View source for US Patent Application 18324002. METHOD FOR AUTOMATIC OBJECT AND/OR SEGMENT LABELING OF SENSOR TARGET DATA, USE OF SUCH LABELED SENSOR TARGET DATA, COMPUTER PROGRAM, AND CONTROL DEVICE OR CENTRAL OR ZONAL COMPUTING MODULE simplified abstract

=METHOD FOR AUTOMATIC OBJECT AND/OR SEGMENT LABELING OF SENSOR TARGET DATA, USE OF SUCH LABELED SENSOR TARGET DATA, COMPUTER PROGRAM, AND CONTROL DEVICE OR CENTRAL OR ZONAL COMPUTING MODULE=


==Organization Name==

[[:Category:Robert Bosch GmbH|Robert Bosch GmbH]]


[[Category:Robert Bosch GmbH]]

==Inventor(s)==

[[:Category:Jerg Pfeil of Cleebronn (DE)|Jerg Pfeil of Cleebronn (DE)]][[Category:Jerg Pfeil of Cleebronn (DE)]]

==METHOD FOR AUTOMATIC OBJECT AND/OR SEGMENT LABELING OF SENSOR TARGET DATA, USE OF SUCH LABELED SENSOR TARGET DATA, COMPUTER PROGRAM, AND CONTROL DEVICE OR CENTRAL OR ZONAL COMPUTING MODULE - A simplified explanation of the abstract==


This abstract first appeared for US patent application 18324002 titled 'METHOD FOR AUTOMATIC OBJECT AND/OR SEGMENT LABELING OF SENSOR TARGET DATA, USE OF SUCH LABELED SENSOR TARGET DATA, COMPUTER PROGRAM, AND CONTROL DEVICE OR CENTRAL OR ZONAL COMPUTING MODULE

==Simplified Explanation==

- The patent application describes a method for automatically labeling objects and segments in sensor data from vehicle target sensors.
- The method involves capturing a sequence of camera images and generating an environment representation of the vehicle based on these images.
- The method then uses a machine recognition method to recognize objects in the environment and estimate their positions based on the camera images.
- The method classifies points in the environment representation based on the recognized objects and estimated positions.
- Distance data is captured using distance sensors, and the environment representation is adjusted based on this data.
- Finally, a synthetic image of the environment is calculated from a virtual perspective of observation using the adjusted environment representation.


==Original Abstract Submitted==

A method for automatic object and/or segment labeling of sensor target data of at least one vehicle target sensor. The method comprises first capturing of at least one sequence of camera images; generating an environment representation of the vehicle as a function of the captured sequence; recognizing at least one object in the environment by a learned machine recognition method as a function of a captured camera image; ascertaining an estimated position of the object as a function of the camera image; classifying a point of the environment representation based on the recognized object and the ascertained estimated position; and second capturing of distance data using at least one distance sensor. The generated environment representation is adjusted as a function of the captured distance data. A calculation of a synthetic image of the environment from a virtual perspective of observation takes place based on the adjusted environment representation.



[[Category:G06V20/56]]
[[Category:G06V10/764]]
[[Category:G06V10/774]]
[[Category:G06T7/70]]