Invention Grant
- Patent Title: Transparency-characteristic based object classification for automated vehicle
-
Application No.: US15467037Application Date: 2017-03-23
-
Publication No.: US10366294B2Publication Date: 2019-07-30
- Inventor: Junqing Wei , Wenda Xu
- Applicant: Aptiv Technologies Limited
- Applicant Address: BB St. Michael
- Assignee: Aptiv Technologies Limited
- Current Assignee: Aptiv Technologies Limited
- Current Assignee Address: BB St. Michael
- Agency: Fish & Richardson P.C.
- Main IPC: G06K9/00
- IPC: G06K9/00 ; G05D1/00 ; G05D1/02 ; G08G1/16 ; G06K9/46 ; G06K9/62 ; G01S17/08 ; G01S17/02 ; G01S17/87 ; G01S17/93 ; G01S7/40 ; G01S13/93

Abstract:
An object classification system for an automated vehicle includes a lidar and/or a camera, and a controller. The controller determines a lidar-outline and/or a camera-outline of an object. Using the lidar, the controller determines a transparency-characteristic of the object based on instances of spot-distances from within the lidar-outline of the object that correspond to a backdrop-distance. Using the camera, the controller determines a transparency-characteristic of the object based on instances of pixel-color within the camera-outline that correspond to a backdrop-color. The transparency-characteristic may also be determined based on a combination of information from the lidar and the camera. The controller operates the host-vehicle to avoid the object when the transparency-characteristic is less than a transparency-threshold.
Public/Granted literature
- US20180276484A1 TRANSPARENCY-CHARACTERISTIC BASED OBJECT CLASSIFICATION FOR AUTOMATED VEHICLE Public/Granted day:2018-09-27
Information query