Abstract: A client device receives a map tile associated with a map feature having a geometric base and a height value, the geometric base represented by a set of vertices, and map terrain data including elevation values. The client device determines whether all vertices of the map feature are in the set of vertices and selects a sample point on the map tile using the set of vertices, determine a sample elevation value for the sample point using the map terrain data, determine vertex elevation values for respective vertices in the set of vertices using the sample elevation value, and render the map tile with the map feature using the vertex elevation values. The map feature is rendered with a top surface graphical component having an elevation corresponding to a sum of the sample elevation value and the height value. The client device displays the rendering.

Abstract: Techniques are described for building extrusion in an electronic map. A client map application receives geographic location information, such as from a user input. The client map application retrieves electronic map data representing a specific building at a location indicated by the received geographic location information. The client map application generates an extruded building graphical element illustrating the specific building in three dimensions based on the retrieved electronic map data representing the specific building. The client map application renders an electronic map representing electronic map data according to a first style. The electronic map also includes the extruded building graphical element such that the specific building is rendered in three dimensions according to a second style.

Abstract: A method for correcting speed estimates for route planning using a machine-learned speed correction model trained on aggregated road data. Location and movement data collected from a plurality of mobile computing devices is aggregated on a server computer and used to train a speed correction model to correct estimated speeds corresponding to roads in one or more geographic regions. Speeds estimates for a road segment in a geographic region are corrected using a speed correction model trained on road data describing road segments in the same geographic region. In some embodiments, road data corresponding to one or more geographic regions is assigned to groups in training the speed correction model. The road data may be anonymized or segmented such that an originating device or route is unidentifiable. More fine-grained speed correction models may also be trained for different or additional factors than geographic region, such as day and/or time.

Abstract: A first set of traces is received by a device, each trace including a set of location data. Each set of location data includes an origin point, a plurality of intermediate points, and a destination point in an ordered sequence. A bounding box surrounding a trace of the first set of traces is drawn, the bounding box corresponding to endpoints of the trace. A function determines a threshold and classifies a trace as stationary or non-stationary. A trace is classified as stationary or non-stationary by comparing a ratio represented by the trace length divided by the bounding box diagonal length to the determined threshold. In response to classifying the trace as stationary, the trace is labeled as stationary. In response to classifying the trace as non-stationary, extremities are extracted from the trace for re-classifying.

Abstract: Digital data corresponding to a trace by a device is received. The digital data includes location data and time data obtained over a period of time and representing movement of the device. A set of derived values for the plurality of metrics is calculated from the sets of values corresponding to the metrics, and the set of derived values is input to a machine-learned classification model to determine whether the received digital data represents movement of a first type. In response to determining that the digital data represents movement of the first type, the received digital data corresponding to the trace is labelled with the first type of movement.

Abstract: A mobile computing device receives an image from a camera physically located within a vehicle. The mobile computing device inputs the image into a convolutional model that generates a set of object detections and a set of segmented environment blocks in the image. The convolutional model includes subsets of encoding and decoding layers, as well as parameters associated with the layers. The convolutional model relates the image and parameters to the sets of object detections and segmented environment blocks. A server that stores object detections and segmented environment blocks is updated with the sets of object detections and segmented environment blocks detected in the image.

Abstract: A system determines spatial locations of pixels of an image. The system includes a processor configured to: receive location data from devices located within a hotspot; generate a density map for the hotspot including density pixels associated with spatial locations defined by the location data, each density pixel having a value indicating an amount of location data received from an associated spatial location; match the density pixels of the density map to at least a portion of the pixels of the image; and determine spatial locations of the at least a portion of the pixels of the image based on the spatial locations of the matching density pixels of the density map. In some embodiments, the image and density map are converted to edge maps, and a convolution is applied to the edge maps to match the density map to the pixels of the image.

Abstract: In order to determine navigation instructions, a mapping server can use digital map data, which can include a set of intersections connected by road segments. In some cases, the mapping server generates navigation guidance data for one or more intersections of digital map data. Responsive to a request for navigation instructions from a client device, the mapping server can generate a navigation route between an origin and destination location, and a corresponding set of navigation instructions can be generated based on the navigation route. In some implementations, the generated navigation instructions are then modified to clarify or otherwise improve the quality of the navigation instructions. Finally, the finalized navigation instructions can be sent to a client device for display to the requesting user.

Abstract: A process, and corresponding system for performing the process, is described for editing and representing property values for a digital map. The process retrieves electronic map data from a style sheet corresponding to a vector map tile set for a digital map and displays map layer styles in a property editor panel, where each map style layer references a portion of the data and one or more visual style rules to apply to the portion. The process also displays a visual map with the map layer styles applied. Upon receiving selection of one or more map style layers from the plurality of map style layers, the process displays style layer properties and values corresponding to each. The process renders in the visual map only the selected one or more style layers marked as visible. As edits are made in the property editor panel, the visual map is updated to reflect the changes.

Abstract: An augmented reality (AR) module receives a location notification including a current location from a client device. The AR module determines an AR tabletop view based on the current location. The AR tabletop view includes geographic information representing a geographic area that includes the current location. The AR module sends the AR tabletop view to the client device. The AR module receives a destination notification including a destination location within the represented geographic area. The AR module transitions the AR tabletop view into an AR world view based on the current location and the destination location. The AR module sends the AR world view to the client device for display.

Abstract: A mapping system prevents mismatching placement of map labels between adjacent map tiles for rendering digital maps. In particular, the mapping system performs a label placement process that considers labels of a map tile shared with adjacent map tiles in order to prevent labels that intersect the map tile and the adjacent map tiles having mismatching placements. During the label placement process, the mapping system performs a first placement phase to place shared labels of a map tile according to label placement constraints. Based on the first phase, the mapping system identifies intersecting labels that can be successfully placed. Using the successfully placed intersecting labels, the mapping system performs a second label placement phase. In particular, the mapping system first places the successfully placed set of intersecting labels, and subsequently places non-intersecting labels for the map tile according to the label placement constraints.

