Abstract: Implementations are described herein for using one or more transformer networks to generate inferred image data based on processing image data capturing a particular geographic area during a particular time period, including first image data captured in a first spectral band and at a first spatial (and/or temporal) resolution and second image data captured in a second spectral band and at a second spatial (and/or temporal) resolution. The inferred image data can include second spectral information at the first spatial (and/or temporal) resolution, or vice versa. Thus, the spatial and/or temporal resolution of image data of a certain spectral band can be improved, allowing for more effective usage of satellite imagery in agricultural settings.

Abstract: Implementations set forth herein relate to determining causal relationships between covariates and value metrics for geographic regions for training one or more machine learning models. Causal relationships between different subsets of covariates and value metrics can be determined for various durations of time and for various geographic regions. For example, a value metric may exhibit a causal relationship to certain covariates for a first geographic region during a first duration of time, but may exhibit a different causal relationship to other covariates for a second geographic region for a second duration of time. Models can be trained and utilized to predict changes in value metrics for geographic regions, thereby enabling forecasting notifications to be provided to persons who may be negatively impacted by changes to those geographic regions.

Abstract: Implementations are described herein for training and applying machine learning models to digital images capturing plants, and to other data indicative of attributes of individual plants captured in the digital images, to recognize individual plants in distinction from other individual plants. In various implementations, a digital image that captures a first plant of a plurality of plants may be applied, along with additional data indicative of an additional attribute of the first plant observed when the digital image was taken, as input across a machine learning model to generate output. Based on the output, an association may be stored in memory, e.g., of a database, between the digital image that captures the first plant and one or more previously-captured digital images of the first plant.

Abstract: Implementations are described herein for training and applying machine learning models to digital images capturing plants, and to other data indicative of attributes of individual plants captured in the digital images, to recognize individual plants in distinction from other individual plants. In various implementations, a digital image that captures a first plant of a plurality of plants may be applied, along with additional data indicative of an additional attribute of the first plant observed when the digital image was taken, as input across a machine learning model to generate output. Based on the output, an association may be stored in memory, e.g., of a database, between the digital image that captures the first plant and one or more previously-captured digital images of the first plant.

Abstract: Implementations relate to detecting/replacing transient obstructions from high-elevation digital images, and/or to fusing data from high-elevation digital images having different spatial, temporal, and/or spectral resolutions. In various implementations, first and second temporal sequences of high-elevation digital images capturing a geographic area may be obtained. These temporal sequences may have different spatial, temporal, and/or spectral resolutions (or frequencies). A mapping may be generated of the pixels of the high-elevation digital images of the second temporal sequence to respective sub-pixels of the first temporal sequence. A point in time at which a synthetic high-elevation digital image of the geographic area may be selected. The synthetic high-elevation digital image may be generated for the point in time based on the mapping and other data described herein.

Abstract: Implementations are described herein for predicting a future climate condition in an agricultural area. In various implementations, a teleconnection model may be applied to a dataset of remote climate conditions such as water surface temperatures to identify one or more of the most influential remote climate conditions on the future climate condition in the agricultural area. A trained machine learning model may be applied to the one or more most influential remote climate conditions and to historical climate data for the agricultural area to generate data indicative of the predicted future climate condition. Based on the data indicative of the predicted future climate condition, one or more output components may be caused to render output that conveys the predicted future climate condition.

Abstract: Methods, systems, and apparatuses, including computer programs encoded on a computer storage medium, for machine learning classification based on separate processing of multiple views. In some implementations, a system obtains image data for multiple images showing different views of an object. A machine learning model is used to generate a separate output based on each the multiple images individually. The outputs for the respective images are combined to generate a combined output. A predicted characteristic of the object is determined based on the combined output. An indication of the predicted characteristic of the object is provided.

Abstract: Methods, systems, apparatus, and computer-readable media for data band selection using machine learning. In some implementations, image data comprising information for each of multiple wavelength bands is obtained. A multi-layer neural network is trained using the image data to perform one or more classification or regression tasks. A proper subset of the wavelength bands is selected based on parameters of a layer of the trained multi-layer neural network, where the parameters were determined through training of the multi-layer neural network using the image data. Output is provided indicating that the selected wavelength bands are selected for the one or more classification or regression tasks.

Abstract: Implementations relate to detecting/replacing transient obstructions from high-elevation digital images, and/or to fusing data from high-elevation digital images having different spatial, temporal, and/or spectral resolutions. In various implementations, first and second temporal sequences of high-elevation digital images capturing a geographic area may be obtained. These temporal sequences may have different spatial, temporal, and/or spectral resolutions (or frequencies). A mapping may be generated of the pixels of the high-elevation digital images of the second temporal sequence to respective sub-pixels of the first temporal sequence. A point in time at which a synthetic high-elevation digital image of the geographic area may be selected. The synthetic high-elevation digital image may be generated for the point in time based on the mapping and other data described herein.

