Merge pull request nutonomy#11 from nutonomy/dev

Visualization scripts and minor refactoring

README.md

@@ -8,6 +8,7 @@ Welcome to the devkit of the [nuScenes](https://www.nuscenes.org) dataset.
 - [Dataset download](#dataset-download)
 - [Devkit setup](#devkit-setup)
 - [Getting started](#getting-started)
+- [Frequently asked questions](#frequently-asked-questions)
 - [Setting up a new virtual environment](#setting-up-a-new-virtual-environment)
 
 ## Changelog
@@ -57,11 +58,13 @@ In case you want to avoid downloading and setting up the data, you can also take
 Github](https://github.com/nutonomy/nuscenes-devkit/blob/master/python-sdk/tutorial.ipynb). To learn more about the dataset, go to [nuScenes.org](https://www.nuscenes.org) or take a look at the [database schema](https://github.com/nutonomy/nuscenes-devkit/blob/master/schema.md) and [annotator instructions](https://github.com/nutonomy/nuscenes-devkit/blob/master/instructions.md).
 
 ## Frequently asked questions
-- *How come some objects visible in the camera images are not annotated?* In the [annotator instructions](https://github.com/nutonomy/nuscenes-devkit/blob/master/instructions.md) we specify that an object should only be annotated if it is covered by at least one LIDAR point. This is done to have precise location annotations, speedup the annotation process and remove faraway objects.
+1) *How come some objects visible in the camera images are not annotated?* In the [annotator instructions](https://github.com/nutonomy/nuscenes-devkit/blob/master/instructions.md) we specify that an object should only be annotated if it is covered by at least one LIDAR point. This is done to have precise location annotations, speedup the annotation process and remove faraway objects.
 
-- *I have found an incorrect annotation. Can you correct it?* Please make sure that the annotation is indeed incorrect according to the [annotator instructions](https://github.com/nutonomy/nuscenes-devkit/blob/master/instructions.md). Then send an email to nuScenes@nutonomy.com. 
+2) *I have found an incorrect annotation. Can you correct it?* Please make sure that the annotation is indeed incorrect according to the [annotator instructions](https://github.com/nutonomy/nuscenes-devkit/blob/master/instructions.md). Then send an email to nuScenes@nutonomy.com.
 
-- *How can I use the RADAR data?* We recently [added features to parse and visualize RADAR point-clouds](https://github.com/nutonomy/nuscenes-devkit/pull/6). More visualization tools will follow.
+3) *How can I use the RADAR data?* We recently [added features to parse and visualize RADAR point-clouds](https://github.com/nutonomy/nuscenes-devkit/pull/6). More visualization tools will follow.
+
+4) *Why are there less sample pointclouds than samples?* See [this issue](https://github.com/nutonomy/nuscenes-devkit/issues/8). Scenes 169 and 170 overlap and going forward we will remove scene 169.
 
 ## Setting up a new virtual environment
 

python-sdk/export/export_egoposes_on_map.py

@@ -0,0 +1,24 @@
+"""
+Exports an image for each map location with all the ego poses drawn on the map.
+"""
+import os
+
+import matplotlib.pyplot as plt
+import numpy as np
+
+from nuscenes_utils.nuscenes import NuScenes
+
+# Load NuScenes class
+nusc = NuScenes()
+locations = np.unique([l['location'] for l in nusc.log])
+
+# Create output directory
+out_dir = os.path.expanduser('~/nuscenes-visualization/map-poses')
+if not os.path.isdir(out_dir):
+    os.makedirs(out_dir)
+
+for location in locations:
+    nusc.render_egoposes_on_map(log_location=location)
+    out_path = os.path.join(out_dir, 'egoposes-%s.png' % location)
+    plt.tight_layout()
+    plt.savefig(out_path)

python-sdk/examples/export_pointclouds_as_obj.py → python-sdk/export/export_pointclouds_as_obj.py

@@ -1,3 +1,8 @@
+"""
+Export fused point clouds of a scene to a Wavefront OBJ file.
+This point-cloud can be viewed in your favorite 3D rendering tool, e.g. Meshlab or Maya.
+"""
+
 import os
 import os.path as osp
 import argparse
@@ -10,22 +15,19 @@
 
 from nuscenes_utils.data_classes import PointCloud
 from nuscenes_utils.geometry_utils import view_points
-from nuscenes_utils.nuscenes import NuScenes, NuScenesExplorer
+from nuscenes_utils.nuscenes import NuScenes
 
