Baidu Online Network Technology (beijing) Co., Ltd.

Samsung Electronics Co., Ltd.

Apollo Intelligent Driving Technology (beijing) Co., Ltd.

Beijing Baidu Netcom Science and Technology Co., Ltd.

Xun Sun

Tian Xia

Han Gao

Ji Yeun Kim

Guo Qiang Li

Maolin Chen

Lujin Gong

Zhilan Hu

Ming Wang

Rui Wang

Fan Zhang

Rong Liu

Yuqiang Zhai

Yucheng Wang

Chang Kyu Choi

Hao Wang

Liang Wang

Kee Chang Lee

Shufu Xie

Ji Wan

Dongming Chen

Guo Li

Mingjian Liu

Embodiments of the present provide a cross-camera obstacle tracking method, system and medium. The method includes: obtaining obstacle tracking results of images captured by at least two cameras, wherein each obstacle tracking result comprises results after identifying and tracking at least one obstacle in the image captured by each camera, and each obstacle is labeled with a local identifier for each camera; in response to a triggering condition of a main camera in the cameras to fusion of cross-camera obstacles being satisfied, establishing a mapping relation between local identifiers of the obstacles in the cameras and global identifiers according to the obstacle tracking result of each of the cameras; determining similarities of obstacles according to the obstacle tracking result of each of the cameras, and performing fusion of the global identifiers on the same obstacle according to the similarities; and obtaining final obstacle tracking results labeled with the global identifiers.

Cross-camera obstacle tracking method, system and medium

A method, device, apparatus and a computer-readable storage medium for detecting a height of an obstacle are provided. The method can include: acquiring observation data of a plurality of reference obstacles from a frame; according to the observation data of each of the reference obstacles, fitting a function F: Z=F(ymax), wherein the observation data of a reference obstacle comprises a longitudinal coordinate of a bottom of the reference obstacle in the frame, and a distance between the reference obstacle and a camera capturing the frame; and determining a distance between the obstacle to be detected and the camera, according to the longitudinal coordinate of the bottom of the obstacle in the frame and the function F; and determining an evaluation value of the height of the obstacle to be detected according to the distance between the obstacle to be detected and the camera.

Method, device, apparatus and storage medium for detecting a height of an obstacle

Embodiments of the present disclosure disclose a method and an apparatus for detecting an obstacle, an electronic device, a vehicle and a storage medium. The method includes: obtaining a depth map around a vehicle; performing a terrain fitting based on the depth map to determine a terrain equation; determining a set of candidate obstacle points based on the terrain equation and the depth map; clustering candidate obstacle points in the set of candidate obstacle points to obtain at least one independent obstacle; and identifying and filtering a falsely detected obstacle from the at least one independent obstacle to obtain a depth-map-based obstacle detection result.

Method and apparatus for detecting obstacle, electronic device, vehicle and storage medium

Embodiments of the present disclosure disclose a method for position detection, a device, and a storage medium. The method includes: detecting a first lane line in a current image captured by a camera; performing an optimization on an initial transformation matrix reflecting a mapping relationship between a world coordinate system and a camera coordinate system based on a detection result of the first lane line; and obtaining a first 3D coordinate of a target object in the current image according to a transformation matrix optimized, and determining an ultimate 3D coordinate of the target object according to the first 3D coordinate.

Method for position detection, device, and storage medium

A method and device of multi-focal sensing of an obstacle. A method includes acquiring detection results of obstacles at multiple moments by utilizing a camera with long focus lens and a camera with short focus lens; performing a target tracking to the acquired detection results, to obtain at least two tracking sequences, wherein each tracking sequence includes detection results acquired at the multiple moments for a same obstacle; and matching two random tracking sequences of the at least two tracking sequences, and combining the two random tracking sequences into a combined tracking sequence, if the matching is successful.

