Velocity determination with a scanned lidar system
A scanning imaging sensor is configured to sense an environment through which a vehicle is moving. A method for determining one or velocities associated with objects in the environment includes generating features from the first set of scan lines and the second set of scan lines, the two sets corresponding to two instances in time. The method further includes generating a collection of candidate velocities based on feature locations and time differences, the features selected pairwise with one from the first set and another from the second set. Furthermore, the method includes analyzing the distribution of candidate velocities, for example, by identifying one or more modes from the collection of the candidate velocities.
1. A method of determining a velocity of an object using an imaging sensor, the method comprising:
generating, by one or more processors, a first set of features of the object and a second set of features of the object using data collected by the imaging sensor during a first time interval and a second time interval, respectively, wherein generating a feature in the first set or second set of features includes detecting a center of a line segment formed by consecutive points in a scan line to identify a flat surface of the object;
computing, by the one or more processors, a plurality of candidate velocities based on differences in spatial positions of features selected pairwise from the first set and the second set; and
computing, by the one or more processors, the velocity of the object based on the distribution of the plurality of candidate velocities.
2. The method of claim 1 , wherein generating the first set of features and the second set of features is based on a first set of scan lines of the imaging sensor and a second set of scan lines of the imaging sensor.
3. The method of claim 2 , wherein the imaging sensor is a lidar sensor.
4. The method of claim 2 , wherein:
the first set of scan lines is captured when the imaging sensor traverses a field of regard in a first instance; and
the second set of scan lines is captured when the imaging sensor traverses the field of regard in a second instance immediately following the first instance.
5. The method of claim 4 , wherein in the first and second instances the imaging sensor traverses the field of regard in opposite directions.
6. The method of claim 2 , wherein:
the first set of scan lines corresponds to a first portion of a field of regard of the imaging sensor; and
the second set of scan lines belongs to a second portion of the field of regard partially overlapping the first portion of the field of regard.
7. The method of claim 1 , wherein generating a feature in the first set or second set of features further includes:
detecting a discontinuity of positions of points in the scan line to identify an edge of the object.
8. The method of claim 1 , wherein generating a feature in the first set or second set of features further includes:
detecting a discontinuity of slope of points in the scan line to identify a corner of the object.
9. The method of claim 1 , wherein:
generating the first set of features and the second set of features further includes determining a type of each of the features; and
computing the plurality of the candidate velocities includes restricting the pairwise comparisons to the features of the same type.
10. The method of claim 1 , wherein computing the velocity of the object includes:
determining a mode in the distribution of the plurality of candidate velocities.
11. The method of claim 10 , wherein determining the mode in the distribution of the candidate velocities includes:
determining how many of the plurality of candidate velocities belong to each of a multiplicity of bins, each bin corresponding to a respective range;
selecting, from the multiplicity of bins, a subset of one or more bins to which more than a threshold number of candidate velocities belong; and
determining the velocity of the object using the selected subset.
12. The method of claim 11 , wherein:
when only one bin in the multiplicity of bins has a count of candidate velocities above the threshold number, the method further comprises:
computing the velocity of the object by averaging the candidate velocities within the one bin.
13. The method of claim 11 , wherein:
when two or more adjacent bins in the multiplicity of bins have counts above the threshold, the method further comprises:
computing the velocity of the object by averaging the candidate velocities within the two or more adjacent bins.
14. The method of claim 10 , wherein determining the mode in the distribution of the candidate velocities comprises clustering the candidate velocities.
15. The method of claim 1 , further comprising:
identifying a first mode and a second mode in the distribution of the candidate velocities; and
computing the velocity of the object and a rotational velocity of the object based at least in part on the first mode and the second mode.
16. The method of claim 1 , further comprising:
identifying a first mode and a second mode in the distribution of the candidate velocities; and
determining, in response to the first and second mode differing by greater than a particular threshold amount, that the object includes two separate objects.
17. An imaging system configured to:
generate, by one or more processors, a first set of features of the object and a second set of features of the object using data collected by the imaging sensor during a first time interval and a second time interval, respectively, wherein generating a feature in the first set or second set of features includes detecting a center of a line segment formed by consecutive points in a scan line to identify a flat surface of the object;
compute, by the one or more processors, a plurality of candidate velocities based on differences in spatial positions of features selected pairwise from the first set and the second set; and
compute, by the one or more processors, the velocity of the object based on the distribution of the plurality of candidate velocities.
18. The imaging system of claim 17 , wherein the imaging system is a lidar system, and generating the first set of features and the second set of features is based on a first set of scan lines of the lidar and a second set of scan lines of the lidar.
19. The imaging system of claim 18 , wherein the lidar system is configured to obtain the first set of scan lines and the second set of scan lines by rescanning at least a portion of a field of regard of the lidar in the course of obtaining a single frame.
20. The imaging system of claim 17 , wherein generating a feature in the first set or second set of features further includes:
detecting a discontinuity of positions of points in the scan line to identify an edge of the object.
21. The imaging system of claim 17 , wherein:
generating the first set of features and the second set of features further includes determining a type of each of the features; and
computing the plurality of the candidate velocities includes restricting the pairwise comparisons to the features of the same type.