 
-def export_scene_pointcloud(explorer: NuScenesExplorer, out_path: str, scene_token: str, channel: str='LIDAR_TOP',
+def export_scene_pointcloud(nusc: NuScenes, out_path: str, scene_token: str, channel: str='LIDAR_TOP',
                             min_dist: float=3.0, max_dist: float=30.0, verbose: bool=True) -> None:
     """
-    Export fused point clouds of a scene to a Wavefront OBJ file.
-    This point-cloud can be viewed in your favorite 3D rendering tool, e.g. Meshlab or Maya.
-    :param explorer: NuScenesExplorer instance.
+    :param nusc: NuScenes instance.
     :param out_path: Output path to write the point-cloud to.
     :param scene_token: Unique identifier of scene to render.
     :param channel: Channel to render.
     :param min_dist: Minimum distance to ego vehicle below which points are dropped.
     :param max_dist: Maximum distance to ego vehicle above which points are dropped.
     :param verbose: Whether to print messages to stdout.
-    :return: <None>
     """
 
     # Check inputs.
@@ -35,15 +37,15 @@ def export_scene_pointcloud(explorer: NuScenesExplorer, out_path: str, scene_tok
     assert channel in valid_channels, 'Input channel {} not valid.'.format(channel)
 
     # Get records from DB.
-    scene_rec = explorer.nusc.get('scene', scene_token)
-    start_sample_rec = explorer.nusc.get('sample', scene_rec['first_sample_token'])
-    sd_rec = explorer.nusc.get('sample_data', start_sample_rec['data'][channel])
+    scene_rec = nusc.get('scene', scene_token)
+    start_sample_rec = nusc.get('sample', scene_rec['first_sample_token'])
+    sd_rec = nusc.get('sample_data', start_sample_rec['data'][channel])
 
     # Make list of frames
     cur_sd_rec = sd_rec
     sd_tokens = []
     while cur_sd_rec['next'] != '':
-        cur_sd_rec = explorer.nusc.get('sample_data', cur_sd_rec['next'])
+        cur_sd_rec = nusc.get('sample_data', cur_sd_rec['next'])
         sd_tokens.append(cur_sd_rec['token'])
 
     # Write point-cloud.
@@ -53,11 +55,11 @@ def export_scene_pointcloud(explorer: NuScenesExplorer, out_path: str, scene_tok
         for sd_token in tqdm(sd_tokens):
             if verbose:
                 print('Processing {}'.format(sd_rec['filename']))
-            sc_rec = explorer.nusc.get('sample_data', sd_token)
-            sample_rec = explorer.nusc.get('sample', sc_rec['sample_token'])
+            sc_rec = nusc.get('sample_data', sd_token)
+            sample_rec = nusc.get('sample', sc_rec['sample_token'])
             lidar_token = sd_rec['token']
-            lidar_rec = explorer.nusc.get('sample_data', lidar_token)
-            pc = PointCloud.from_file(osp.join(explorer.nusc.dataroot, lidar_rec['filename']))
+            lidar_rec = nusc.get('sample_data', lidar_token)
+            pc = PointCloud.from_file(osp.join(nusc.dataroot, lidar_rec['filename']))
 
             # Get point cloud colors.
             coloring = np.ones((3, pc.points.shape[1])) * -1
@@ -68,7 +70,7 @@ def export_scene_pointcloud(explorer: NuScenesExplorer, out_path: str, scene_tok
 
             # Points live in their own reference frame. So they need to be transformed via global to the image plane.
             # First step: transform the point cloud to the ego vehicle frame for the timestamp of the sweep.
-            cs_record = explorer.nusc.get('calibrated_sensor', lidar_rec['calibrated_sensor_token'])
+            cs_record = nusc.get('calibrated_sensor', lidar_rec['calibrated_sensor_token'])
             pc.rotate(Quaternion(cs_record['rotation']).rotation_matrix)
             pc.translate(np.array(cs_record['translation']))
 
