| CPC G01S 7/497 (2013.01) [G01S 17/86 (2020.01); G01S 17/89 (2013.01); G06T 7/50 (2017.01); G06T 2207/10028 (2013.01)] | 20 Claims |
|
1. A method comprising:
obtaining, using a first, calibrated sensor, a first plurality of raw sensor measurements of an environment, wherein each raw sensor measurement comprises a depth value being related to a distance from the sensor to a point on a surface in the environment in a sensor coordinate frame;
determining, from the first plurality of raw sensor measurements, a Truncated Signed Distance Field (TSDF)-based model of surfaces in the environment;
obtaining, using a second sensor, a second plurality of raw sensor measurements of the environment;
determining, based on the second plurality of raw sensor measurements and initial values of a set of calibration parameters of the second sensor, an initial multi-dimensional point cloud representation of the environment, the multi-dimensional point cloud representation comprising a plurality of data points each being defined by multiple values representing respective coordinates in a vehicle coordinate frame; and
determining refined values of the set of calibration parameters of the second sensor, including evaluating an objective function that measures a difference between, for each of one or more of the plurality of data points in the multi-dimensional point cloud representation, (i) the multiple values that define the data point and (ii) multiple values that define a target data point derived from the Truncated Signed Distance Field (TSDF)-based model of surfaces in the environment.
|