Google Inc.

Alan Sheridan

Daniel Joseph Filip

Aaron Michael Donsbach

Jeremy Bryant Pack

Scott Benjamin Satkin

Vivek Verma

Kyle John Krafka

Marek Gorecki

Andre Le

Juan David Hincapie

Charles Robert Armstrong

Mirko Ranieri

Michael Vitus

David Santana

Wenli Zhao

Bo Zhang

Matt Seegmiller

Tony Shih

Mohamed Suhail Mohamed Yousuf Sait

In one aspect, a request to generate an automated tour based on a set of panoramic images is received. Each particular panoramic image is associated with geographic location information and linking information linking the particular panoramic image with one or more other panoramic images in the set. A starting panoramic image and a second panoramic image are determined based at least in part on the starting panoramic image and the linking information associated with the starting and second panoramic images. A first transition between the starting panoramic image and the second panoramic image is also determined based at least in part on the linking information for these panoramic images. Additional panoramic images as well as a second transition for between the additional panoramic images are also determined. The determined panoramic images and transitions are added to the tour according to an order of the tour.

Automatically generating panorama tours

To present augmented reality features without localizing a user, a client device receives a request for presenting augmented reality features in a camera view of a computing device of the user. Prior to localizing the user, the client device obtains sensor data indicative of a pose of the user, and determines the pose of the user based on the sensor data with a confidence level that exceeds a confidence threshold which indicates a low accuracy state. Then the client device presents one or more augmented reality features in the camera view in accordance with the determined pose of the user while in the low accuracy state.

Geographic augmented reality design for low accuracy scenarios

Systems and methods for adjusting an image are provided. In one example embodiment, a computer-implemented method includes obtaining data representing an image. The method includes inputting the image into a machine learned model. The method includes obtaining, in response to inputting the image into the machine learned model, an output of the machine learned model that includes an estimated tilt of the image. The method includes correcting a tilt of the image based at least in part on the estimated tilt obtained from the machine learned model.

Systems and methods for leveling images

Aspects of the disclosure generally relate to connecting panoramic images. One or more computing devices may load and display a first panoramic image captured at a first location and receive a selection of an area on the first panoramic image, the area corresponding to where a connection to other panoramic images may be made. The one or more computing devices may identify and display one or more nearby panoramic images which were captured near the first location and receive a selection of one of the one or more nearby panoramic images. The one or more computing devices may display the selected panoramic image and the first panoramic image and align the selected panoramic image with the first panoramic image such that the selected panoramic image is oriented in the same direction as the first panoramic image. The one or more computing devices may connect the selected panoramic image with the first panoramic image.

Method and system for obtaining pair-wise epipolar constraints and solving for panorama pose on a mobile device

Aspects of the present disclosure relate to generating turn-by-turn direction previews. In one aspect, one or more computing devices, may receive a request for a turn-by-turn direction preview. The one or more computing devices may generate a set of directions from a first geographic location to a second geographic location. Each turn of the directions may be associated with a corresponding waypoint. The one or more computing devices may then identify street level images corresponding with the waypoints, and display a preview of their associated turns by displaying their street level images and a map in between.

Generating turn-by-turn direction previews

Aspects of the present disclosure relate to generating turn-by-turn direction previews. In one aspect, one or more computing devices, may receive a request for a turn-by-turn direction preview. The one or more computing devices may generate a set of turn-by-turn directions based on a series of road segments connecting a first geographic location and a second geographic location. Each direction in the set of turn-by-turn directions may be associated with a corresponding waypoint. The one or more computing devices then identify a set of images corresponding the series of road segments between two adjacent waypoints of the set of turn-by-turn directions, and determine a subset of the set of images to include in the turn-by-turn direction preview. Subsequently, the one or more computing devices may generate the turn-by-turn direction preview based on at least in part on the determined subset of the set of images.

Aspects of the disclosure generally relate to determining the location and orientation of panoramic images by a computing apparatus. One or more computing devices may receive alignment data between a first panoramic image and second panoramic image and original location data for the first panoramic image and the second panoramic image. The one or more computing devices may determine relative orientations between the pair of panoramic images based on the alignment data and calculate a heading from the first panoramic image to the second panoramic image based on the original location data. The location data and alignment data may be optimized by the one or more computing devices based on the relative orientations between the pair of panoramic images and the original location data. The one or more computing devices may replace the original location data and relative orientations with the optimized relative orientations and optimized location data.

A system and method provide for determining regions of interest within an image based on viewer interaction with the image. At least one image associated with a location is provided for display in a viewport, and pose data related to user interaction with the at least one image is identified. Weights are assigned to portions of the at least one image based on the pose data, the weights indicating at least a period of time the portion of the at least one image is generally at a center of the viewport. Based on the assigned weights, image regions of interest of the at least one image are determined.

Determining regions of interest based on user interaction

Aspects of the disclosure provide for selecting viewports. A viewport is a sub-portion of an image. For example, a request to select a viewport for a particular image may be received. A set of potential viewports are identified within the particular image, and features are identified within the viewports of the set of potential viewports. A score is calculated for each given viewport of the set of potential viewports based at least in part on a weight vector having a set of feature weights and the identified features within that given viewport of the set of potential viewports. A viewport of the set of potential viewports is selected based on the calculated scores. The selected viewport is then provided in response to the request to select the viewport for the particular image.

Growing community of inventors

Apex, NC, United States of America

Alan Sheridan