Abstract: A computer-implemented method for determining field boundaries and crop forecasts in each field is provided. The method includes deriving vegetation indices for each geo-spatial pixel of each image of multi-spectral imagery at a plurality of points in time, constructing minimum bounding boxes for each image according to the vegetation indices, and generating, based on a neural network analysis of each image and the minimum bounding boxes, a geo-spatial plot of crops including a predicted plot of future crop usage for an area including each field in the multi-spectral imagery.

Abstract: A method for clustering time stamps in time series data includes receiving a one-dimensional array of ordered timestamps and an expected frequency; determining a set of time intervals; determining a first binary array that indicates whether each time interval in the set of time intervals is greater than or less than the expected frequency; determining a second binary array of differences between a corresponding pair of adjacent elements of the first binary array; appending an ith timestamp to one of a set of opening interval bounds, a set of closing interval bounds, or a set of isolated points; merging the set of set of opening interval bounds and the set of closing interval bounds into a set of cluster intervals ?; and outputting the set of cluster intervals and the set of isolated points.

Abstract: A computer-implemented method for determining farm boundary delineations within a target geographic area, comprising extracting data from pixels of a satellite image of the target geographic area, evaluating the data using a classification algorithm to generate one or more line segments between adjacent pixels, the one or more line segments being representative of a portion of a boundary delineation, connecting the one or more line segments to an adjacent line segment to form a boundary delineation defining at least one parcel of land within the target geographic area, and generating a boundary delineation map including the boundary delineation.

Abstract: Scalable feature classification for 3D point cloud data is provided. In one aspect, a method for rasterizing 3D point cloud data includes: obtaining the 3D point cloud data; generating a digital elevation model (DEM) from the 3D point cloud data; decomposing the DEM into local and global fluctuations to obtain a local DEM; generating geo-referenced shapes by automatically thresholding the local DEM; cropping and normalizing the local DEM using minimum bounding boxes derived from the geo-referenced shapes and manual annotations from subject matter experts to create a cropped DEM; and linking geo-spatially tagged labels from the subject matter experts to the cropped DEM. These data can be then directly fed into a system having an ensemble of artificial neural networks. By way of example, a scalable ecosystem is presented on the basis of the geo-spatial platform IBM PAIRS.

Abstract: A computer-implemented method includes accessing, by a processing unit, an existing layer representing geospatial-temporal data at a selected timestamp. A first overview layer of the existing layer is generated by iteratively aggregating each cluster of cells of the existing layer into a corresponding lower-resolution cell of the first overview layer. The first overview layer therefore has a lower resolution than the existing layer. A query is received related to the geospatial-temporal data in the existing layer, and the query is processed with reference to the first overview layer.

Abstract: A system and method to enhance observation resolution using continuous learning include obtaining a first image of a surface area from a first satellite, and obtaining a second image of the surface area from a second satellite. The first image has a lower spatial resolution than the second image, and temporal resolution of the first images obtained by the first satellite is higher than temporal resolution of the second images obtained by the second satellite. The method also includes determining a convolution matrix A or training a neural network, obtaining additional one or more of the first images prior to obtaining an additional one of the second images, and generating a new image from each of the one or more of the first images using the convolution matrix A or the neural network. The new image has a higher spatial resolution than the one or more of the first images.

Abstract: A computer-implemented method is provided for improving farm management by determining farm boundary delineations within a target geographic area based on crop recognition. The method includes evaluating the data using a classification algorithm to generate one or more line segments at an interface separating each pixel in at least one pair of adjacent pixels. It also includes connecting the one or more line segments to an adjacent line segment to form a boundary delineation. The method further includes generating a boundary delineation map including the boundary delineation as well as generating a farm management plan including the boundary delineation map and a recommended crop type. The farm management plan can be generated based on at least a type of vegetation present.

Abstract: A computer-implemented method for determining farm boundary delineations within a target geographic area, comprising extracting data from pixels of a satellite image of the target geographic area, evaluating the data using a classification algorithm to generate one or more line segments between adjacent pixels, the one or more line segments being representative of a portion of a boundary delineation, connecting the one or more line segments to an adjacent line segment to form a boundary delineation defining at least one parcel of land within the target geographic area, and generating a boundary delineation map including the boundary delineation.

