Abstract: The present disclosure relates to a computer implemented method (100) for attribution of Land Use/Land Cover, LULC, said method comprising the steps of: providing (S1) a plurality of images covering a transport network part, identifying (S2) objects in at least a subset of said plurality of images, identifying (S4) transport routes at least in said at least one satellite image of the plurality of images, and determining (S5) characteristics of said transport network part based on the identified objects at said identified (S4) transport routes.

Abstract: The disclosure relates to a computer implemented method for determining 3D coordinates describing a scene, obtaining first, second and third images covering the scene captured from different viewpoints, said images being associated to a timing of capture, wherein the scene is formed in an overlapping geographical area of the first, second and third images.

Abstract: The present disclosure relates to a method for 3D reconstruction from satellite imagery using deep learning, said method comprising providing (101) at least two overlapping 2D satellite images, providing (102) imaging device parameters for the at least two overlapping 2D satellite images, providing (103) at least one trained Machine Learning Network, MLN, able to predict depth maps, said trained MLN being trained on a training set comprising multi-view geocoded 3D ground truth data and predicting (104) a depth map of the at provided at least two 2D satellite images using the trained at least one MLN and based on the corresponding imaging device parameters.

Abstract: Described are a system (200) and method arranged to provide improved geocoded reference data to a 3D map representation. The system comprises a storage (201) having stored thereupon a 3D map representation comprising a textured 3D representation provided with geocoded reference data and formed based on imagery, the imagery being associated to information relating to at least one imaging device which has captured the imaging. The system comprises further a processor (208) arranged to receive at least one new image associated to information related to an imaging device which has captured the new image, perform registration of the new image to the 3D map representation, determine corresponding points in the new image and the 3D map representation, and determine displacement data for a plurality of 3D positions in the 3D map representation based on the determined corresponding points in the new image and the 3D map representation.

Abstract: The present disclosure relates to a method for 3D reconstruction from satellite imagery using deep learning, said method comprising providing (101) at least two overlapping 2D satellite images, providing (102) imaging device parameters for the at least two overlapping 2D satellite images, providing (103) at least one trained Machine Learning Network, MLN, able to predict depth maps, said trained MLN being trained on a training set comprising multi-view geocoded 3D ground truth data and predicting (104) a depth map of the at provided at least two 2D satellite images using the trained at least one MLN and based on the corresponding imaging device parameters.

Abstract: The present invention relates to a method and an arrangement for providing a 3D model of an environment. The method comprises the step of forming a mesh modelling the environment in three dimensions, said mesh comprising nodes interconnected by edges and having surfaces boarded by the edges, wherein each node is associated to a 3D coordinate in a geographical coordinate system, determining for a plurality of the nodes and/or surfaces and/or edges in the mesh a mesh uncertainty and associating the determined mesh uncertainty to the corresponding node and/or surface and/or edge.

Abstract: A method for resolution enhancement of images is described comprising the steps of providing (101) at least a first two dimensional (2D) test image, providing (102) a high-resolution 3D map, providing (103) a Machine Learning Network (MLN), extracting (104), from the high-resolution 3D map, a 2D submap, comprising geocoded 2D coordinate data and texture information, extracting (105) a 2D subimage from the 2D test image, which 2D subimage is an image of the same area as the 2D submap, and training the MLN, using the high-resolution 2D submap and the 2D subimage.

Abstract: The present disclosure relates to a method and system for navigation of an aerial vehicle. The method (100) comprises providing (110) a sensor image from an aerial vehicle sensor and repeatedly, until at least one predetermined criterion is reached, performing the steps: setting (120) input data comprising information related to pitch angle, roll angle, yaw angle and three-dimensional position of the aerial vehicle; providing (130) at least one two-dimensional perspective view image based on the input data, where the at least one two-dimensional perspective view image is obtained from a database comprising three-dimensional geo-referenced information of the environment, said three-dimensional geo-referenced information comprising texture data; and comparing (140) the sensor image and the at least one two-dimensional perspective view image.