapper\u8fdb\u7a0b\u4e4b\u95f4\u7684\u6709\u6548\u534f\u8c03\u548c\u540c\u6b65\u3002<\/p>\n\n\n\n

\u901a\u8fc7\u8fd9\u4e9b\u673a\u5236\uff0cXRDSLAM\u80fd\u591f\u5728\u4e0d\u540c\u8fdb\u7a0b\u4e4b\u95f4\u4fdd\u6301\u6570\u636e\u548c\u7b97\u6cd5\u72b6\u6001\u7684\u4e00\u81f4\u6027\uff0c\u540c\u65f6\u4f7f\u7528\u7f13\u51b2\u533a\u548c\u4e8b\u4ef6\u540c\u6b65\u6765\u6709\u6548\u7ba1\u7406\u6570\u636e\u6d41\u548c\u8fdb\u7a0b\u4e4b\u95f4\u7684\u534f\u8c03\u3002<\/p>\n\n\n\n

4.SLAM\u7ec4\u4ef6<\/h2>\n\n\n\n

XRDSLAM\u91c7\u7528\u6a21\u5757\u5316\u8bbe\u8ba1\uff0c\u5141\u8bb8\u6bcf\u4e2aSLAM\u7ec4\u4ef6\u72ec\u7acb\u914d\u7f6e\u548c\u6269\u5c55\uff0c\u4ece\u800c\u63d0\u9ad8\u7cfb\u7edf\u7684\u6548\u7387\u548c\u53ef\u7ef4\u62a4\u6027\u3002\u6574\u4e2a\u7cfb\u7edf\u5305\u62ec\u6838\u5fc3\u7b97\u6cd5\u7ec4\u4ef6\u548c\u901a\u7528\u7ec4\u4ef6\u3002\u4e3a\u4e86\u51cf\u5c11\u5197\u4f59\u5f00\u53d1\uff0c\u901a\u7528\u6a21\u5757\u7684\u8bc4\u4f30\u548c\u53ef\u89c6\u5316\u88ab\u62bd\u8c61\u4e3autility\u7c7b\u3002<\/p>\n\n\n\n

4.1 \u57fa\u7c7b<\/h3>\n\n\n\n

\u57fa\u672c\u7ec4\u4ef6\u5305\u62ec\u7528\u4e8e\u914d\u7f6e\u53c2\u6570\u3001\u6570\u636e\u8bfb\u53d6\u3001\u4f4d\u59ff\u53d8\u6362\u548c\u76f8\u673a\u6a21\u578b\u7b49\u529f\u80fd\u7684\u5e38\u7528\u57fa\u7c7b\u3002<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

\u25b2\u56fe2\uff5cXRDSLAM\u7684\u7ec4\u6210\u90e8\u5206<\/p>\n\n\n\n

1\uff09Config\uff1a<\/strong>\u4e0eNeRFStudio\u4e00\u6837\uff0cXRDSLAM\u4f7f\u7528\u914d\u7f6e\u7c7b\u6765\u7ec4\u7ec7\u548c\u7ba1\u7406SLAM\u7b97\u6cd5\u7684\u53c2\u6570\u548c\u7ec4\u4ef6\u3002\u8be5\u914d\u7f6e\u7c7b\u7ee7\u627f\u81eaInstantiateConfig\uff0c\u5e76\u4e14\u901a\u8fc7setup\u65b9\u6cd5\u6765\u52a8\u6001\u5b9e\u4f8b\u5316\u76f8\u5e94\u7684\u5bf9\u8c61\u3002\u7cfb\u7edf\u4e2d\u7684\u6bcf\u4e2a\u7ec4\u4ef6\u5747\u53ef\u4ee5\u901a\u8fc7\u7edf\u4e00\u7684\u65b9\u5f0f\u6765\u914d\u7f6e\u548c\u5b9e\u4f8b\u5316\u3002XRDSLAMerConfig\u96c6\u4e2d\u7ba1\u7406\u4e0eSLAM\u8fd0\u884c\u76f8\u5173\u7684\u6240\u6709\u53c2\u6570\uff0c\u5e76\u4e14\u901a\u8fc7\u8be5\u914d\u7f6e\u7c7b\uff0c\u7528\u6237\u53ef\u4ee5\u65b9\u4fbf\u5730\u63a7\u5236SLAM\u7cfb\u7edf\u7684\u884c\u4e3a\uff0c\u7b80\u5316\u53c2\u6570\u7ba1\u7406\u548c\u8c03\u8bd5\u8fc7\u7a0b\u3002\u914d\u7f6e\u7c7b\u7684\u4f7f\u7528\u4e0d\u4ec5\u63d0\u9ad8\u4e86\u4ee3\u7801\u7684\u53ef\u8bfb\u6027\uff0c\u8fd8\u589e\u5f3a\u4e86\u7cfb\u7edf\u7684\u7075\u6d3b\u6027\u548c\u53ef\u6269\u5c55\u6027\u3002<\/p>\n\n\n\n

