METHOD AND CONTROL UNIT FOR TRAINING AN OBJECT DETECTOR

• Main Authors: ULRICH MICHAEL, GLAESER CLAUDIUS, DREWS FLORIAN, FAION FLORIAN, ROSENBAUM LARS, TIMM FABIAN
• Format: Patent
• Language: English; Korean
• Published: 01.08.2023
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE ORDIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; COUNTING; PERFORMING OPERATIONS; PHYSICS; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TOTHE CONTROL OF A PARTICULAR SUB-UNIT; TRANSPORTING; VEHICLES IN GENERAL

The present invention relates to a method for training an object detector designed to detect an object in sensor data from a sensor. This method (1, 20) comprises the following steps of: providing first sensor data from the sensor (3, 22); providing (4, 23) an object representation assigned to the first sensor data; transferring the object representation to a sensor model designed to map the object representation to the sensor data from the sensor (5, 24); assigning the object representation to second sensor data by the sensor model (6, 25); and training an object detector based on the second sensor data (7, 26). 본 발명은 센서로부터의 센서 데이터에서 물체를 검출하도록 설계된 물체 검출기를 트레이닝시키는 방법에 관한 것이며, 이 방법(1, 20)은 센서로부터 제 1 센서 데이터를 제공하는 단계(3, 22); 상기 제 1 센서 데이터에 할당된 물체 표현을 제공하는 단계(4, 23); 물체 표현을 센서의 센서 데이터에 매핑하도록 설계된 센서 모델로 상기 물체 표현을 전송하는 단계(5, 24); 상기 센서 모델에 의해 상기 물체 표현을 제 2 센서 데이터에 할당하는 단계(6, 25); 및 상기 제 2 센서 데이터에 기초하여 물체 검출기를 트레이닝시키는 단계(7, 26)를 포함한다.