@@ -81,7 +83,7 @@ def export_scene_pointcloud(explorer: NuScenesExplorer, out_path: str, scene_tok
                 print('Distance filter: Keeping %d of %d points...' % (keep.sum(), len(keep)))
 
             # Second step: transform to the global frame.
-            poserecord = explorer.nusc.get('ego_pose', lidar_rec['ego_pose_token'])
+            poserecord = nusc.get('ego_pose', lidar_rec['ego_pose_token'])
             pc.rotate(Quaternion(poserecord['rotation']).rotation_matrix)
             pc.translate(np.array(poserecord['translation']))
 
@@ -94,13 +96,14 @@ def export_scene_pointcloud(explorer: NuScenesExplorer, out_path: str, scene_tok
                     f.write("v {v[0]:.8f} {v[1]:.8f} {v[2]:.8f} {c[0]:.4f} {c[1]:.4f} {c[2]:.4f}\n".format(v=v, c=c/255.0))
 
             if not sd_rec['next'] == "":
-                sd_rec = explorer.nusc.get('sample_data', sd_rec['next'])
+                sd_rec = nusc.get('sample_data', sd_rec['next'])
 
 
-def pointcloud_color_from_image(nusc, pointsensor_token: str, camera_token: str) -> Tuple[np.array, np.array]:
+def pointcloud_color_from_image(nusc: NuScenes, pointsensor_token: str, camera_token: str) -> Tuple[np.array, np.array]:
     """
     Given a point sensor (lidar/radar) token and camera sample_data token, load point-cloud and map it to the image
     plane, then retrieve the colors of the closest image pixels.
+    :param nusc: NuScenes instance.
     :param pointsensor_token: Lidar/radar sample_data token.
     :param camera_token: Camera sample data token.
     :return (coloring <np.float: 3, n>, mask <np.bool: m>). Returns the colors for n points that reproject into the
@@ -165,10 +168,10 @@ def pointcloud_color_from_image(nusc, pointsensor_token: str, camera_token: str)
     parser = argparse.ArgumentParser(description='Export a scene in Wavefront point cloud format.',
                                      formatter_class=argparse.ArgumentDefaultsHelpFormatter)
     parser.add_argument('--scene', default='scene-0061', type=str, help='Name of a scene, e.g. scene-0061')
-    parser.add_argument('--out_dir', default='', type=str, help='Output folder')
+    parser.add_argument('--out_dir', default='~/nuscenes-visualization/pointclouds', type=str, help='Output folder')
     parser.add_argument('--verbose', default=0, type=int, help='Whether to print outputs to stdout')
     args = parser.parse_args()
-    out_dir = args.out_dir
+    out_dir = os.path.expanduser(args.out_dir)
     scene_name = args.scene
     verbose = bool(args.verbose)
 
@@ -188,4 +191,4 @@ def pointcloud_color_from_image(nusc, pointsensor_token: str, camera_token: str)
     scene_tokens = [s['token'] for s in nusc.scene if s['name'] == scene_name]
     assert len(scene_tokens) == 1, 'Error: Invalid scene %s' % scene_name
 
-    export_scene_pointcloud(nusc.explorer, out_path, scene_tokens[0], channel='LIDAR_TOP', verbose=verbose)
+    export_scene_pointcloud(nusc, out_path, scene_tokens[0], channel='LIDAR_TOP', verbose=verbose)

python-sdk/export/export_scene_videos.py

@@ -0,0 +1,23 @@
+"""
+Exports a video of each scene (with annotations) to disk.
+"""
+import os
+
+from nuscenes_utils.nuscenes import NuScenes
+
+# Load NuScenes class
+nusc = NuScenes()
+scene_tokens = [s['token'] for s in nusc.scene]
+
+# Create output directory
+out_dir = os.path.expanduser('~/nuscenes-visualization/scene-videos')
+if not os.path.isdir(out_dir):
+    os.makedirs(out_dir)
+
+# Write videos to disk
+for scene_token in scene_tokens:
+    scene = nusc.get('scene', scene_token)
+    print('Writing scene %s' % scene['name'])
+    out_path = os.path.join(out_dir, scene['name']) + '.avi'
+    if not os.path.exists(out_path):
+        nusc.render_scene(scene['token'], out_path=out_path)