UPLOAD DATA
...
Guides
More examples
Lidars and Cameras Sequence
4min
A LidarsAndCamerasSeq consists of a sequence of camera images and lidar point clouds, where each frame consists on 1-9 camera images as well as 1-20 point clouds. For more documentation on what each field corresponds to in the LidarsAndCamerasSeq object please check the section related to Scene Overview.
Python
1from __future__ import absolute_import
2
3from datetime import datetime
4from typing import Optional
5from uuid import uuid4
6
7import kognic.io.model.scene.lidars_and_cameras_sequence as LCSM
8from examples.calibration.calibration import create_sensor_calibration
9from kognic.io.client import KognicIOClient
10from kognic.io.logger import setup_logging
11from kognic.io.model import CreateSceneResponse, Image, PointCloud
12
13
14def run(client: KognicIOClient, dryrun: bool = True, **kwargs) -> Optional[CreateSceneResponse]:
15 print("Creating Lidar and Camera Sequence Scene...")
16
17 lidar_sensor1 = "lidar"
18 cam_sensor1 = "RFC01"
19 cam_sensor2 = "RFC02"
20 metadata = {"location-lat": 27.986065, "location-long": 86.922623, "vehicle_id": "abg"}
21
22 # Create calibration
23 # (Please refer to the API documentation about calibration for more details)
24 calibration_spec = create_sensor_calibration(
25 f"Collection {datetime.now()}",
26 [lidar_sensor1],
27 [cam_sensor1, cam_sensor2],
28 )
29 created_calibration = client.calibration.create_calibration(calibration_spec)
30
31 scene = LCSM.LidarsAndCamerasSequence(
32 external_id=f"LCS-example-{uuid4()}",
33 frames=[
34 LCSM.Frame(
35 frame_id="1",
36 relative_timestamp=0,
37 point_clouds=[
38 PointCloud(
39 filename="./examples/resources/point_cloud_RFL01.las",
40 sensor_name=lidar_sensor1,
41 ),
42 ],
43 images=[
44 Image(
45 filename="./examples/resources/img_RFC01.jpg",
46 sensor_name=cam_sensor1,
47 ),
48 Image(
49 filename="./examples/resources/img_RFC02.jpg",
50 sensor_name=cam_sensor2,
51 ),
52 ],
53 metadata={"dut_status": "active"},
54 ),
55 LCSM.Frame(
56 frame_id="2",
57 relative_timestamp=100,
58 point_clouds=[
59 PointCloud(
60 filename="./examples/resources/point_cloud_RFL02.las",
61 sensor_name=lidar_sensor1,
62 ),
63 ],
64 images=[
65 Image(
66 filename="./examples/resources/img_RFC11.jpg",
67 sensor_name=cam_sensor1,
68 ),
69 Image(
70 filename="./examples/resources/img_RFC12.jpg",
71 sensor_name=cam_sensor2,
72 ),
73 ],
74 metadata={"dut_status": "active"},
75 ),
76 ],
77 calibration_id=created_calibration.id,
78 metadata=metadata,
79 )
80 # Create scene
81 return client.lidars_and_cameras_sequence.create(scene, dryrun=dryrun, **kwargs)
82
83
84if __name__ == "__main__":
85 setup_logging(level="INFO")
86 client = KognicIOClient()
87
88 # Project - Available via `client.project.get_projects()`
89 project = "<project-identifier>"
90
91 run(client, project)

Use dryrun to validate scene
Setting dryrun parameter to true in the method call, will validate the scene using the API but not create it.
Reuse calibration
Note that you can, and should, reuse the same calibration for multiple scenes if possible.
Ego vehicle motion (i.e. the position and rotation of the ego vehicle) is optional information that can be provided when creating LidarsAndCamerasSeqs. This information can enable a massive reduction in the time it takes to annotate static objects. Ego vehicle motion information is provided by passing a EgoVehicleMotion object to each Framein the scene.