2\uff09BaseDataset\uff1a<\/strong>\u57fa\u4e8eNICE-SLAM\u4ee3\u7801\uff0cXRDSLAM\u901a\u8fc7get_dataset\u65b9\u6cd5\u6765\u652f\u6301\u591a\u79cd\u7c7b\u578b\u7684\u6570\u636e\u96c6\uff0c\u5e76\u4e14\u5141\u8bb8\u901a\u8fc7\u914d\u7f6e\u8f7b\u677e\u6269\u5c55\u6216\u8005\u5207\u6362\u6570\u636e\u6e90\uff0c\u4ee5\u6ee1\u8db3\u4e0d\u540c\u7684\u6570\u636e\u9700\u6c42\u3002\u6570\u636e\u96c6\u52a0\u8f7d\u7531\u7ee7\u627f\u81eaBaseDataset\u7684\u591a\u79cd\u6570\u636e\u96c6\u7c7b\u548c\u8f85\u52a9\u51fd\u6570\u6765\u5b9e\u73b0\u3002\u8fd9\u4e9butility\u7c7b\u4e0e\u5404\u79cd\u6570\u636e\u683c\u5f0f\u548c\u6765\u6e90\u517c\u5bb9\uff0c\u63d0\u4f9b\u4e86\u4e00\u4e2a\u7edf\u4e00\u7684\u76f8\u673a\u6a21\u578b\u548c\u6570\u636e\u52a0\u8f7d\u63a5\u53e3\uff0c\u8fd9\u6709\u52a9\u4e8e\u96c6\u6210\u5230SLAM\u6846\u67b6\u4e2d\u3002\u76ee\u524d\u652f\u6301\u7684\u6570\u636e\u96c6\u5305\u62ecReplica\u3001ScanNet\u3001TUM RGBD\u30017-Scenes\u3001Co-Fusion\u548cEuroc\uff0c\u5176\u4e3b\u8981\u529f\u80fd\u5305\u62ec\u8bfb\u53d6RGB\u548c\u6df1\u5ea6\u56fe\u50cf\u3001IMU\u6570\u636e\u548c\u76f8\u673a\u4f4d\u59ff\uff0c\u4ee5\u53ca\u6267\u884c\u5fc5\u8981\u7684\u9884\u5904\u7406\uff08\u4f8b\u5982\u53bb\u7578\u53d8\u3001\u964d\u91c7\u6837\uff09\u3002<\/p>\n\n\n\n

3\uff09OptimizablePose\uff1a<\/strong>OptimizablePose\u7c7b\u7528\u4e8eSLAM\u4e2d\u7684\u4f4d\u59ff\u4f18\u5316\uff0c\u5904\u7406\u65cb\u8f6c\u548c\u5e73\u79fb\u3002\u672c\u6587\u6269\u5c55\u4e86Vox-Fusion\u4ee3\u7801\u4e2d\u7684OptimizablePose\u7c7b\uff0c\u4ee5\u652f\u6301\u4e24\u79cd\u65cb\u8f6c\u8868\u793a\u65b9\u6cd5\uff1a\u89d2\u8f74\u548c\u56db\u5143\u6570<\/a>\u3002\u5b83\u5229\u7528PyTorch\u7684\u81ea\u52a8\u5fae\u5206\u529f\u80fd\u6765\u5b9e\u73b0\u68af\u5ea6\u4f18\u5316\uff0c\u9ad8\u6548\u5730\u5728\u4f4d\u59ff\u53c2\u6570\u548c\u65cb\u8f6c\u77e9\u9635\u4e4b\u95f4\u8fdb\u884c\u8f6c\u6362\u3002OptimizablePose\u7c7b\u5305\u542b\u5728Frame\u7c7b\u4e2d\u3002<\/p>\n\n\n\n

4\uff09Frame\uff1a<\/strong>Frame\u7c7b\u7528\u4e8e\u7ba1\u7406SLAM\u7cfb\u7edf\u4e2d\u7684\u5355\u5e27\u6570\u636e\uff0c\u5305\u62ecRGB\u56fe\u50cf\u3001\u6df1\u5ea6\u56fe\u50cf\u548c\u4f4d\u59ff\u3002\u5b83\u5c01\u88c5\u4e86\u8fd9\u4e9b\u4fe1\u606f\uff0c\u5e76\u4e14\u63d0\u4f9b\u4e86\u8bbe\u7f6e\u3001\u83b7\u53d6\u548c\u4f18\u5316\u7684\u65b9\u6cd5\u3002\u901a\u8fc7\u7ee7\u627ftorch.nn.Module\uff0c\u5b83\u6709\u52a9\u4e8e\u5728\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\u5185\u8fdb\u884c\u7ba1\u7406\u548c\u4f18\u5316\u3002<\/p>\n\n\n\n

5\uff09Camera\uff1a<\/strong>Camera\u7528\u4e8e\u7ba1\u7406XRDSLAM\u4e2d\u7684\u76f8\u673a\u6a21\u578b\u3002\u76ee\u524d\uff0c\u652f\u6301\u7684\u6a21\u578b\u662f\u9488\u5b54\u6a21\u578b\u3002\u76f8\u673a\u7684\u5185\u53c2\u548c\u5206\u8fa8\u7387\u4fe1\u606f\u5b58\u50a8\u5728Camera\u7c7b\u4e2d\u3002<\/p>\n\n\n\n

4.2 \u6838\u5fc3\u6a21\u5757<\/h3>\n\n\n\n

1\uff09Algorithm\uff1a<\/strong>Algorithm\u7c7b\u662f\u4e00\u4e2a\u62bd\u8c61\u7c7b\uff0c\u5b83\u5b9a\u4e49\u4e86\u5b9e\u73b0SLAM\u7b97\u6cd5\u6240\u9700\u7684\u6838\u5fc3\u63a5\u53e3\u548c\u65b9\u6cd5\u3002\u5b83\u63d0\u4f9b\u4e86\u9009\u62e9\u4f18\u5316\u7684\u5e27\u3001\u83b7\u53d6\u6a21\u578b\u8f93\u5165\u3001\u8ba1\u7b97\u635f\u5931\u3001\u9884\u5904\u7406\u548c\u540e\u5904\u7406\u7b49\u529f\u80fd\uff0c\u4e3aSLAM\u7cfb\u7edf\u6784\u5efa\u4e86\u901a\u7528\u7684\u8fed\u4ee3\u4f18\u5316\u8fc7\u7a0b\u3002\u6b64\u5916\uff0c\u8be5\u7c7b\u5305\u62ec\u7528\u4e8e\u4f18\u5316\u5668\u914d\u7f6e\u3001\u7cfb\u7edf\u72b6\u6001\u7ef4\u62a4\u548c\u66f4\u65b0\u7684\u62bd\u8c61\u65b9\u6cd5\u3002<\/p>\n\n\n\n

