公开(公告)号：US11961243B2

公开(公告)日：2024-04-16

申请号：US17187228

申请日：2021-02-26

Applicant: NVIDIA Corporation

Inventor： Dong Zhang ,  Sangmin Oh ,  Junghyun Kwon ,  Baris Evrim Demiroz ,  Tae Eun Choe ,  Minwoo Park ,  Chethan Ningaraju ,  Hao Tsui ,  Eric Viscito ,  Jagadeesh Sankaran ,  Yongqing Liang

IPC: G06T7/00 ,  B60W60/00 ,  G06F18/214 ,  G06N3/08 ,  G06T7/246 ,  G06V10/25 ,  G06V10/75 ,  G06V20/58 ,  G06V20/56

CPC classification number: G06T7/246 ,  B60W60/001 ,  G06F18/2148 ,  G06N3/08 ,  G06V10/25 ,  G06V10/751 ,  G06V20/58 ,  G06V20/56

Abstract: A geometric approach may be used to detect objects on a road surface. A set of points within a region of interest between a first frame and a second frame are captured and tracked to determine a difference in location between the set of points in two frames. The first frame may be aligned with the second frame and the first pixel values of the first frame may be compared with the second pixel values of the second frame to generate a disparity image including third pixels. One or more subsets of the third pixels that have a value above a first threshold may be combined, and the third pixels may be scored and associated with disparity values for each pixel of the one or more subsets of the third pixels. A bounding shape may be generated based on the scoring.

公开(公告)号：US11906660B2

公开(公告)日：2024-02-20

申请号：US17005788

申请日：2020-08-28

Applicant: NVIDIA Corporation

Inventor： Tilman Wekel ,  Sangmin Oh ,  David Nister ,  Joachim Pehserl ,  Neda Cvijetic ,  Ibrahim Eden

IPC: G01S7/00 ,  G01S7/48 ,  G01S17/894 ,  G01S7/481 ,  G01S17/931 ,  G06V10/764 ,  G06V10/80 ,  G06V10/82 ,  G06V20/58 ,  G01S7/28

CPC classification number: G01S7/4802 ,  G01S7/481 ,  G01S17/894 ,  G01S17/931 ,  G06V10/764 ,  G06V10/80 ,  G06V10/82 ,  G06V20/58 ,  G01S7/28

Abstract: In various examples, a deep neural network (DNN) may be used to detect and classify animate objects and/or parts of an environment. The DNN may be trained using camera-to-LiDAR cross injection to generate reliable ground truth data for LiDAR range images. For example, annotations generated in the image domain may be propagated to the LiDAR domain to increase the accuracy of the ground truth data in the LiDAR domain—e.g., without requiring manual annotation in the LiDAR domain. Once trained, the DNN may output instance segmentation masks, class segmentation masks, and/or bounding shape proposals corresponding to two-dimensional (2D) LiDAR range images, and the outputs may be fused together to project the outputs into three-dimensional (3D) LiDAR point clouds. This 2D and/or 3D information output by the DNN may be provided to an autonomous vehicle drive stack to enable safe planning and control of the autonomous vehicle.

公开(公告)号：US11790230B2

公开(公告)日：2023-10-17

申请号：US17723195

申请日：2022-04-18

Applicant: NVIDIA Corporation

Inventor： Yilin Yang ,  Bala Siva Sashank Jujjavarapu ,  Pekka Janis ,  Zhaoting Ye ,  Sangmin Oh ,  Minwoo Park ,  Daniel Herrera Castro ,  Tommi Koivisto ,  David Nister

IPC: G06K9/00 ,  G06N3/08 ,  B60W30/14 ,  B60W60/00 ,  G06V20/56 ,  G06F18/214 ,  G06V10/762

CPC classification number: G06N3/08 ,  B60W30/14 ,  B60W60/0011 ,  G06F18/2155 ,  G06V10/763 ,  G06V20/56

Abstract: In various examples, a deep neural network (DNN) is trained to accurately predict, in deployment, distances to objects and obstacles using image data alone. The DNN may be trained with ground truth data that is generated and encoded using sensor data from any number of depth predicting sensors, such as, without limitation, RADAR sensors, LIDAR sensors, and/or SONAR sensors. Camera adaptation algorithms may be used in various embodiments to adapt the DNN for use with image data generated by cameras with varying parameters—such as varying fields of view. In some examples, a post-processing safety bounds operation may be executed on the predictions of the DNN to ensure that the predictions fall within a safety-permissible range.