Method and device of multi-focal sensing of an obstacle and non-volatile computer-readable storage medium

A method, device, apparatus and a computer-readable storage medium for detecting a height of an obstacle are provided. The method can include: acquiring observation data of a plurality of reference obstacles from a frame; for each reference obstacle, determining a first evaluation value of the height of the obstacle to be detected, according to the observation data of the reference obstacle, an observed height of the obstacle to be detected in a frame, an actual height of the reference obstacle, and a distance between a bottom of the obstacle to be detected in the frame and an optical axis of a camera capturing the frame; and determining a second evaluation value of the height of the obstacle to be detected by using a plurality of first evaluation values for the plurality of reference obstacles.

Embodiments of a method and system for sensing an obstacle, a computer device, and a computer-readable storage medium are provided. The method can include: capturing continuously, by first and second cameras adjacently arranged on a motor vehicle, obstacles around the motor vehicle, to obtain at least a first obstacle image and a second obstacle image respectively; associating the first obstacle image with the second obstacle image; determining whether the first obstacle image and the second obstacle image comprise the same obstacle. By means of some of the disclosed methods and systems, an obstacle can be real-time sensed by a motor vehicle in a large field of view, such as, a range of 360&#xb0; around the vehicle.

Method and system for sensing an obstacle, and storage medium

Embodiments of the present disclosure provide a method and an apparatus for generating three-dimensional data, a device, and a storage medium. The method comprises: obtaining a profile information annotation result for a target object in a two-dimensional image; calculating depth of field of the target object and a three-dimensional orientation of the target object according to the profile information annotation result and a standard three-dimensional size of the target object; determining a three-dimensional coordinate of a reference two-dimensional image point in the target object according to a two-dimensional coordinate of the reference two-dimensional image point in the two-dimensional image and the depth of field; and generating three-dimensional data matching the target object according to the three-dimensional coordinate of the reference two-dimensional image point and the three-dimensional orientation of the target object.

Method and apparatus for generating three-dimensional data, device, and storage medium

The present disclosure provides a method and apparatus of obtaining obstacle information, a device and a computer storage medium, wherein the method of obtaining obstacle information comprises: obtaining scenario data synchronously collected by a laser radar device and a camera device; using point cloud top view data in the scenario data to perform obstacle recognition to obtain a point cloud set of candidate obstacles; using fusion features of the scenario data corresponding to the point cloud set of the candidate obstacles to perform obstacle recognition to obtain specific obstacle information. The technical solution according to the present disclosure enables accurate obtainment of specific obstacle information, thereby enhances the sensing capability of the self-driving vehicle and improves safety of the self-driving vehicle.

Method and apparatus of obtaining obstacle information, device and computer storage medium

The present disclosure provides a three-dimensional posture estimating method and apparatus, a device and a computer storage medium, wherein the method comprises: obtaining two-dimensional posture information of an object in an image and three-dimensional size information of the object; determining coordinates of key points of the object in an object coordinate system according to the three-dimensional size information of the object; determining a transformation relationship between a camera coordinate system and the object coordinate system according to a geometrical relationship between coordinates of key points of the object in the object coordinate system and the two-dimensional posture information of the object. Application of this manner to the field of autonomous driving may implement mapping a detection result of a two-dimensional obstacle to a three-dimensional space to obtain its posture.

Three-dimensional posture estimating method and apparatus, device and computer storage medium

An apparatus for detecting a body part from a user image may include an image acquirer to acquire a depth image, an extractor to extract the user image from a foreground of the acquired depth image, and a body part detector to detect the body part from the user image, using a classifier trained based on at least one of a single-user image sample and a multi-user image sample. The single-user image may be an image representing non-overlapping users, and the multi-user image may be an image representing overlapping users.

Apparatus and method for detecting body parts from user image

An apparatus and method for parsing a human body image may be implemented by acquiring a depth image including a human body, and detecting a plurality of points in the acquired depth image by conducting a minimum energy skeleton scan on the depth image.

Apparatus and method for parsing human body image

A device and a method for image processing include an image processing device that may extract a foreground moving object from a depth map of a three-dimensional (3D) image that may include an image depth map acquirer to obtain the depth map of a successive 3D image over a period of time, a moving object segmenter to segment a moving object from the obtained depth map, and a moving object tracker to identify and track the segmented moving object.

Growing community of inventors

Beijing, China

Xun Sun