{
  "id": 109415,
  "title": "Points' Intensity and channel information missing?",
  "url": "/competitions/3d-object-detection-for-autonomous-vehicles/discussion/109415",
  "author_name": "soulkwon",
  "post_date": "2019-09-19T01:00:47.329000",
  "votes": 13,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hello. </p>\n\n<p>I think points' channel and intenslty information are missing in Lyft dataset compared to the nuScenes dataset of which Lyft dataset follows the format. The codes below show the issue:</p>\n\n<p>```</p>\n\n<h2>Pick a random point cloud file from each dataset</h2>\n\n<p>file_path = '../lyft/train/lidar/host-a009_lidar1_1236020739601140606.bin'\nlyft_points = load_pcl(file_path)\n​\nfile_path = '../nuscenes/trainval/samples/LIDAR_TOP/n015-2018-08-01-16-32-59+0800__LIDAR_TOP__1533112832196540.pcd.bin'\nnuscenes_points = load_pcl(file_path)\n​\nprint('lyft'); print(lyft_points)\nprint('nuscenes'); print(nuscenes_points)</p>\n\n<h2>Results</h2>\n\n<p>lyft\n[[ 58.14607      0.21440078   7.303754   100.           1.        ]\n [ 59.453796     0.4321603    7.4708657  100.           1.        ]\n [ 62.629566     0.66805637   7.876157   100.           1.        ]\n ...\n [  5.897753    -0.04891394  -1.6242886  100.           1.        ]\n [  5.897241    -0.02844508  -1.6243095  100.           1.        ]\n [  5.8966975   -0.00694822  -1.6243294  100.           1.        ]]\nnuscenes\n[[-2.9440382  -0.3490765  -1.7511001  15.          0.        ]\n [-3.0973759  -0.34716725 -1.7453089  18.          1.        ]\n [-3.2428663  -0.34577352 -1.729147   21.          2.        ]\n ...\n [-6.527841   -0.00455226  0.9174293  38.         29.        ]\n [-6.5363693  -0.0022812   1.0738865  48.         30.        ]\n [-6.6156087  -0.00000249  1.2464437  30.         31.        ]]\n```</p>\n\n<p>The 4th and 5th column should be intensity and channel, respectively. Lyft points have consistently 100 and 1. Lyft dataset doesn't provide? Can someone in Lyft confirm?</p>",
  "messages": [
    {
      "id": 629605,
      "postDate": "2019-09-19T01:00:47.330Z",
      "content": "<p>Hello. </p>\n\n<p>I think points' channel and intenslty information are missing in Lyft dataset compared to the nuScenes dataset of which Lyft dataset follows the format. The codes below show the issue:</p>\n\n<p>```</p>\n\n<h2>Pick a random point cloud file from each dataset</h2>\n\n<p>file_path = '../lyft/train/lidar/host-a009_lidar1_1236020739601140606.bin'\nlyft_points = load_pcl(file_path)\n​\nfile_path = '../nuscenes/trainval/samples/LIDAR_TOP/n015-2018-08-01-16-32-59+0800__LIDAR_TOP__1533112832196540.pcd.bin'\nnuscenes_points = load_pcl(file_path)\n​\nprint('lyft'); print(lyft_points)\nprint('nuscenes'); print(nuscenes_points)</p>\n\n<h2>Results</h2>\n\n<p>lyft\n[[ 58.14607      0.21440078   7.303754   100.           1.        ]\n [ 59.453796     0.4321603    7.4708657  100.           1.        ]\n [ 62.629566     0.66805637   7.876157   100.           1.        ]\n ...\n [  5.897753    -0.04891394  -1.6242886  100.           1.        ]\n [  5.897241    -0.02844508  -1.6243095  100.           1.        ]\n [  5.8966975   -0.00694822  -1.6243294  100.           1.        ]]\nnuscenes\n[[-2.9440382  -0.3490765  -1.7511001  15.          0.        ]\n [-3.0973759  -0.34716725 -1.7453089  18.          1.        ]\n [-3.2428663  -0.34577352 -1.729147   21.          2.        ]\n ...\n [-6.527841   -0.00455226  0.9174293  38.         29.        ]\n [-6.5363693  -0.0022812   1.0738865  48.         30.        ]\n [-6.6156087  -0.00000249  1.2464437  30.         31.        ]]\n```</p>\n\n<p>The 4th and 5th column should be intensity and channel, respectively. Lyft points have consistently 100 and 1. Lyft dataset doesn't provide? Can someone in Lyft confirm?</p>",