Abstract: A computer-implemented method includes receiving one or more training images depicting one or more training geographical regions. One or more environmental characteristic (EC) values are determined for the training images. The EC values include at least one EC value for each of the training images. One or more models are generated for mapping an EC value of an image to a determination of whether an artifact is present in a geographical region depicted by the image, based on the EC values and based on knowledge of which of the training images depict training geographical regions having artifacts present. A new image is received depicting a new geographical region. The models are applied to the new image. A probability that a new artifact is present in the new geographical region depicted in the new image is determined, based on the applying the models to the new image.

Abstract: A system and method to enhance observation resolution using continuous learning include obtaining a first image of a surface area from a first satellite, and obtaining a second image of the surface area from a second satellite. The first image has a lower spatial resolution than the second image, and temporal resolution of the first images obtained by the first satellite is higher than temporal resolution of the second images obtained by the second satellite. The method also includes determining a convolution matrix A or training a neural network, obtaining additional one or more of the first images prior to obtaining an additional one of the second images, and generating a new image from each of the one or more of the first images using the convolution matrix A or the neural network. The new image has a higher spatial resolution than the one or more of the first images.

Abstract: A method for clustering time stamps in time series data includes receiving a one-dimensional array of ordered timestamps and an expected frequency; determining a set of time intervals; determining a first binary array that indicates whether each time interval in the set of time intervals is greater than or less than the expected frequency; determining a second binary array of differences between a corresponding pair of adjacent elements of the first binary array; appending an ith timestamp to one of a set of opening interval bounds, a set of closing interval bounds, or a set of isolated points; merging the set of set of opening interval bounds and the set of closing interval bounds into a set of cluster intervals ?; and outputting the set of cluster intervals and the set of isolated points.

Abstract: A database system includes a processing sub-system having an electronic hardware controller that receives first formatted data, and an electronic file formatting sub-system having an electronic hardware controller that assigns coordinate values corresponding to a second data format to the first formatted data. The file formatting sub-system generates a dual-format data file that fuses together the first formatted data with the coordinate values corresponding to a second data format. The database system further includes a storage sub-system having a data storage unit configured to store the dual-format data file.

Abstract: A database system includes a processing sub-system having an electronic hardware controller that receives first formatted data, and an electronic file formatting sub-system having an electronic hardware controller that assigns coordinate values corresponding to a second data format to the first formatted data. The file formatting sub-system generates a dual-format data file that fuses together the first formatted data with the coordinate values corresponding to a second data format. The database system further includes a storage sub-system having a data storage unit configured to store the dual-format data file.

Abstract: A computer-implemented method includes receiving one or more training images depicting one or more training geographical regions. One or more environmental characteristic (EC) values are determined for the training images. The EC values include at least one EC value for each of the training images. One or more models are generated for mapping an EC value of an image to a determination of whether an artifact is present in a geographical region depicted by the image, based on the EC values and based on knowledge of which of the training images depict training geographical regions having artifacts present. A new image is received depicting a new geographical region. The models are applied to the new image. A probability that a new artifact is present in the new geographical region depicted in the new image is determined, based on the applying the models to the new image.

Abstract: A computer-implemented method for determining farm boundary delineations within a target geographic area, comprising extracting data from pixels of a satellite image of the target geographic area, evaluating the data using a classification algorithm to generate one or more line segments between adjacent pixels, the one or more line segments being representative of a portion of a boundary delineation, connecting the one or more line segments to an adjacent line segment to form a boundary delineation defining at least one parcel of land within the target geographic area, and generating a boundary delineation map including the boundary delineation.

Abstract: A computer-implemented method for determining farm boundary delineations within a target geographic area, comprising extracting data from pixels of a satellite image of the target geographic area, evaluating the data using a classification algorithm to generate one or more line segments between adjacent pixels, the one or more line segments being representative of a portion of a boundary delineation, connecting the one or more line segments to an adjacent line segment to form a boundary delineation defining at least one parcel of land within the target geographic area, and generating a boundary delineation map including the boundary delineation.

Abstract: Methods and systems for detecting emission sites include identifying a set of known emitters having visible features and a spectroscopic signature that correspond to sites that emit a substance to form a training set. A classifier is generated based on the training set. New emitters are identified based on the classifier, a spectroscopic signature map, and a map of visible features. An alert is provided responsive to the identification of a new emitter.