公开(公告)号：EP3649579A1

公开(公告)日：2020-05-13

申请号：EP18759476.7

申请日：2018-08-06

Applicant: Siemens Aktiengesellschaft

Inventor： GONG, Yunye ,  WU, Ziyan ,  ERNST, Jan

IPC: G06K9/62 ,  G06K9/66 ,  G06K9/00 ,  G06N3/08 ,  G06N3/04

公开(公告)号：EP3580693A1

公开(公告)日：2019-12-18

申请号：EP18715433.1

申请日：2018-03-16

Applicant: Siemens Aktiengesellschaft

Inventor： PENG, Kuan-Chuan ,  WU, Ziyan ,  ERNST, Jan

IPC: G06K9/32 ,  G06K9/62

公开(公告)号：EP3824405A1

公开(公告)日：2021-05-26

申请号：EP18782543.5

申请日：2018-08-21

Applicant: Siemens Aktiengesellschaft

Inventor： INNANJE, Arun ,  PENG, Kuan-Chuan ,  WU, Ziyan ,  ERNST, Jan

IPC: G06K9/00 ,  G06K9/62 ,  G06K9/66

公开(公告)号：EP3568830A1

公开(公告)日：2019-11-20

申请号：EP17809154.2

申请日：2017-11-27

Applicant: Siemens Aktiengesellschaft

Inventor： CHANG, Ti-Chiun ,  ERNST, Jan ,  WILES, Jeremy Ralph ,  BAMBERGER, Joachim ,  SZABO, Andrei

IPC: G06T7/11 ,  G06T7/143 ,  G06T7/246

公开(公告)号：EP3404582A1

公开(公告)日：2018-11-21

申请号：EP18170902.3

申请日：2018-05-04

Applicant: Siemens Aktiengesellschaft

Inventor： PENG, Kuan-Chuan ,  WU, Ziyan ,  ERNST, Jan

IPC: G06K9/46 ,  G06K9/62

CPC classification number: G06K9/6256 ,  G06K9/00201 ,  G06K9/4628 ,  G06K9/6267 ,  G06K9/6288 ,  G06N3/04 ,  G06N3/08

Abstract: Embodiments of the present invention provide a computer-implemented method for training an RGB-D classifier for a scene classification task. The method receives task-relevant labeled depth data, task-irrelevant RGB-D data, and a given trained representation in RGB. The method simulates an RGB representation using only the task-irrelevant RGB-D data. The method builds a joint neural network using only the task-irrelevant RGB-D data and the task-relevant labeled depth data.

公开(公告)号：EP3631687A1

公开(公告)日：2020-04-08

申请号：EP18745755.1

申请日：2018-07-03

Applicant: Siemens Aktiengesellschaft

Inventor： KARANAM, Srikrishna ,  WU, Ziyan ,  PENG, Kuan-Chuan ,  ERNST, Jan

IPC: G06K9/46

公开(公告)号：EP3132291B1

公开(公告)日：2018-06-06

申请号：EP15717774.2

申请日：2015-04-09

Applicant: Siemens Aktiengesellschaft

Inventor： SUN, Shanhui ,  ERNST, Jan ,  SAPKOTA, Archana ,  RITZHAUPT-KLEISSL, Eberhard ,  WILES, Jeremy Ralph ,  CHEN, Terrence

IPC: G01W1/10

CPC classification number: G01W1/10 ,  G01J1/4204 ,  G01W1/12

Abstract: A method for predicting short-term cloud coverage includes a computer calculating an estimated cloud velocity field at a current time value based on sky images. The computer determines a segmented cloud model based on the sky images, a future sun location corresponding to a future time value, and sun pixel locations at the future time value based on the future sun location. Next, the computer applies a back-propagation algorithm to the sun pixel locations using the estimated cloud velocity field to yield propagated sun pixel locations corresponding to a previous time value. Then, the computer predicts cloud coverage for the future sun location based on the propagated sun pixel locations and the segmented cloud model.

公开(公告)号：EP3788551A1

公开(公告)日：2021-03-10

申请号：EP18735743.9

申请日：2018-05-31

Applicant: Siemens Aktiengesellschaft

Inventor： CHANG, Ti-Chiun ,  REEB, Patrick ,  BAMBERGER, Joachim ,  PENG, Kuan-Chuan ,  ERNST, Jan

IPC: G06N3/04 ,  G06N3/08

公开(公告)号：EP3710977A1

公开(公告)日：2020-09-23

申请号：EP17830077.8

申请日：2017-12-21

Applicant: Siemens Aktiengesellschaft

Inventor： CHANG, Ti-chiun ,  ERNST, Jan ,  REEB, Patrick ,  BAMBERGER, Joachim

IPC: G06K9/00 ,  G06K9/46 ,  G06N3/04 ,  G06N3/08

公开(公告)号：EP3132291A1

公开(公告)日：2017-02-22

申请号：EP15717774.2

申请日：2015-04-09

Applicant: Siemens Aktiengesellschaft

Inventor： SUN, Shanhui ,  ERNST, Jan ,  SAPKOTA, Archana ,  RITZHAUPT-KLEISSL, Eberhard ,  WILES, Jeremy Ralph ,  CHEN, Terrence

IPC: G01W1/10

CPC classification number: G01W1/10 ,  G01J1/4204 ,  G01W1/12

Abstract: A method for predicting short-term cloud coverage includes a computer calculating an estimated cloud velocity field at a current time value based on sky images. The computer determines a segmented cloud model based on the sky images, a future sun location corresponding to a future time value, and sun pixel locations at the future time value based on the future sun location. Next, the computer applies a back-propagation algorithm to the sun pixel locations using the estimated cloud velocity field to yield propagated sun pixel locations corresponding to a previous time value. Then, the computer predicts cloud coverage for the future sun location based on the propagated sun pixel locations and the segmented cloud model.

Abstract translation: 一种用于预测短期云覆盖的方法包括计算机基于天空图像计算当前时间值的估计云速度场。 计算机基于天空图像，未来太阳位置对应于未来时间值，以及基于未来太阳位置的未来时间值的太阳像素位置来确定分段云模型。 接下来，计算机使用估计的云速度场向太阳像素位置应用反向传播算法，以产生对应于先前时间值的传播的太阳像素位置。 然后，计算机基于传播的太阳像素位置和分段云模型预测未来太阳位置的云覆盖。