Mobileye Vision Technologies Ltd.

Nadav Shaag

Gideon Stein

Itay Blumenthal

Jeffrey Moskowitz

Ofer Springer

Natalie Carlebach

Yaakov Kingsley

David Neuhof

Gal Topel

Yotam Stern

Erez Dagan

Kevin Rosenblum

Roy Maor Lotan

Shahar Harouche

Alon Ziv

Ishay Levi

Abraham Hendler

Yahel Guberman

David Boublil

Tomer Hochbaum

Nir Zabari

Chaim Blau

Gal Alkon

Yonatan Shemesh

Paz Lavi Bunis

Noa Schaffer

Daniel Einy

Mordechai Ben Zecharia

Ido Kenan

Jeremy Atlan

Techniques are disclosed for improving upon the usage of Light Detection and Ranging (LIDAR) supervision to perform image depth estimation. The techniques use a generator and adversary network to generate respective models that 'compete' against one another to enable the generator model to output a desired output image that compensates for a LIDAR image having a structured or lined data pattern. The techniques described herein may be suitable for use by vehicles and/or other agents operating in a particular environment as part of machine vision algorithms that are implemented to perform autonomous and/or semi-autonomous functions.

Adversarial approach to usage of lidar supervision to image depth estimation

System and techniques for vehicle environment modeling with a camera are described herein. A device for modeling an environment comprises: a hardware sensor interface to obtain a sequence of unrectified images representative of a road environment, the sequence of unrectified images including a first unrectified image, a previous unrectified image, and a previous-previous unrectified image; and processing circuitry to: provide the first unrectified image, the previous unrectified image, and the previous-previous unrectified image to an artificial neural network (ANN) to produce a three-dimensional structure of a scene; determine a selected homography; and apply the selected homography to the three-dimensional structure of the scene to create a model of the road environment.

Vehicle environment modeling with a camera

The present disclosure is directed to searching an image database. A system may include at least one processor comprising circuitry and a memory, wherein the memory includes instructions that, when executed by the circuitry cause the at least one processor to: receive a first user input including one or more scenario descriptors, identify in a database an initial plurality of images based on the first user input and display the initial plurality of images to a user. The processor may be further configured to receive a second user input wherein the second user input identifies at least one of the initial plurality of images. The at least one processor may identify a refined plurality of images in the database based on the one or more scenario descriptors in combination with the second user input and display the refined plurality of images to the user.

Systems and methods for searching an image database

A navigation system for a host vehicle may include a processor programmed to: receive from a camera onboard the host vehicle at least one captured image representative of an environment of the host vehicle, wherein the camera is positioned at a first location relative to the host vehicle; receive point cloud information from a LIDAR system onboard the host vehicle, wherein the LIDAR system is positioned at a second location relative to the host vehicle; analyze the at least one captured image and the received point cloud information to detect one or more objects in the shared field of view region; determine whether a vantage point difference between the first location of the camera and the second location of the LIDAR system accounts for the one or more detected objects being represented in only one of the at least one captured image or the received point cloud information.

LIDAR-camera fusion where LIDAR and camera validly see different things

Systems and methods for navigating a host vehicle are disclosed. In one implementation, a system includes a processor configured to receive a first image acquired by a first camera and a second image acquired by a second camera onboard the host vehicle; identify a first representation of an object in the first image and a second representation of the object in the second image; input to a first trained model at least a portion of the first image; input to a second trained model at least a portion of the second image; receive the first signature encoding determined by the first trained model and the second signature encoding determined by the second trained model; input to a third trained model the first signature encoding and the second signature encoding; and receive an indicator of a location of the object determined by the third trained model.

Stereo-assist network for determining an object's location

System and techniques for vehicle environment modeling with a camera are described herein. A time-ordered sequence of images representative of a road surface may be obtained. An image from this sequence is a current image. A data set may then be provided to an artificial neural network (ANN) to produce a three-dimensional structure of a scene. Here, the data set includes a portion of the sequence of images that includes the current image, motion of the sensor from which the images were obtained, and an epipole. The road surface is then modeled using the three-dimensional structure of the scene.

Various systems and methods for modeling a scene. A device for modeling a scene includes a hardware interface to obtain a time-ordered sequence of images representative of a scene, the time-ordered sequence including a plurality of images, one of the sequence of images being a current image, the scene captured by a monocular imaging system; and processing circuitry to: provide a data set to an artificial neural network (ANN) to produce a three-dimensional structure of the scene, the data set including: a portion of the sequence of images, the portion of the sequence of images including the current image; and motion of a sensor that captured the sequence of images; and model the scene using the three-dimensional structure of the scene, wherein the three-dimensional structure is determined for both moving and fixed objects in the scene.

Vehicle environment modeling with cameras

A navigation system for a host vehicle may include a processor programmed to: receive from a center camera onboard the host vehicle a captured center image including a representation of at least a portion of an environment of the host vehicle, receive from a left surround camera onboard the host vehicle a captured left surround image including a representation of at least a portion of the environment of the host vehicle, and receive from a right surround camera onboard the host vehicle a captured right surround image including a representation of at least a portion of the environment of the host vehicle; provide the center image, the left surround image, and the right surround image to an analysis module configured to generate an output relative to the at least one captured center image; and cause a navigational action by the host vehicle based on the generated output.

Pseudo lidar

Systems and methods are provided for determining a road profile along a predicted path. In one implementation, a system includes at least one image capture device configured to acquire a plurality of images of an area in a vicinity of a user vehicle; a data interface; and at least one processing device configured to receive the plurality of images captured by the image capture device through the data interface; and compute a profile of a road along one or more predicted paths of the user vehicle. At least one of the one or more predicted paths is predicted based on image data.

Road profile along a predicted path

In some embodiments, a first homography, created from two images of a roadway, is decomposed to determine an ego-motion, and the ego-motion is used to adjust a previous estimate of a road plane. The adjusted previous estimate of the road plane is combined with the current estimate of the plane to create a second homography, and the second homography is used to determine residual motion and vertical deviation in the surface of the roadway. In some embodiments, multiple road profiles each corresponding to a common portion of a roadway are adjusted in slope and offset by optimizing a function having a data term, a smoothness term and a regularization term; and the adjusted road profiles are combined into a multi-frame road profile. In some embodiments, road profile information for a predetermined number of data points is transmitted in periodic data bursts, with more than one data point per data burst.

Road vertical contour detection using a stabilized coordinate frame

Growing community of inventors

Jerusalem, Israel

Nadav Shaag