Python
1from __future__ import absolute_import
2
3from datetime import datetime
4from typing import Optional
5from uuid import uuid4
6
7import kognic.io.model.scene.lidars_and_cameras_sequence as LCSM
8from examples.calibration.calibration import create_sensor_calibration
9from kognic.io.client import KognicIOClient
10from kognic.io.logger import setup_logging
11from kognic.io.model import CreateSceneResponse, EgoVehiclePose, Image, PointCloud, Position, RotationQuaternion
12
13
14def run(client: KognicIOClient, dryrun: bool = True, **kwargs) -> Optional[CreateSceneResponse]:
15 print("Creating Lidar and Camera Sequence Scene...")
16
17 lidar_sensor1 = "lidar"
18 cam_sensor1 = "RFC01"
19 cam_sensor2 = "RFC02"
20 cam_sensor3 = "RFC03"
21 metadata = {"location-lat": 27.986065, "location-long": 86.922623, "vehicle_id": "abg"}
22
23 # Create calibration
24 calibration_spec = create_sensor_calibration(f"Collection {datetime.now()}", [lidar_sensor1], [cam_sensor1, cam_sensor2, cam_sensor3])
25 created_calibration = client.calibration.create_calibration(calibration_spec)
26
27 scene = LCSM.LidarsAndCamerasSequence(
28 external_id=f"LCS-full-example-{uuid4()}",
29 frames=[
30 LCSM.Frame(
31 frame_id="1",
32 relative_timestamp=0,
33 point_clouds=[
34 PointCloud(filename="./examples/resources/point_cloud_RFL01.las", sensor_name=lidar_sensor1),
35 ],
36 images=[
37 Image(filename="./examples/resources/img_RFC01.jpg", sensor_name=cam_sensor1),
38 Image(filename="./examples/resources/img_RFC02.jpg", sensor_name=cam_sensor2),
39 ],
40 metadata={"dut_status": "active"},
41 ego_vehicle_pose=EgoVehiclePose(
42 position=Position(x=1.0, y=1.0, z=1.0), rotation=RotationQuaternion(w=0.01, x=1.01, y=1.01, z=1.01)
43 ),
44 ),
45 LCSM.Frame(
46 frame_id="2",
47 relative_timestamp=500,
48 point_clouds=[
49 PointCloud(filename="./examples/resources/point_cloud_RFL02.las", sensor_name=lidar_sensor1),
50 ],
51 images=[
52 Image(filename="./examples/resources/img_RFC11.jpg", sensor_name=cam_sensor1),
53 Image(filename="./examples/resources/img_RFC12.jpg", sensor_name=cam_sensor2),
54 ],
55 ego_vehicle_pose=EgoVehiclePose(
56 position=Position(x=2.0, y=2.0, z=2.0), rotation=RotationQuaternion(w=0.01, x=2.01, y=2.01, z=2.01)
57 ),
58 ),
59 ],
60 calibration_id=created_calibration.id,
61 metadata=metadata,
62 )
63 # Create scene
64 return client.lidars_and_cameras_sequence.create(scene, dryrun=dryrun, **kwargs)
65
66
67if __name__ == "__main__":
68 setup_logging(level="INFO")
69 client = KognicIOClient()
70
71 # Project - Available via `client.project.get_projects()`
72 project = "<project-identifier>"
73
74 run(client, project=project)

Coordinate systems
Note that both position and rotation for ego vehicle pose are with respect to the local coordinate system.
Shutter timings are optional metadata that may be provided when creating an Image within a Frame. Timings are two values: shutter start and end timestamp in nanoseconds since unix epoch and are specified for each image in each frame.
Python
1from __future__ import absolute_import
2
3import os.path
4from datetime import datetime
5from typing import Optional
6from uuid import uuid4
7
8import kognic.io.model.scene.lidars_and_cameras_sequence as LCSM
9from examples.calibration.calibration import create_sensor_calibration
10from examples.imu_data.create_imu_data import create_dummy_imu_data
11from kognic.io.client import KognicIOClient
12from kognic.io.logger import setup_logging
13from kognic.io.model import CreateSceneResponse, Image, ImageMetadata, PointCloud
14
15
16def run(client: KognicIOClient, dryrun: bool = True, **kwargs) -> Optional[CreateSceneResponse]:
17 print("Creating Lidar and Camera Sequence Scene...")
