Method and system for generating ground truth data for machine learning of recognizer

A method comprising:

• acquiring first surrounding environment data by a first sensor, wherein the first sensor is a LiDAR;
• acquiring second surrounding environment data by a second sensor at a same time as an acquisition time of the first surrounding environment data and in a same region as an acquisition region of the first surrounding environment data;
• acquiring a first recognition result by inputting the first surrounding environment data into a first recognizer machine-learned using surrounding environment data acquired by the first sensor;
• projecting the first recognition result onto the second surrounding environment data by transformation from a coordinate system of the first sensor to a coordinate system of the second sensor;
• generating ground truth data for machine-learning a second recognizer configured to recognize surrounding environment data acquired by the second sensor, based on the second surrounding environment data onto which the first recognition result is projected;
• inputting the first surrounding environment data into a velocity estimator configured to estimate velocity of point cloud data acquired by the first sensor to acquire a velocity estimation result;
• and
• further projecting the velocity estimation result onto the second surrounding environment data onto which the first recognition result is projected,
• wherein the generating the ground truth data comprises generating the ground truth data based on the second surrounding environment data onto which the first recognition result and the velocity estimation result are projected.

A method comprising:

• acquiring first surrounding environment data by a first sensor, wherein the first sensor is a LiDAR;
• acquiring second surrounding environment data by a second sensor at a same time as an acquisition time of the first surrounding environment data and in a same region as an acquisition region of the first surrounding environment data;
• acquiring a first recognition result by inputting the first surrounding environment data into a first recognizer machine-learned using surrounding environment data acquired by the first sensor;
• projecting the first recognition result onto the second surrounding environment data by transformation from a coordinate system of the first sensor to a coordinate system of the second sensor;
• generating ground truth data for machine-learning a second recognizer configured to recognize surrounding environment data acquired by the second sensor, based on the second surrounding environment data onto which the first recognition result is projected;
• extracting a contour shape of an object from point cloud data acquired by the first sensor;
• and
• further projecting the contour shape onto the second surrounding environment data onto which the first recognition result is projected,
• wherein the generating the ground truth data comprises generating the ground truth data based on the second surrounding environment data onto which the first recognition result and the contour shape are projected.