\u901a\u8fc7\u7ee7\u627fAlgorithm\u7c7b\uff0c\u53ef\u4ee5\u5b9e\u73b0\u4e0d\u540c\u7684SLAM\u7b97\u6cd5\uff0c\u4ee5\u5b9e\u73b0\u6ee1\u8db3\u7279\u5b9a\u8981\u6c42\u7684\u81ea\u5b9a\u4e49\u529f\u80fd\u3002XRDSLAM\u6846\u67b6\u5df2\u7ecf\u96c6\u6210\u4e86\u5404\u79cd\u7b97\u6cd5\uff0c\u4f8b\u5982NICE-SLAM\u3001Co-SLAM\u3001Vox-Fusion\u3001Point-SLAM\u548cSplaTAM\u3002\u6bcf\u79cd\u7b97\u6cd5\u5728\u7ee7\u627fAlgorithm\u7c7b\u7684\u57fa\u7840\u4e0a\u6269\u5c55\u4e86\u7279\u5b9a\u7684\u4f18\u5316\u548c\u5904\u7406\u8fc7\u7a0b\uff0c\u5c55\u73b0\u4e86\u6846\u67b6\u7684\u7075\u6d3b\u6027\u548c\u53ef\u6269\u5c55\u6027\u3002<\/p>\n\n\n\n

\u6b64\u5916\uff0cAlgorithm\u7c7b\u5c06\u8ddf\u8e2a\u548c\u5efa\u56fe\u8fdb\u7a0b\u89e3\u8026\uff0c\u5141\u8bb8\u7528\u6237\u66ff\u6362\u548c\u7ed3\u5408\u4e0d\u540c\u7684\u8ddf\u8e2a\u548c\u5efa\u56fe\u65b9\u6cd5\uff0c\u4ee5\u5feb\u901f\u9a8c\u8bc1\u5404\u79cd\u89e3\u51b3\u65b9\u6848\u3002<\/p>\n\n\n\n

\u5177\u4f53\u800c\u8a00\uff0c\u5f53\u8ddf\u8e2a\u529f\u80fd\u76f4\u63a5\u4f7f\u7528\u8f93\u5165\u771f\u503c\u4f4d\u59ff\u65f6\uff0c\u7cfb\u7edf\u53ef\u4ee5\u4ece\u4f20\u7edf\u7684SLAM\u6a21\u5f0f\u8fc7\u6e21\u5230\u589e\u91cf\u91cd\u5efa\u8fc7\u7a0b\u3002NeuralRecon\u901a\u8fc7\u8be5\u65b9\u6cd5\u96c6\u6210\u5230XRDSLAM\u4e2d\u3002\u5bf9\u4e8e\u4e0d\u9700\u8981\u5efa\u56fe\u7684\u7b97\u6cd5\uff08\u4f8b\u5982DPVO\uff09\uff0c\u5f00\u53d1\u8005\u53ef\u4ee5\u8df3\u8fc7Mapper\u6a21\u5757\uff0c\u4ec5\u5728Tracker\u6a21\u5757\u4e2d\u5b9e\u73b0\u4f4d\u59ff\u8ddf\u8e2a\u529f\u80fd\u3002<\/p>\n\n\n\n

2\uff09Model\uff1a<\/strong>Model\u7c7b<\/a>\u662f\u6240\u6709\u7279\u5b9a\u7f51\u7edc\u6a21\u578b\u7684\u57fa\u7c7b\uff0c\u4e3a\u6a21\u578b\u63d0\u4f9b\u4e86\u57fa\u672c\u7ed3\u6784\u548c\u901a\u7528\u63a5\u53e3\u3002\u5b83\u7ee7\u627f\u81eaPyTorch\u7684nn.Module\uff0c\u7528\u4e8e\u5b9a\u4e49\u548c\u7ba1\u7406\u6df1\u5ea6\u5b66\u4e60\u6a21\u578b\u7684\u4e3b\u8981\u7ec4\u4ef6\u3002Model\u7c7b\u7684\u6838\u5fc3\u4efb\u52a1\u5305\u62ec\u524d\u5411\u4f20\u64ad\u3001\u635f\u5931\u8ba1\u7b97\u548c\u83b7\u53d6\u4f18\u5316\u7684\u53c2\u6570\u3002<\/p>\n\n\n\n