公开(公告)号：US20230099494A1

公开(公告)日：2023-03-30

申请号：US17489346

申请日：2021-09-29

Applicant: NVIDIA Corporation

Inventor： Mehmet Kocamaz ,  Neeraj Sajjan ,  Sangmin Oh ,  David Nister ,  Junghyun Kwon ,  Minwoo Park

IPC: G06K9/00 ,  G06K9/32 ,  G06N3/08

Abstract: In various examples, live perception from sensors of an ego-machine may be leveraged to detect objects and assign the objects to bounded regions (e.g., lanes or a roadway) in an environment of the ego-machine in real-time or near real-time. For example, a deep neural network (DNN) may be trained to compute outputs—such as output segmentation masks—that may correspond to a combination of object classification and lane identifiers. The output masks may be post-processed to determine object to lane assignments that assign detected objects to lanes in order to aid an autonomous or semi-autonomous machine in a surrounding environment.

公开(公告)号：US11531088B2

公开(公告)日：2022-12-20

申请号：US16836618

申请日：2020-03-31

Applicant: NVIDIA Corporation

Inventor： Alexander Popov ,  Nikolai Smolyanskiy ,  Ryan Oldja ,  Shane Murray ,  Tilman Wekel ,  David Nister ,  Joachim Pehserl ,  Ruchi Bhargava ,  Sangmin Oh

IPC: G01S7/41 ,  G01S13/86 ,  G06N3/04 ,  G06N3/08 ,  G01S13/89

Abstract: In various examples, a deep neural network(s) (e.g., a convolutional neural network) may be trained to detect moving and stationary obstacles from RADAR data of a three dimensional (3D) space. In some embodiments, ground truth training data for the neural network(s) may be generated from LIDAR data. More specifically, a scene may be observed with RADAR and LIDAR sensors to collect RADAR data and LIDAR data for a particular time slice. The RADAR data may be used for input training data, and the LIDAR data associated with the same or closest time slice as the RADAR data may be annotated with ground truth labels identifying objects to be detected. The LIDAR labels may be propagated to the RADAR data, and LIDAR labels containing less than some threshold number of RADAR detections may be omitted. The (remaining) LIDAR labels may be used to generate ground truth data.

公开(公告)号：US20220297706A1

公开(公告)日：2022-09-22

申请号：US17698695

申请日：2022-03-18

Applicant: NVIDIA Corporation

Inventor： Hans Jonas Nilsson ,  Michael Cox ,  Sangmin Oh ,  Joachim Pehserl ,  Aidin Ehsanibenafati

IPC: B60W50/00 ,  G05B13/02 ,  B60W60/00

Abstract: In various examples, systems and methods are disclosed that perform sensor fusion using rule-based and learned processing methods to take advantage of the accuracy of learned approaches and the decomposition benefits of rule-based approaches for satisfying higher levels of safety requirements. For example, in-parallel and/or in-serial combinations of early rule-based sensor fusion, late rule-based sensor fusion, early learned sensor fusion, or late learned sensor fusion may be used to solve various safety goals associated with various required safety levels at a high level of accuracy and precision. In embodiments, learned sensor fusion may be used to make more conservative decisions than the rule-based sensor fusion (as determined using, e.g., severity (S), exposure (E), and controllability (C) (SEC) associated with a current safety goal), but the rule-based sensor fusion may be relied upon where the learned sensor fusion decision may be less conservative than the corresponding rule-based sensor fusion.

