A method of creating a orthomosaic of a corridor area, the corridor area at least partially described by a corridor path, the method comprising flying an aircraft along a primary flight line approximating the corridor path and capturing a sequence of primary images; flying the aircraft along a secondary flight line substantially parallel to the corridor path and capturing a sequence of secondary images; identifying, in the primary images and secondary images, common features corresponding to common ground points; estimating, via bundle adjustment and from the common ground points, an exterior orientation associated with each primary image and a three-dimensional position associated with each ground point; orthorectifying, using at least some of the exterior orientations and at least some of the three-dimensional ground point positions, at least some of the primary images; and merging the orthorectified primary images to create the orthomosaic.

The present invention discloses an artificial intelligence (AI)-based urban design multi-plan generation method for a regulatory plot. The method includes operation steps performed by the following modules: (1) plot space calculation sand table module; (2) regulatory characteristic parameter input module; (3) characteristic case intelligent learning module; (4) plot road intelligent generation module; (5) plot landscape intelligent generation module; (6) plot architecture intelligent generation module; and (7) outputted plan auxiliary drawing module. By means of the AI-based urban design multi-plan generation method for a regulatory plot of the present invention, the planners realize real-time, accurate, and efficient feedback adjustment to a plurality of urban design plans under regulatory conditions, resolving a plurality of problems such as a long working period of a conventional design plan.

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.

A system and a method are provided that are capable of providing map data for supporting a variety of user network environments and selecting data zones freely. navigation terminal includes a reception unit adapted to receive a file in which map data of a specific zone is stored, from a map provision server; and an execution unit adapted to execute a navigation function on the specific zone using the file. The file is produced by an individual unit with respect to each of geographic areas divided by a mesh unit having a variable size. The size of the mesh unit is decided according to the amount of information included in the geographic area such that the file has an equalized size.

An approach is provided for predicting a pose error for a sensor system based on a trained machine learning model. The approach, for example, involves receiving images depicting a survey point with a known physical location. The approach also involves determining meta-data associated with the sensor system used to capture the images. The approach further involves generating a ray from the capture location through a pixel location of the survey point on an image plane of each image. The approach further involves calculating an error between the ray generated for the image and the known physical location. The approach further involves training a machine learning model to predict a pose error from image data captured using the sensor system based on the error in combination with features extracted from the image and the meta-data for the image. The approach further involves providing the trained machined learning as an output.

A method of documenting a position of an underground utility which involves placing temporary markings on a ground surface designating an underground location of the underground utility. A photograph is then taken providing an overhead panoramic view of the temporary markings designating the underground location of the underground utility. The photograph includes visual landmarks, such as sidewalks, driveways, trees and hedges, thereby providing a visual reference. After excavation has taken place and the temporary markings have been removed along with the ground surface, using the photograph to reconfirm the underground location of the underground utility by using the visual reference provided by the relationship that the temporary markings depicted in the photograph bear to the visual landmarks that remain after excavation.

A transportation management system analyzes data characterizing trips associated with an access point where a rider is picked up by, or dropped off by, a driver. The transportation management system determines trip metrics for the access point based on the data. The transportation management system identifies an access point defect for the access point using the trip metrics. The transportation management system flags the access point as associated with a missing road segment responsive to identifying the access point defect. The transportation management system performs a resolution action to address the access point defect.

Efficient 3D geospatial mapping is disclosed. A 3D geospatial map of an area of interest is generated from 2D satellite imagery. The 2D imagery is preprocessed to generate a point cloud of the area of interest. The point cloud is optimized by removing atmospheric clouds and shadows. A 3D geographical information system (GIS) map with multiple levels of details (LOD) is generated.

Techniques are disclosed for mapping in a movable object environment. A method of mapping may include obtaining mapping data from a scanning sensor of a compact payload coupled to an unmanned aerial vehicle (UAV) the compact payload comprising the scanning sensor, one or more cameras, and an inertial navigation system (INS) configured to be synchronized using a reference clock signal, obtaining feature data from a first camera of the one or more cameras, obtaining positioning data from the INS, associating the mapping data with the positioning data based at least on the reference clock signal to generate geo-referenced data, and storing the geo-referenced data and the feature data to a removable storage medium.

Systems and methods are described for identifying and validating the routes and characteristics of roads unknown to a road mapping database. The systems and methods may combine feature recognition analysis of aerial images with other information sources such as location tracking information from a mobile device or client in order to improve the accuracy of road information stored within a road mapping database. The systems and methods may also facilitate the collection of additional information regarding the characteristics of the identified roads from a client device or user thereof.