\u5b50\u7c7b\u6839\u636e\u6a21\u578b\u7684\u7279\u5b9a\u7f51\u7edc\u7ed3\u6784\uff08\u5305\u62ec\u7f16\u7801\u5668\u3001\u89e3\u7801\u5668\u7b49\uff09\u8fdb\u884c\u914d\u7f6e\uff0c\u7ba1\u7406\u6a21\u578b\u8f93\u5165\u548c\u8f93\u51fa\u3001\u635f\u5931\u8ba1\u7b97\u3001\u5730\u56fe\u66f4\u65b0\u4ee5\u53caNeRF\u548c3DGS\u6e32\u67d3\u7b49\u8fc7\u7a0b\u3002\u8fd9\u4e9b\u5b50\u7c7b\u5b9e\u73b0\u4e86\u4e0d\u540c\u5730\u56fe\u8868\u793a\u7684\u7ba1\u7406\u548c\u66f4\u65b0\uff0c\u4f8b\u5982\u591a\u5206\u8fa8\u7387\u7279\u5f81\u7f51\u683c\u3001\u7a00\u758f\u4f53\u7d20\u516b\u53c9\u6811\u3001\u54c8\u5e0c\u7f51\u683c\u3001\u795e\u7ecf\u70b9\u4e91\u30013DGS\u7b49\u3002<\/p>\n\n\n\n

3\uff09Visualizer\u548cOfflineViewer\uff1a<\/strong>Visualizer\u7c7b\u8d1f\u8d23SLAM\u8fc7\u7a0b\u4e2d\u751f\u6210\u7684\u56fe\u50cf\u548c3D\u6570\u636e\u7684\u5b9e\u65f6\u53ef\u89c6\u5316\u3002\u5b83\u4eceSLAM\u7cfb\u7edf\u63a5\u6536\u5e76\u4e14\u5904\u7406\u6570\u636e\uff08\u4f8b\u5982\u4f4d\u59ff\u3001RGB\u56fe\u50cf\u3001\u6df1\u5ea6\u56fe\u50cf\u3001\u7f51\u683c\u548c\u70b9\u4e91\uff09\uff0c\u5e76\u4e14\u4f7f\u7528Open3D\u548cMatplotlib\u6765\u5b9e\u65f6\u663e\u793a\u3002\u6838\u5fc3\u529f\u80fd\u5305\u62ec\u8f68\u8ff9\u663e\u793a\u3001\u56fe\u50cf\u8d28\u91cf\u6307\u6807\u7684\u8ba1\u7b97\u548c\u5448\u73b0\u4ee5\u53ca\u53ef\u89c6\u5316\u7ed3\u679c\u7684\u4fdd\u5b58\u3002\u5373\u65f6\u53ef\u89c6\u5316\u6548\u679c\u4f7f\u7814\u7a76\u8005\u80fd\u591f\u5feb\u901f\u8c03\u8bd5\u548c\u4f18\u5316\u7b97\u6cd5\u3002<\/p>\n\n\n\n

OfflineViewer\u7528\u4e8e\u56de\u653eSLAM\u7b97\u6cd5\u7684\u7ed3\u679c\u3002\u5b83\u8bfb\u53d6\u8f68\u8ff9\u6570\u636e\u5e76\u4e14\u76f8\u5e94\u5730\u66f4\u65b0\u754c\u9762\u4e2d\u7684\u56fe\u50cf\u3001\u70b9\u4e91\u548c\u7f51\u683c\u3002\u6839\u636e\u914d\u7f6e\uff0cOfflineViewer\u8fd8\u53ef\u4ee5\u5c06\u6e32\u67d3\u7ed3\u679c\u4fdd\u5b58\u4e3a\u89c6\u9891\u3002<\/p>\n\n\n\n

4\uff09Metrics\u548cEvaluation\uff1a<\/strong>XRDSLAM\u652f\u6301\u83b7\u53d6\u548c\u4fdd\u5b58\u8ddf\u8e2a\u5e27\u7684\u4f4d\u59ff\u3001\u6e32\u67d3\u7684\u56fe\u50cf\u4ee5\u53ca\u573a\u666f\u7684\u70b9\u4e91\u6216\u8005\u7f51\u683c\uff0c\u5e76\u4e14\u6839\u636e\u771f\u503c\u6570\u636e\u6765\u8ba1\u7b97\u76f8\u5173\u7684\u8bc4\u4f30\u6307\u6807\u3002EvalMetrics\u7c7b\u4e3aSLAM\u7ed3\u679c\u63d0\u4f9b\u4e86\u4e00\u4e2a\u8bc4\u4f30\u8fdb\u7a0b\u3002SLAM\u7cfb\u7edf\u7684\u603b\u4f53\u8bc4\u4f30\u5305\u62ec\u4e09\u4e2a\u4e3b\u8981\u65b9\u9762\uff1a\u6e32\u67d3\u8d28\u91cf\u30013D\u91cd\u5efa\u8d28\u91cf\u548c\u8f68\u8ff9\u7cbe\u5ea6\uff0c\u5177\u4f53\u6307\u6807\u5305\u62ec\uff1a<\/p>\n\n\n\n

\u8f68\u8ff9\u7cbe\u5ea6\uff1a<\/strong>\u5728SLAM\u9886\u57df\u4e2d\uff0cATE\u662f\u4e00\u79cd\u5e7f\u6cdb\u4f7f\u7528\u7684\u7cbe\u5ea6\u6307\u6807\uff0c\u672c\u6587\u7528\u8be5\u6307\u6807\u6765\u8bc4\u4f30\u8f68\u8ff9\u7cbe\u5ea6\uff1b<\/p>\n\n\n\n

\u6e32\u67d3\u6307\u6807\uff1a<\/strong>\u672c\u6587\u91c7\u7528PSNR\uff08\u5cf0\u503c\u4fe1\u566a\u6bd4<\/a>\uff09\u3001SSIM\uff08\u7ed3\u6784\u76f8\u4f3c\u6027\u6307\u6570\u5ea6\u91cf\uff09\u548cLPIPS\uff08\u5b66\u4e60\u611f\u77e5\u56fe\u50cf\u5757\u76f8\u4f3c\u5ea6\uff09\u6765\u8bc4\u4f30\u65b0\u89c6\u56fe\u6e32\u67d3\u7684\u8d28\u91cf\u3002<\/p>\n\n\n\n

