Automatic labeling of objects in sensor data
Aspects of the disclosure provide for automatically generating labels for sensor data. For instance, first sensor data, for a vehicle may be identified. This first sensor data may have been captured by a first sensor of the vehicle at a first location during a first point in time and may be associated with a first label for an object. Second sensor data for the vehicle may be identified. The second sensor data may have been captured by a second sensor of the vehicle at a second location at a second point in time outside of the first point in time. The second location is different from the first location. A determination may be made as to whether the object is a static object. Based on the determination that the object is a static object, the first label may be used to automatically generate a second label for the second sensor data.
1. A vehicle comprising:
a perception system including:
a first sensor rotating at a frequency and configured to generate first sensor data during a single rotation of the first sensor, the first sensor data identifying first location and orientation of the vehicle at a first point in time, and
a second sensor configured to generate second sensor data identifying a second location and orientation of the vehicle at a second point in time different from the first point in time, wherein a difference in time between the first point in time and the second point in time is based on the frequency of rotation of the first sensor; and
one or more computing devices having one or more processors configured to control the vehicle in an autonomous driving mode based on the first sensor data and the second sensor data.
2. The vehicle of claim 1 , wherein the second location and orientation of the vehicle are different from the first location and orientation of the vehicle.
3. The vehicle of claim 1 , wherein the first sensor data includes light detection and ranging (LIDAR) sensor data and the second sensor data includes one or more images.
4. The vehicle of claim 1 , wherein the first sensor data is associated with a first label for an object.
5. The vehicle of claim 4 , wherein the perception system is configured to determine that a localized position of the object did not change more than a predetermined amount during a time period between the first point in time and the second point in time.
6. The vehicle of claim 5 , wherein the first label is used to automatically generate a second label for the second sensor data.
7. The vehicle of claim 1 , wherein the frequency is 10 Hz or 10 revolutions per second.
8. The vehicle of claim 1 , wherein the first sensor data includes bounding boxes and one or more associated labels for objects detected by the perception system.
9. The vehicle of claim 8 , wherein the bounding boxes represent a three-dimensional space that includes all data points generated by a sensor corresponding to each of the objects.
10. The vehicle of claim 8 , wherein the one or more associated labels identify information about each of the objects.
11. A method of controlling a vehicle in an autonomous driving mode, the method comprising:
receiving, by one or more processors, first sensor data generated during a single rotation of a first sensor rotating at a frequency, wherein the first sensor data identifies a first location and orientation of the vehicle at a first point in time;
receiving, by the one or more processors, second sensor data generated by a second sensor at a second location and orientation of the vehicle at a second point in time different from the first point in time, wherein a difference in time between the first point in time and the second point in time is based on the frequency of rotation of the first sensor; and
controlling, by the one or more processors, the vehicle in the autonomous driving mode based on the first sensor data and the second sensor data.
12. The method of claim 11 , wherein the second location and orientation of the vehicle are different from the first location and orientation of the vehicle.
13. The method of claim 11 , wherein the first sensor data includes light detection and ranging (LIDAR) sensor data and the second sensor data includes one or more images.
14. The method of claim 11 , wherein the first sensor data is associated with a first label for an object.
15. The method of claim 14 , further comprising:
determining, by the one or more processors, that a localized position of the object did not change more than a predetermined amount during a time period between the first point in time and the second point in time.
16. The method of claim 15 , wherein the first label is used to automatically generate a second label for the second sensor data.
17. The method of claim 11 , wherein the frequency is 10 Hz or 10 revolutions per second.
18. The method of claim 11 , wherein the first sensor data includes bounding boxes and one or more associated labels for objects detected by at least one of the first sensor or the second sensor.
19. The method of claim 18 , wherein the bounding boxes represent a three-dimensional space that includes all data points generated by a sensor corresponding to each of the objects.
20. The method of claim 18 , wherein the one or more associated labels identify information about each of the objects.