Abstract: Methods and system are provided for training and using a model to determine states of lanes of interest. For instance, image data including an image and an associated label identifying at least one traffic light, a state of the at least one traffic light, and a lane controlled by the at least one traffic light are received and used to train the model such that the model is configured to, in response to receiving an image and a lane of interest included in the image, output a lane state for the lane of interest. This model is then used by a vehicle in order to determine a state of a lane of interest. This state is then used to control the vehicle in an autonomous driving mode based on the state of the lane of interest.

Abstract: Aspects of the disclosure relate to training and using a phrase recognition model to identify phrases in images. As an example, a selected phrase list may include a plurality of phrases is received. Each phrase of the plurality of phrases includes text. An initial plurality of images may be received. A training image set may be selected from the initial plurality of images by identifying the phrase-containing images that include one or more phrases from the selected phrase list. Each given phrase-containing image of the training image set may be labeled with information identifying the one or more phrases from the selected phrase list included in the given phrase-containing images. The model may be trained based on the training image set such that the model is configured to, in response to receiving an input image, output data indicating whether a phrase of the plurality of phrases is included in the input image.

Abstract: Methods and system are provided for training and using a model to determine states of lanes of interest. For instance, image data including an image and an associated label identifying at least one traffic light, a state of the at least one traffic light, and a lane controlled by the at least one traffic light are received and used to train the model such that the model is configured to, in response to receiving an image and a lane of interest included in the image, output a lane state for the lane of interest. This model is then used by a vehicle in order to determine a state of a lane of interest. This state is then used to control the vehicle in an autonomous driving mode based on the state of the lane of interest.

Abstract: Methods and system are provided for training and using a model to determine states of lanes of interest. For instance, image data including an image and an associated label identifying at least one traffic light, a state of the at least one traffic light, and a lane controlled by the at least one traffic light are received and used to train the model such that the model is configured to, in response to receiving an image and a lane of interest included in the image, output a lane state for the lane of interest. This model is then used by a vehicle in order to determine a state of a lane of interest. This state is then used to control the vehicle in an autonomous driving mode based on the state of the lane of interest.

Abstract: Aspects of the disclosure relate to training and using a phrase recognition model to identify phrases in images. As an example, a selected phrase list may include a plurality of phrases is received. Each phrase of the plurality of phrases includes text. An initial plurality of images may be received. A training image set may be selected from the initial plurality of images by identifying the phrase-containing images that include one or more phrases from the selected phrase list. Each given phrase-containing image of the training image set may be labeled with information identifying the one or more phrases from the selected phrase list included in the given phrase-containing images. The model may be trained based on the training image set such that the model is configured to, in response to receiving an input image, output data indicating whether a phrase of the plurality of phrases is included in the input image.

Abstract: Aspects of the disclosure relate to detecting and responding to stop signs. An object detected in a vehicle's environment having location coordinates may be identified as a stop sign and, it may be determined whether the location coordinates of the identified stop sign correspond to a location of a stop sign in detailed map information. Then, whether the identified stop sign applies to the vehicle may be determined based on the detailed map information or on a number of factors. Then, if the identified stop sign is determined to apply to the vehicle, responses of the vehicle to the stop sign may be determined, and, the vehicle may be controlled based on the determined responses.

Abstract: Aspects of the disclosure relate to training and using a phrase recognition model to identify phrases in images. As an example, a selected phrase list may include a plurality of phrases is received. Each phrase of the plurality of phrases includes text. An initial plurality of images may be received. A training image set may be selected from the initial plurality of images by identifying the phrase-containing images that include one or more phrases from the selected phrase list. Each given phrase-containing image of the training image set may be labeled with information identifying the one or more phrases from the selected phrase list included in the given phrase-containing images. The model may be trained based on the training image set such that the model is configured to, in response to receiving an input image, output data indicating whether a phrase of the plurality of phrases is included in the input image.

Abstract: Aspects of the disclosure relate to training and using a phrase recognition model to identify phrases in images. As an example, a selected phrase list may include a plurality of phrases is received. Each phrase of the plurality of phrases includes text. An initial plurality of images may be received. A training image set may be selected from the initial plurality of images by identifying the phrase-containing images that include one or more phrases from the selected phrase list. Each given phrase-containing image of the training image set may be labeled with information identifying the one or more phrases from the selected phrase list included in the given phrase-containing images. The model may be trained based on the training image set such that the model is configured to, in response to receiving an input image, output data indicating whether a phrase of the plurality of phrases is included in the input image.

Abstract: Aspects of the disclosure relate to training and using a phrase recognition model to identify phrases in images. As an example, a selected phrase list may include a plurality of phrases is received. Each phrase of the plurality of phrases includes text. An initial plurality of images may be received. A training image set may be selected from the initial plurality of images by identifying the phrase-containing images that include one or more phrases from the selected phrase list. Each given phrase-containing image of the training image set may be labeled with information identifying the one or more phrases from the selected phrase list included in the given phrase-containing images. The model may be trained based on the training image set such that the model is configured to, in response to receiving an input image, output data indicating whether a phrase of the plurality of phrases is included in the input image.

Abstract: Aspects of the disclosure relate to training and using a model for determine states of lanes of interest. For instance, image data including an image and an associated label identifying at least one traffic light, a state of the at least one traffic light, and a lane controlled by the at least one traffic light may be received and used to train the mode such that the model is configured to, in response to receiving an image and a lane of interest included in the image, output a lane state for the lane of interest. This model may then be used by a vehicle in order to determine a state of a lane of interest. This state may then be used to control the vehicle in an autonomous driving mode based on the state of the lane of interest.