\u91cd\u5efa\u8d28\u91cf\uff1a<\/strong>\u672c\u6587\u4f7f\u75287\u4e2a\u6307\u6807\u6765\u5168\u9762\u8bc4\u4f30\u91cd\u5efa\u7684\u8d28\u91cf\uff0c\u5305\u62ec\u7cbe\u5ea6\u3001\u5b8c\u6574\u5ea6\u3001\u5b8c\u6210\u7387\u3001\u51c6\u786e\u7387\u3001\u53ec\u56de\u7387\u3001F-Score\u3001L1-Depth\u3002<\/p>\n\n\n\n

5.\u5b9e\u9a8c<\/h2>\n\n\n\n

5.1 SLAM\u5b9e\u73b0<\/h3>\n\n\n\n

\u8868\u683c1\u5c55\u73b0\u4e86\u57283D visualizer\u4e2d\u8fd0\u884c\u7684\u5404\u79cd\u7b97\u6cd5\u3002\u5176\u4e2d\u5305\u542b5\u79cdSLAM\u65b9\u6cd5\u3001\u4e00\u79cd\u91cc\u7a0b\u8ba1\u65b9\u6cd5\u548c\u4e00\u79cd\u589e\u91cf\u5f0f\u91cd\u5efa\u65b9\u6cd5\uff0c\u6240\u6709\u5747\u96c6\u6210\u5728XRDSLAM\u4e2d\u3002<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

\u25b2\u8868\u683c1\uff5c\u96c6\u6210\u4e0d\u540c\u7c7b\u578b\u7684SLAM\uff0c\u5305\u62ec\u57fa\u4e8eNeRF\/3DGS\u7684SLAM\u3001\u91cc\u7a0b\u8ba1\u548c\u91cd\u5efa\uff0c\u6db5\u76d6\u4e86\u4e0d\u540c\u7684\u573a\u666f\u7f16\u7801\u548c\u51e0\u4f55\u8868\u793a<\/p>\n\n\n\n

\u6b64\u5916\uff0cXRDSLAM\u53ef\u4ee5\u7075\u6d3b\u652f\u6301\u5404\u79cd\u573a\u666f\u7f16\u7801\u683c\u5f0f\u548c\u5efa\u56fe\u65b9\u6cd5\uff0c\u4f8b\u5982NICE-SLAM\u7684\u591a\u5206\u8fa8\u7387\u7279\u5f81\u7f51\u683c\u3001Co-SLAM\u7684\u54c8\u5e0c\u7f51\u683c\u3001Vox-Fusion\u7684\u7a00\u758f\u4f53\u7d20\u516b\u53c9\u6811\u3001Point-SLAM\u7684\u795e\u7ecf\u70b9\u4e91\u548cSplaTAM\u76843DGS\u3002<\/p>\n\n\n\n

5.2 \u6d4b\u8bd5\u6570\u636e\u96c6<\/h3>\n\n\n\n

\u4e3a\u4e86\u4fbf\u4e8e\u5bf9\u73b0\u6709SLAM\u7b97\u6cd5\u8fdb\u884c\u8bc4\u4f30\uff0c\u672c\u6587\u9009\u62e9Replica\u6570\u636e\u96c6\u4f5c\u4e3a\u57fa\u51c6\u3002Replica\u6570\u636e\u96c6\u63d0\u4f9b\u4e86\u771f\u5b9e\u3001\u9ad8\u8d28\u91cf\u7684\u7a7a\u95f4\u91cd\u5efa\u7ed3\u679c\u3001\u771f\u503c\u8f68\u8ff9\u548cRGBD\u6570\u636e\uff0c\u6ee1\u8db3\u4e86\u6d4b\u8bd5\u548c\u8bc4\u4f30\u5927\u591a\u6570\u5f53\u524dSLAM\u7b97\u6cd5\u7684\u9700\u6c42\u3002\u672c\u6587\u4f7f\u75288\u4e2a\u5ba4\u5185RGBD\u5e8f\u5217\u8fdb\u884c\u8bc4\u4f30\u3002\u6b64\u5916\uff0c\u5f00\u53d1\u8005\u53ef\u4ee5\u53c2\u8003\u539f\u59cb\u8bba\u6587\u4e2d\u63d0\u4f9b\u7684\u914d\u7f6e\u53c2\u6570\uff0c\u7528\u4e8e\u5728\u5176\u5b83\u6570\u636e\u96c6\u4e0a\u6bd4\u8f83\u8bc4\u4f30\u3002<\/p>\n\n\n\n