      "rawMarkdown": "Hello. \n\nI think points' channel and intenslty information are missing in Lyft dataset compared to the nuScenes dataset of which Lyft dataset follows the format. The codes below show the issue:\n\n```\n## Pick a random point cloud file from each dataset\nfile_path = '../lyft/train/lidar/host-a009_lidar1_1236020739601140606.bin'\nlyft_points = load_pcl(file_path)\n​\nfile_path = '../nuscenes/trainval/samples/LIDAR_TOP/n015-2018-08-01-16-32-59+0800__LIDAR_TOP__1533112832196540.pcd.bin'\nnuscenes_points = load_pcl(file_path)\n​\nprint('lyft'); print(lyft_points)\nprint('nuscenes'); print(nuscenes_points)\n\n## Results\nlyft\n[[ 58.14607      0.21440078   7.303754   100.           1.        ]\n [ 59.453796     0.4321603    7.4708657  100.           1.        ]\n [ 62.629566     0.66805637   7.876157   100.           1.        ]\n ...\n [  5.897753    -0.04891394  -1.6242886  100.           1.        ]\n [  5.897241    -0.02844508  -1.6243095  100.           1.        ]\n [  5.8966975   -0.00694822  -1.6243294  100.           1.        ]]\nnuscenes\n[[-2.9440382  -0.3490765  -1.7511001  15.          0.        ]\n [-3.0973759  -0.34716725 -1.7453089  18.          1.        ]\n [-3.2428663  -0.34577352 -1.729147   21.          2.        ]\n ...\n [-6.527841   -0.00455226  0.9174293  38.         29.        ]\n [-6.5363693  -0.0022812   1.0738865  48.         30.        ]\n [-6.6156087  -0.00000249  1.2464437  30.         31.        ]]\n```\n\nThe 4th and 5th column should be intensity and channel, respectively. Lyft points have consistently 100 and 1. Lyft dataset doesn't provide? Can someone in Lyft confirm?",
      "votes": 13
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "629605": "Hello. \n\nI think points' channel and intenslty information are missing in Lyft dataset compared to the nuScenes dataset of which Lyft dataset follows the format. The codes below show the issue:\n\n```\n## Pick a random point cloud file from each dataset\nfile_path = '../lyft/train/lidar/host-a009_lidar1_1236020739601140606.bin'\nlyft_points = load_pcl(file_path)\n​\nfile_path = '../nuscenes/trainval/samples/LIDAR_TOP/n015-2018-08-01-16-32-59+0800__LIDAR_TOP__1533112832196540.pcd.bin'\nnuscenes_points = load_pcl(file_path)\n​\nprint('lyft'); print(lyft_points)\nprint('nuscenes'); print(nuscenes_points)\n\n## Results\nlyft\n[[ 58.14607      0.21440078   7.303754   100.           1.        ]\n [ 59.453796     0.4321603    7.4708657  100.           1.        ]\n [ 62.629566     0.66805637   7.876157   100.           1.        ]\n ...\n [  5.897753    -0.04891394  -1.6242886  100.           1.        ]\n [  5.897241    -0.02844508  -1.6243095  100.           1.        ]\n [  5.8966975   -0.00694822  -1.6243294  100.           1.        ]]\nnuscenes\n[[-2.9440382  -0.3490765  -1.7511001  15.          0.        ]\n [-3.0973759  -0.34716725 -1.7453089  18.          1.        ]\n [-3.2428663  -0.34577352 -1.729147   21.          2.        ]\n ...\n [-6.527841   -0.00455226  0.9174293  38.         29.        ]\n [-6.5363693  -0.0022812   1.0738865  48.         30.        ]\n [-6.6156087  -0.00000249  1.2464437  30.         31.        ]]\n```\n\nThe 4th and 5th column should be intensity and channel, respectively. Lyft points have consistently 100 and 1. Lyft dataset doesn't provide? Can someone in Lyft confirm?"
  }
}