Abstract: A feature extraction system extracts map features from an aerial image. The feature extraction system receives an aerial image having pixels and predicts, for each pixel, a probability that the pixel corresponds to a map feature based on a machine learning model. The machine learning model is trained to determine a probability that a pixel corresponds to the map feature based on a training dataset comprising pairs of aerial images and corresponding mask images that describe known instances of the map feature. The feature extraction system identifies a subset of pixels of the plurality of pixels. Each pixel in the subset has a predicted probability that is greater than or equal to a threshold probability that a pixel corresponds to the map feature. The feature extraction system further determines a bounded geometry enclosing the identified subset of pixels, the bounding geometry encompassing an instance of the map feature.

Abstract: A process, and corresponding system for performing the process, is described for editing and representing property values for a digital map. The process retrieves electronic map data from a style sheet corresponding to a vector map tile set for a digital map. The process displays a plurality of templates and receives selection of a template. The process displays a visual map representing the electronic map with the set of map style layers of the selected template. The process displays a theme and a plurality of components. Each component is a set of map style layers and is a subset of the set of map style layers of the template. The process receives selection of a component. The process displays a plurality of property values associated with the selected component. The process receives an edit to a property value. The process updates the visual map to reflect the edited property value.

Abstract: AR elements are occluded in video image frames. A depth map is determined for an image frame of a video received from a video capture device. An AR graphical element for overlaying over the image frame is received. An element distance for AR graphical elements relative to a position of a user of the video capture device (e.g., the geographic position of the video capture device) is also received. Based on the depth map for the image frame, a pixel distance is determined for each pixel in the image frame. The pixel distances of the pixels in the image frame are compared to the element distance, and in response to a pixel distance for a given pixel being less than the element distance, the pixel of the image frame is displayed rather than a corresponding pixel of the AR graphical element.

Abstract: A computer-implemented method includes receiving a video comprising image frames depicting multiple objects. The video is captured by a video capture device moving relative to the surface of the Earth while the video is captured. A geographic location of the video capture device is received for each of the image frames and an angular orientation of the video capture device is determined based on the image frames. The determining the angular orientation includes determining a line in the image frames of the video for each object of a plurality of the multiple objects. The determined line corresponds to two-dimensional positions of the object in the image frames. The computer-implemented method includes determining a vanishing point of the image frames based on the determined lines and determining the angular orientation of the video capture device based on the determined vanishing point.

Abstract: An oblique imagery application receives an oblique image captured by an oblique camera at a non-orthogonal angle with respect to a ground plane and map data including a map tile corresponding to geographic coordinates. A principal axis is determined that is orthogonal to an image plane defined by the oblique image and intersecting a center of the oblique image. For each pixel of the oblique image, a pixel vector is determined and a set of deviation coordinates based on a deviation of the pixel vector from the principal axis is determined for the pixel, with the pixel vector of a pixel passing through a focal point of the oblique camera and ending at the pixel. The map tile is associated to the pixels of the oblique image based on the camera parameters, the deviation coordinates of the pixels, the oblique camera parameters, and the geographic coordinates of the map tile.

Abstract: Techniques are described for building extrusion in an electronic map. A client map application receives geographic location information, such as from a user input. The client map application retrieves electronic map data representing a specific building at a location indicated by the received geographic location information. The client map application generates an extruded building graphical element illustrating the specific building in three dimensions based on the retrieved electronic map data representing the specific building. The client map application renders an electronic map representing electronic map data according to a first style. The electronic map also includes the extruded building graphical element such that the specific building is rendered in three dimensions according to a second style.

Abstract: A first set of traces is received by a device, each trace including a set of location data. Each set of location data includes an origin point, a plurality of intermediate points, and a destination point in an ordered sequence. A bounding box surrounding a trace of the first set of traces is drawn, the bounding box corresponding to endpoints of the trace. A function determines a threshold and classifies a trace as stationary or non-stationary. A trace is classified as stationary or non-stationary by comparing a ratio represented by the trace length divided by the bounding box diagonal length to the determined threshold. In response to classifying the trace as stationary, the trace is labeled as stationary. In response to classifying the trace as non-stationary, extremities are extracted from the trace for re-classifying.

Abstract: A process, and corresponding system for performing the process, is described for editing and representing property values for a digital map. The process stores a library of expressions for styling a digital map. The process retrieves electronic map data and displays a user interface including a property editor panel and a visual map portion. The property editor panel displays a plurality of map styles. The visual map portion displays a representation of the electronic map data. For a selected map style, the process displays a plurality of style properties and a text field for receiving expressions corresponding to a style property. The process receives, via the text field, an expression and determines a set of suggested expressions corresponding to the style property. The process applies the received expression to the electronic map data and renders in the visual map portion an edited representation of the electronic map data.

Abstract: A polygon Boolean operation and topology correction algorithm generates valid polygons despite polygon data containing self-intersecting polygons, overlapping polygons, polygons including a chain of holes, and/or polygons including complex intersections. A polygon Boolean operation and topology correction algorithm uses two variants of the Vatti algorithm along with a topology correction algorithm to address problematic areas within input data, such as the overlapping area shared between one or more polygons or a polygon having no area at all. Upon performing a variant of the Vatti algorithm to identify hot pixels within an arrangement of polygons, the polygon Boolean operation and topology correction algorithm runs a second variant of the Vatti algorithm to generate polygons within an integer coordinate system using snap-rounding.