5.3 \u8bc4\u4f30<\/h3>\n\n\n\n

1\uff09\u7cbe\u5ea6\u8bc4\u4f30\uff1a<\/strong>XRDSLAM\u63d0\u4f9b\u4e86\u4e00\u4e2a\u7528\u4e8e\u8be6\u7ec6\u8bc4\u4f30\u7684\u8bc4\u4f30\u6a21\u5757\u3002\u672c\u6587\u4f7f\u7528Replica\u6570\u636e\u96c6\u7684\u539f\u59cb\u5206\u8fa8\u7387\u548c\u539f\u59cb\u7b97\u6cd5\u4ed3\u5e93\u63d0\u4f9b\u7684\u9ed8\u8ba4\u53c2\u6570\u6765\u6bd4\u8f83\u7b97\u6cd5\u7684\u51c6\u786e\u6027\u3002\u6bcf\u4e2aReplica\u6570\u636e\u96c6\u5305\u542b2000\u5e27\u56fe\u50cf\uff0c\u6bcf50\u5e27\u8fdb\u884c\u4e00\u6b21\u573a\u666f\u6e32\u67d3\uff0c\u4ee5\u83b7\u5f97RGB\u548c\u6df1\u5ea6\u56fe\u50cf\u3002Tracker\u8fdb\u7a0b\u5b9e\u65f6\u66f4\u65b0\u8f68\u8ff9\u4f30\u8ba1\uff0c\u5e76\u4e14\u5728\u8fd0\u884c\u5b8c\u6210\u540e\u81ea\u52a8\u4fdd\u5b583D\u7f51\u683c\u3002\u5728\u8bc4\u4f30\u8fc7\u7a0b\u4e2d\uff0cDPVO\u9700\u8981\u5c3a\u5ea6\u6821\u6b63\uff0c\u8fd9\u662f\u56e0\u4e3aDPVO\u4ec5\u4f7f\u7528RGB\u6570\u636e\u800c\u6ca1\u6709\u771f\u5b9e\u5c3a\u5ea6\u3002<\/p>\n\n\n\n

\u8868\u683c2\u663e\u793a\u4e86\u591a\u79cd\u539f\u59cb\u7b97\u6cd5\u53ca\u5176\u96c6\u6210\u7248\u672c\uff08\u5e26\u201c*\u201d\uff09\u57288\u4e2aReplica\u6570\u636e\u96c6\u4e0a\u7684\u6307\u6807\u5e73\u5747\u7ed3\u679c\u3002\u539f\u59cb\u7b97\u6cd5\u548cXRDSLAM\u96c6\u6210\u7684\u7b97\u6cd5\u4e4b\u95f4\u66f4\u597d\u7684\u7ed3\u679c\u4ee5\u7c97\u4f53\u8868\u793a\u3002\u5b8c\u6574\u7684\u8bc4\u4f30\u8868\u683c\u53ef\u4ee5\u5728github\u4ed3\u5e93\u4e2d\u67e5\u770b\u3002<\/p>\n\n\n\n

\"\"
\u25b2\u8868\u683c2\uff5c\u4ece\u4e09\u4e2a\u65b9\u9762\u8bc4\u4f30\u57fa\u4e8e\u6307\u6807\u7684\u96c6\u6210\u65b9\u6cd5\uff1a\u8f68\u8ff9\u3001\u6e32\u67d3\u548c\u91cd\u5efa<\/figcaption><\/figure>\n\n\n\n

\u4ece\u8868\u683c2\u4e2d\u53ef\u4ee5\u660e\u663e\u770b\u51fa\uff0c\u96c6\u6210\u5230XRDSLAM\u4e2d\u7684\u7b97\u6cd5\u5728\u7cbe\u5ea6\u65b9\u9762\u4e0e\u539f\u59cb\u7248\u672c\u76f8\u5f53\u3002\u57283D\u91cd\u5efa\u8d28\u91cf\u65b9\u9762\uff0cPoint-SLAM\u5728\u7cbe\u5ea6\u6307\u6807\u4e0a\u8868\u73b0\u6700\u4f73\u3002\u5728\u56fe\u50cf\u6e32\u67d3\u8d28\u91cf\u65b9\u9762\uff0cPoint-SLAM\u548cSplaTAM\u5747\u5c55\u73b0\u51fa\u51fa\u8272\u7684\u6027\u80fd\u3002\u5728\u5b9a\u4f4d\u6307\u6807\u65b9\u9762\uff0cDPVO\u7684\u7cbe\u5ea6\u6700\u9ad8\uff0cSplaTAM\u7684\u5b9a\u4f4d\u7cbe\u5ea6\u4f18\u4e8e\u5176\u5b83\u57fa\u4e8e\u795e\u7ecf\u5efa\u56fe\u7684\u65b9\u6cd5\u3002<\/p>\n\n\n\n

\u56fe3\u663e\u793a\u4e86\u96c6\u6210\u5230XRDSLAM\u4e2d\u7684\u82e5\u5e72\u79cdSLAM\u65b9\u6cd5\u5728Replica\/room0\u6570\u636e\u96c6\u4e0a\u8fdb\u884c\u7f51\u683c\u91cd\u5efa\u7684\u5b9a\u6027\u7ed3\u679c\uff0c\u5176\u6027\u80fd\u9010\u6b65\u63d0\u5347\uff0c\u4eceNICE-SLAM\u3001CO-SLAM\u3001Vox-Fusion\u5230Point-SLAM\u3002<\/p>\n\n\n\n

\"\"<\/figure>\n\n\n\n

\u25b2\u56fe3\uff5c\u4e0d\u540c\u96c6\u6210\u7b97\u6cd5\u5728Replica\u6570\u636e\u96c6\u4e0a\u7f51\u683c\u91cd\u5efa\u7684\u6bd4\u8f83<\/p>\n\n\n\n