公开(公告)号：US11195331B2

公开(公告)日：2021-12-07

申请号：US16820164

申请日：2020-03-16

Applicant: NVIDIA Corporation

Inventor： Dongwoo Lee ,  Junghyun Kwon ,  Sangmin Oh ,  Wenchao Zheng ,  Hae-Jong Seo ,  David Nister ,  Berta Rodriguez Hervas

IPC: G06T17/30 ,  G06T7/40

Abstract: A neural network may be used to determine corner points of a skewed polygon (e.g., as displacement values to anchor box corner points) that accurately delineate a region in an image that defines a parking space. Further, the neural network may output confidence values predicting likelihoods that corner points of an anchor box correspond to an entrance to the parking spot. The confidence values may be used to select a subset of the corner points of the anchor box and/or skewed polygon in order to define the entrance to the parking spot. A minimum aggregate distance between corner points of a skewed polygon predicted using the CNN(s) and ground truth corner points of a parking spot may be used simplify a determination as to whether an anchor box should be used as a positive sample for training.

公开(公告)号：US11182916B2

公开(公告)日：2021-11-23

申请号：US16728598

申请日：2019-12-27

Applicant: NVIDIA Corporation

Inventor： Yilin Yang ,  Bala Siva Sashank Jujjavarapu ,  Pekka Janis ,  Zhaoting Ye ,  Sangmin Oh ,  Minwoo Park ,  Daniel Herrera Castro ,  Tommi Koivisto ,  David Nister

IPC: G06K9/00 ,  G06T7/536

Abstract: In various examples, a deep neural network (DNN) is trained to accurately predict, in deployment, distances to objects and obstacles using image data alone. The DNN may be trained with ground truth data that is generated and encoded using sensor data from any number of depth predicting sensors, such as, without limitation, RADAR sensors, LIDAR sensors, and/or SONAR sensors. Camera adaptation algorithms may be used in various embodiments to adapt the DNN for use with image data generated by cameras with varying parameters—such as varying fields of view. In some examples, a post-processing safety bounds operation may be executed on the predictions of the DNN to ensure that the predictions fall within a safety-permissible range.

公开(公告)号：US11170299B2

公开(公告)日：2021-11-09

申请号：US16813306

申请日：2020-03-09

Applicant: NVIDIA Corporation

Inventor： Junghyun Kwon ,  Yilin Yang ,  Bala Siva Sashank Jujjavarapu ,  Zhaoting Ye ,  Sangmin Oh ,  Minwoo Park ,  David Nister

IPC: G06N3/08 ,  B60W30/14 ,  B60W60/00 ,  G06K9/62 ,  G06K9/00

Abstract: In various examples, a deep neural network (DNN) is trained—using image data alone—to accurately predict distances to objects, obstacles, and/or a detected free-space boundary. The DNN may be trained with ground truth data that is generated using sensor data representative of motion of an ego-vehicle and/or sensor data from any number of depth predicting sensors—such as, without limitation, RADAR sensors, LIDAR sensors, and/or SONAR sensors. The DNN may be trained using two or more loss functions each corresponding to a particular portion of the environment that depth is predicted for, such that—in deployment—more accurate depth estimates for objects, obstacles, and/or the detected free-space boundary are computed by the DNN. In some embodiments, a sampling algorithm may be used to sample depth values corresponding to an input resolution of the DNN from a predicted depth map of the DNN at an output resolution of the DNN.

公开(公告)号：US20210156960A1

公开(公告)日：2021-05-27

申请号：US16836583

申请日：2020-03-31

Applicant: NVIDIA Corporation

Inventor： Alexander Popov ,  Nikolai Smolyanskiy ,  Ryan Oldja ,  Shane Murray ,  Tilman Wekel ,  David Nister ,  Joachim Pehserl ,  Ruchi Bhargava ,  Sangmin Oh

IPC: G01S7/41 ,  G06N3/08 ,  G06T7/73 ,  G06T7/246 ,  G01S13/931

Abstract: In various examples, a deep neural network(s) (e.g., a convolutional neural network) may be trained to detect moving and stationary obstacles from RADAR data of a three dimensional (3D) space, in both highway and urban scenarios. RADAR detections may be accumulated, ego-motion-compensated, orthographically projected, and fed into a neural network(s). The neural network(s) may include a common trunk with a feature extractor and several heads that predict different outputs such as a class confidence head that predicts a confidence map and an instance regression head that predicts object instance data for detected objects. The outputs may be decoded, filtered, and/or clustered to form bounding shapes identifying the location, size, and/or orientation of detected object instances. The detected object instances may be provided to an autonomous vehicle drive stack to enable safe planning and control of the autonomous vehicle.