18
19 lidar_sensor1 = "RFL01"
20 lidar_sensor2 = "RFL02"
21 cam_sensor1 = "RFC01"
22 cam_sensor2 = "RFC02"
23 metadata = {"location-lat": 27.986065, "location-long": 86.922623, "vehicleId": "abg"}
24 examples_path = os.path.dirname(__file__)
25
26 # Create calibration
27 calibration_spec = create_sensor_calibration(f"Collection {datetime.now()}", [lidar_sensor1, lidar_sensor2], [cam_sensor1, cam_sensor2])
28 created_calibration = client.calibration.create_calibration(calibration_spec)
29
30 # Generate IMU data
31 ONE_MILLISECOND = 1000000 # one millisecond, expressed in nanos
32 start_ts = 1648200140000000000
33 end_ts = start_ts + 10 * ONE_MILLISECOND
34 imu_data = create_dummy_imu_data(start_timestamp=start_ts, end_timestamp=end_ts, samples_per_sec=1000)
35
36 scene = LCSM.LidarsAndCamerasSequence(
37 external_id=f"LCS-full-with-imu-and-shutter-example-{uuid4()}",
38 frames=[
39 LCSM.Frame(
40 frame_id="1",
41 unix_timestamp=start_ts + ONE_MILLISECOND,
42 relative_timestamp=0,
43 point_clouds=[
44 PointCloud(filename=examples_path + "/resources/point_cloud_RFL01.csv", sensor_name=lidar_sensor1),
45 PointCloud(filename=examples_path + "/resources/point_cloud_RFL02.csv", sensor_name=lidar_sensor2),
46 ],
47 images=[
48 Image(
49 filename=examples_path + "/resources/img_RFC01.jpg",
50 sensor_name=cam_sensor1,
51 metadata=ImageMetadata(
52 shutter_time_start_ns=start_ts + 0.5 * ONE_MILLISECOND, shutter_time_end_ns=start_ts + 1.5 * ONE_MILLISECOND
53 ),
54 ),
55 Image(
56 filename=examples_path + "/resources/img_RFC02.jpg",
57 sensor_name=cam_sensor2,
58 metadata=ImageMetadata(
59 shutter_time_start_ns=start_ts + 0.5 * ONE_MILLISECOND, shutter_time_end_ns=start_ts + 1.5 * ONE_MILLISECOND
60 ),
61 ),
62 ],
63 ),
64 LCSM.Frame(
65 frame_id="2",
66 unix_timestamp=start_ts + 5 * ONE_MILLISECOND,
67 relative_timestamp=4,
68 point_clouds=[
69 PointCloud(filename=examples_path + "/resources/point_cloud_RFL11.csv", sensor_name=lidar_sensor1),
70 PointCloud(filename=examples_path + "/resources/point_cloud_RFL12.csv", sensor_name=lidar_sensor2),
71 ],
72 images=[
73 Image(
74 filename=examples_path + "/resources/img_RFC11.jpg",
75 sensor_name=cam_sensor1,
76 metadata=ImageMetadata(
77 shutter_time_start_ns=start_ts + 4.5 * ONE_MILLISECOND, shutter_time_end_ns=start_ts + 5.5 * ONE_MILLISECOND
78 ),
79 ),
80 Image(
81 filename=examples_path + "/resources/img_RFC12.jpg",
82 sensor_name=cam_sensor2,
83 metadata=ImageMetadata(
84 shutter_time_start_ns=start_ts + 4.5 * ONE_MILLISECOND, shutter_time_end_ns=start_ts + 5.5 * ONE_MILLISECOND
85 ),
86 ),
87 ],
88 ),
89 ],
90 calibration_id=created_calibration.id,
91 metadata=metadata,
92 imu_data=imu_data,
93 )
94 # Create scene
95 return client.lidars_and_cameras_sequence.create(scene, dryrun=dryrun, **kwargs)
96
97
98if __name__ == "__main__":
99 setup_logging(level="INFO")
100 client = KognicIOClient()
101
102 # Project - Available via `client.project.get_projects()`
103 project = "<project-id>"
104
105 run(client, project=project)


Updated 17 Dec 2024
Did this page help you?