2\uff09\u6027\u80fd\u8bc4\u4f30\uff1a<\/strong>\u4f7f\u7528Python\u7684tqdm\u5e93\u6765\u7edf\u8ba1\u6bcf\u5e27\u7684\u8fd0\u884c\u65f6\u95f4\uff0c\u5e76\u4e14\u8ba1\u7b97\u6bcf\u79cd\u7b97\u6cd5\u7684FPS\u3002\u8868\u683c1\u603b\u7ed3\u4e86\u5728\u964d\u91c7\u6837\u7684Replica\/office0\u6570\u636e\u96c6\u4e0a\u8fd0\u884cXRDSLAM\u7684\u57fa\u51c6\u6d4b\u8bd5\u7ed3\u679c\u3002\u7531\u4e8eNeuralRecon\u4ec5\u652f\u6301640×480\u5206\u8fa8\u7387\uff0c\u56e0\u6b64\u56fe\u50cf\u4f1a\u4ece\u539f\u59cb\u5206\u8fa8\u7387\u88c1\u526a\u5e76\u4e14\u8c03\u6574\u5230\u8fd9\u4e00\u5206\u8fa8\u7387\u8fdb\u884c\u6d4b\u8bd5\u3002<\/p>\n\n\n\n

\u8868\u683c1\u8868\u660e\uff0c\u5728\u96c6\u6210\u5230XRDSLAM\u7684\u7b97\u6cd5\u4e2d\uff0cNeuralRecon\u7684\u6548\u7387\u6700\u9ad8\u3002\u5728SLAM\u7b97\u6cd5\u4e2d\uff0c\u4f7f\u7528\u54c8\u5e0c\u7f51\u683c\u7684Co-SLAM\u5728\u6548\u7387\u65b9\u9762\u8868\u73b0\u6700\u4f73\uff0c\u800cPoint-SLAM\u548cSplatTAM\u5219\u6548\u7387\u76f8\u5bf9\u8f83\u4f4e\u3002\u5c3d\u7ba1SplatTAM\u5728\u56fe\u50cf\u6e32\u67d3\u65b9\u9762\u66f4\u9ad8\u6548\uff0c\u4f46\u662f\u7531\u4e8e\u9700\u8981\u5904\u7406\u6bcf\u4e00\u5e27\u548c\u66f4\u9ad8\u7684\u8fed\u4ee3\u6b21\u6570\uff0c\u7cfb\u7edf\u7684\u6574\u4f53\u6548\u7387\u5e76\u4e0d\u7406\u60f3\u3002<\/p>\n\n\n\n

\u6b64\u5916\uff0cXRDSLAM\u4e2dNICE-SLAM\u7684FPS\u4e0e\u8bba\u6587\u4e2d\u7684\u7ed3\u679c\u975e\u5e38\u4e0d\u540c\uff0c\u4e3b\u8981\u662f\u56e0\u4e3a\u539f\u59cb\u7684NICE-SLAM\u9ed8\u8ba4\u5728\u8fd0\u884c\u8fc7\u7a0b\u4e2d\u63d0\u53d6\u7f51\u683c\uff0c\u8fd9\u4f1a\u5bfc\u81f4\u8017\u65f6\u66f4\u957f\u3002XRDSLAM\u6807\u51c6\u5316\u4e86\u6570\u636e\u4fdd\u5b58\u8fc7\u7a0b\uff0c\u4f7f\u5f97\u7b97\u6cd5\u6027\u80fd\u7684\u6bd4\u8f83\u66f4\u4e3a\u516c\u6b63\u3002\u5f71\u54cd\u7b97\u6cd5\u6548\u7387\u7684\u56e0\u7d20\u4e0d\u4ec5\u5305\u62ec\u6570\u636e\u4fdd\u5b58\u548c\u7b97\u6cd5\u81ea\u8eab\u7684\u8fdb\u7a0b\uff0c\u8fd8\u5305\u62ec\u8fed\u4ee3\u6b21\u6570\u3001\u91c7\u6837\u70b9\u6570\u91cf\u548c\u6570\u636e\u5904\u7406\u95f4\u9694\u5e27\u3002\u4e3a\u4e86\u786e\u4fdd\u4e0e\u539f\u59cb\u7b97\u6cd5\u7684\u7cbe\u5ea6\u4fdd\u6301\u4e00\u81f4\uff0c\u5728\u8bc4\u4f30\u9636\u6bb5\u4f7f\u7528\u4e86\u539f\u59cb\u7b97\u6cd5\u7684\u9ed8\u8ba4\u914d\u7f6e\u3002<\/p>\n\n\n\n

\u8868\u683c1\u8fd8\u63d0\u4f9b\u4e86XRDSLAM\u5728Replica\/office0\u6570\u636e\u96c6\u4e0a\u8fd0\u884c\u65f6\u7684\u5cf0\u503cGPU\u5185\u5b58\u4f7f\u7528\u60c5\u51b5\uff08\u4ee5GB\u4e3a\u5355\u4f4d\uff09\u3002\u5728\u6d4b\u8bd5\u8fc7\u7a0b\u4e2d\uff0cVox-Fusion\u7684GPU\u5185\u5b58\u6700\u521d\u98d9\u5347\u5230\u5927\u7ea610GB\u5de6\u53f3\uff0c\u7136\u540e\u4e0b\u964d\u5e76\u4e14\u7a33\u5b9a\u57284-6GB\u4e4b\u95f4\u3002SplaTAM\u7684\u5185\u5b58\u4f7f\u7528\u4e0e\u5206\u8fa8\u7387\u5bc6\u5207\u76f8\u5173\uff0c\u4e3b\u8981\u662f\u56e0\u4e3a\u5b83\u4f7f\u7528RGB-D\u70b9\u4e91\u8fdb\u884cGS\u521d\u59cb\u5316\u3002\u57fa\u4e8e\u795e\u7ecf\u70b9\u4e91\u7684Point-SLAM\u968f\u7740\u70b9\u4e91\u5730\u56fe\u7684\u6269\u5c55\uff0c\u5185\u5b58\u6d88\u8017\u4e5f\u5728\u589e\u52a0\u3002\u57fa\u4e8e\u56fa\u5b9a\u5206\u8fa8\u7387\u4f53\u7d20\u6805\u683c\u548c\u6bcf\u5e27\u6052\u5b9a\u6570\u91cf\u6837\u672c\u7684CoSLAM\u548cNICE-SLAM\u5177\u6709\u76f8\u5bf9\u7a33\u5b9a\u7684\u5185\u5b58\u4f7f\u7528\u3002NeuralRecon\u662f\u4e00\u79cd\u589e\u91cf\u5f0f\u91cd\u5efa\u65b9\u6cd5\uff0c\u968f\u7740\u5730\u56fe\u89c4\u6a21\u7684\u589e\u52a0\uff0c\u5185\u5b58\u6d88\u8017\u4e5f\u4f1a\u589e\u52a0\u3002<\/p>\n\n\n\n