Abstract: Aspects of the disclosure relate to training and using a phrase recognition model to identify phrases in images. As an example, a selected phrase list may include a plurality of phrases is received. Each phrase of the plurality of phrases includes text. An initial plurality of images may be received. A training image set may be selected from the initial plurality of images by identifying the phrase-containing images that include one or more phrases from the selected phrase list. Each given phrase-containing image of the training image set may be labeled with information identifying the one or more phrases from the selected phrase list included in the given phrase-containing images. The model may be trained based on the training image set such that the model is configured to, in response to receiving an input image, output data indicating whether a phrase of the plurality of phrases is included in the input image.

Abstract: Aspects of the disclosure relate to detecting and responding to stop signs. An object detected in a vehicle's environment having location coordinates may be identified as a stop sign and, it may be determined whether the location coordinates of the identified stop sign correspond to a location of a stop sign in detailed map information. Then, whether the identified stop sign applies to the vehicle may be determined based on the detailed map information or on a number of factors. Then, if the identified stop sign is determined to apply to the vehicle, responses of the vehicle to the stop sign may be determined, and, the vehicle may be controlled based on the determined responses.

Abstract: Systems and methods are provided for manually finding a view for a geographic object in a street level image and associating the view with the geographic object. Information related to a geographic object and a first image related to the geographic object is displayed. User inputs indicating a presence of the geographic object in the image and user input indicating a viewpoint within the image are received and processed. An association of the viewpoint, the image and the geographic object is made and the association is stored in a database. A second image is determined, based on the association, as a default initial image to be displayed for the geographic object in a mapping application.

Abstract: Aspects of the disclosure relate to detecting and responding to stop signs. An object detected in a vehicle's environment having location coordinates may be identified as a stop sign and, it may be determined whether the location coordinates of the identified stop sign correspond to a location of a stop sign in detailed map information. Then, whether the identified stop sign applies to the vehicle may be determined based on the detailed map information or on a number of factors. Then, if the identified stop sign is determined to apply to the vehicle, responses of the vehicle to the stop sign may be determined, and, the vehicle may be controlled based on the determined responses.

Abstract: Aspects of the disclosure relate to detecting and responding to stop signs. An object detected in a vehicle's environment having location coordinates may be identified as a stop sign and, it may be determined whether the location coordinates of the identified stop sign correspond to a location of a stop sign in detailed map information. Then, whether the identified stop sign applies to the vehicle may be determined based on the detailed map information or on a number of factors. Then, if the identified stop sign is determined to apply to the vehicle, responses of the vehicle to the stop sign may be determined, and, the vehicle may be controlled based on the determined responses.

Abstract: The invention relates to collecting different images using different camera parameters. As an example, a single camera may capture two different types of images: a dark exposure for light emitting objects and a normal exposure for passive objects. The camera may first capture an image. This first image may be processed to determine the ideal camera settings for capturing the average intensity of the environment. Fixed offset values may be added to these ideal camera settings and the dark exposure may be captured. The ideal camera settings are then used to capture a normal exposure, which in turn may be processed to determine new ideal camera settings. Again, the fixed offset values may be added to the new ideal camera settings and a dark exposure is captured. This process may repeat continuously and periodically, and the resulting images may be processed to identify emissive and passive objects.

Abstract: Systems and methods for assisting a user in capturing images for three-dimensional (3D) reconstruction are provided. A system includes an input module configured to receive current video image data from a device. The system also includes an analysis module configured to determine a 3D reconstruction indicator based on the current video image data and a previous image (or images) of a target captured by the device. The analysis module is also configured to determine whether the 3D reconstruction indicator is sufficient to allow a 3D reconstruction of the target based on the current video image data and the previous image. The system also includes a guidance module configured to assist the user in capturing a current image of the target from the current video image data based on the determination of whether the 3D reconstruction indicator is sufficient.

Abstract: Systems and methods are provided for manually finding a view for a geographic object in a street level image and associating the view with the geographic object. Information related to a geographic object and a first image related to the geographic object is displayed. User inputs indicating a presence of the geographic object in the image and user input indicating a viewpoint within the image are received and processed. An association of the viewpoint, the image and the geographic object is made and the association is stored in a database. A second image is determined, based on the association, as a default initial image to be displayed for the geographic object in a mapping application.

Abstract: A street-level imagery acquisition and selection process identifies which images are published in a street field view. An imagery database includes panoramas each corresponding to a set of images acquired from a single viewpoint. The panoramas are attached to corresponding positions on a road network graph. The graph is divided into a set of selection paths, each of which includes a topologically linear sequence of road segments. Each selection path is evaluated to select a set of panoramas to be published in the path. Panoramas of interior road segments are selected before panoramas at intersections. Selected panorama identifiers for each interior road segment of the selection paths and each intersection correspond to a position along the road network graph. The selected panorama identifiers are then published in the street field view.

Abstract: Methods and systems for detection of a construction zone sign are described. A computing device, configured to control the vehicle, may be configured to receive, from an image-capture device coupled to the computing device, images of a vicinity of the road on which the vehicle is travelling. Also, the computing device may be configured to determine image portions in the images that may depict sides of the road at a predetermined height range. Further, the computing device may be configured to detect a construction zone sign in the image portions, and determine a type of the construction zone sign. Accordingly, the computing device may be configured to modify a control strategy associated with a driving behavior of the vehicle; and control the vehicle based on the modified control strategy.