6.\u603b\u7ed3<\/h2>\n\n\n\n

\u672c\u6587\u5f15\u5165\u4e86\u4e00\u79cd\u65b0\u7684\u901a\u7528SLAM\u6846\u67b6\u2014\u2014XRDSLAM\u3002\u8be5\u6846\u67b6\u63d0\u4f9b\u4e86\u4e00\u4e2a\u901a\u7528\u7684\u63a5\u53e3\u548c\u5de5\u5177\uff0c\u4f7f\u5e73\u53f0\u6613\u4e8e\u4f7f\u7528\u3002\u6a21\u5757\u5316\u8bbe\u8ba1\u4fbf\u4e8e\u96c6\u6210\u548c\u8fc1\u79fb\u4e0d\u540c\u7684\u7b97\u6cd5\uff0c\u540c\u65f6\u63d0\u4f9b\u7edf\u4e00\u7684\u7ed3\u679c\u8f93\u51fa\u548c\u8bc4\u4f30\u529f\u80fd\uff0c\u5b9e\u73b0\u4e86\u4e0d\u540cSLAM\u7b97\u6cd5\u7684\u516c\u6b63\u6bd4\u8f83\u3002XRDSLAM\u663e\u8457\u964d\u4f4e\u4e86\u4ee3\u7801\u5f00\u53d1\u7684\u6210\u672c\u5e76\u4e14\u63d0\u9ad8\u4e86\u5f00\u53d1\u6548\u7387\uff0c\u4ece\u800c\u4fc3\u8fdbSLAM\u6280\u672f\u7684\u53d1\u5c55\u3002<\/p>\n\n\n\n

\u5728\u8be5\u5e73\u53f0\u4e0a\uff0c\u96c6\u6210\u4e86\u6700\u5148\u8fdb\u7684SLAM\u7b97\u6cd5\uff08Vox-Fusion\u3001NICE-SLAM\u3001SplaTAM\u7b49\uff09\uff0c\u5e76\u4e14\u5728\u7edf\u4e00\u7684\u6846\u67b6\u5185\u8fdb\u884c\u5ba2\u89c2\u8bc4\u4f30\uff0c\u79d1\u5b66\u5730\u5206\u6790\u4e86\u4e0d\u540c\u7b97\u6cd5\u7684\u7cbe\u5ea6\u7279\u6027\u548c\u8ba1\u7b97\u6548\u7387\u3002<\/p>\n\n\n\n

\u672a\u6765\uff0c\u5c06\u7ee7\u7eed\u96c6\u6210\u66f4\u591a\u6700\u5148\u8fdb\u7684SLAM\u7b97\u6cd5\uff0c\u8fdb\u4e00\u6b65\u5b8c\u5584\u6df1\u5ea6\u5b66\u4e60SLAM\u7684\u751f\u6001\u7cfb\u7edf\uff0c\u4e5f\u6b22\u8fce\u66f4\u591a\u7684\u7814\u7a76\u8005\u5728\u8be5\u5e73\u53f0\u4e0a\u5f00\u53d1\uff0c\u8d21\u732e\u9ad8\u8d28\u91cf\u7684\u7b97\u6cd5\u548c\u4ee3\u7801\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"

\u25cf \u8bba\u6587\u94fe\u63a5\uff1ahttps:\/\/arxiv.or […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3,1],"tags":[],"class_list":["post-20442","post","type-post","status-publish","format-standard","hentry","category-technology-frontier","category-home"],"views":5,"_links":{"self":[{"href":"http:\/\/222.128.65.89:18086\/index.php\/wp-json\/wp\/v2\/posts\/20442"}],"collection":[{"href":"http:\/\/222.128.65.89:18086\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/222.128.65.89:18086\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/222.128.65.89:18086\/index.php\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/222.128.65.89:18086\/index.php\/wp-json\/wp\/v2\/comments?post=20442"}],"version-history":[{"count":1,"href":"http:\/\/222.128.65.89:18086\/index.php\/wp-json\/wp\/v2\/posts\/20442\/revisions"}],"predecessor-version":[{"id":20448,"href":"http:\/\/222.128.65.89:18086\/index.php\/wp-json\/wp\/v2\/posts\/20442\/revisions\/20448"}],"wp:attachment":[{"href":"http:\/\/222.128.65.89:18086\/index.php\/wp-json\/wp\/v2\/media?parent=20442"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/222.128.65.89:18086\/index.php\/wp-json\/wp\/v2\/categories?post=20442"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/222.128.65.89:18086\/index.php\/wp-json\/wp\/v2\/tags?post=20442"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}