Abstract: Image data is obtained that indicates an extent to which one or more objects reflect, scatter, or absorb light at each of multiple wavelength bands, where the image data was collected while a conveyor belt was moving the object(s). The image data is preprocessed by performing an analysis across frequencies and/or performing an analysis across a representation of a spatial dimension. A set of feature values is generated using the image preprocessed image data. A machine-learning model generates an output using to the feature values. A prediction of an identity of a chemical in the one or more objects or a level of one or more chemicals in the object(s) is generated using the output. Data is output indicating the prediction of the identity of the chemical in the object(s) or the level of the one or more chemicals in at least one of the one or more objects.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for improved image segmentation using hyperspectral imaging. In some implementations, a system obtains image data of a hyperspectral image, the image data comprising image data for each of multiple wavelength bands. The system accesses stored segmentation profile data for a particular object type that indicates a predetermined subset of the wavelength bands designated for segmenting different region types for images of an object of the particular object type. The system segments the image data into multiple regions using the predetermined subset of the wavelength bands specified in the stored segmentation profile data to segment the different region types. The system provides output data indicating the multiple regions and the respective region types of the multiple regions.

Abstract: Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for image segmentation and chemical analysis using machine learning. In some implementations, a system obtains a hyperspectral image that includes a representation of an object. The system segments the hyperspectral image to identify regions of a particular type on the object. The system generates a set of feature values derived from image data for different wavelength bands that is located in the hyperspectral image in the identified regions of the particular type. The system generates a prediction of a level of one or more chemicals in the object based on an output produced by a machine learning model in response to the set of feature values being provided as input to the machine learning model. The system provides data indicating the prediction of the level of the one or more chemicals in the object.

Abstract: Methods, systems, apparatus, and computer-readable media for data band selection using machine learning. In some implementations, image data comprising information for each of multiple wavelength bands is obtained. A multi-layer neural network is trained using the image data to perform one or more classification or regression tasks. A proper subset of the wavelength bands is selected based on parameters of a layer of the trained multi-layer neural network, where the parameters were determined through training of the multi-layer neural network using the image data. Output is provided indicating that the selected wavelength bands are selected for the one or more classification or regression tasks.

Abstract: Implementations relate to detecting/replacing transient obstructions from high-elevation digital images, and/or to fusing data from high-elevation digital images having different spatial, temporal, and/or spectral resolutions. In various implementations, first and second temporal sequences of high-elevation digital images capturing a geographic area may be obtained. These temporal sequences may have different spatial, temporal, and/or spectral resolutions (or frequencies). A mapping may be generated of the pixels of the high-elevation digital images of the second temporal sequence to respective sub-pixels of the first temporal sequence. A point in time at which a synthetic high-elevation digital image of the geographic area may be selected. The synthetic high-elevation digital image may be generated for the point in time based on the mapping and other data described herein.

Abstract: Implementations are described herein for leveraging teleconnections and location embeddings to predict geospatial measures for a geographic location of interest. In various implementations, a plurality of reference geographic locations may be identified that are disparate from a geographic location of interest and influence a geospatial measure in the geographic location of interest. One or more features may be extracted from each of the plurality of reference geographic locations. The extracted features and a location embedding generated for the geographic location of interest may be encoded into a joint embedding. A sequence encoder may be applied to the joint embedding to generate encoded data indicative of the predicted geospatial measure.

Abstract: Implementations are described herein for generating location embeddings that capture spatial dependence and heterogeneity of data, making the embeddings suitable for downstream statistical analysis and/or machine learning processing. In various implementations, a position coordinate for a geographic location of interest may be processed using a spatial dependence encoder to generate a first location embedding that captures spatial dependence of geospatial measure(s) for the geographic location of interest. The position coordinate may also be processed using a spatial heterogeneity encoder to generate a second location embedding that captures spatial heterogeneity of the geospatial measure(s) for the geographic location. A combined embedding corresponding to the geographic location may be generated based on the first and second location embeddings.

Abstract: Methods, systems, apparatus, and computer-readable media for data band selection using machine learning. In some implementations, image data comprising information for each of multiple wavelength bands is obtained. A multi-layer neural network is trained using the image data to perform one or more classification or regression tasks. A proper subset of the wavelength bands is selected based on parameters of a layer of the trained multi-layer neural network, where the parameters were determined through training of the multi-layer neural network using the image data. Output is provided indicating that the selected wavelength bands are selected for the one or more classification or regression tasks.

Abstract: Implementations are described herein for predicting a future climate condition in an agricultural area. In various implementations, a teleconnection model may be applied to a dataset of remote climate conditions such as water surface temperatures to identify one or more of the most influential remote climate conditions on the future climate condition in the agricultural area. A trained machine learning model may be applied to the one or more most influential remote climate conditions and to historical climate data for the agricultural area to generate data indicative of the predicted future climate condition. Based on the data indicative of the predicted future climate condition, one or more output components may be caused to render output that conveys the predicted future climate condition.

Abstract: Implementations are described herein for predicting a future climate condition in an agricultural area. In various implementations, a teleconnection model may be applied to a dataset of remote climate conditions such as water surface temperatures to identify one or more of the most influential remote climate conditions on the future climate condition in the agricultural area. A trained machine learning model may be applied to the one or more most influential remote climate conditions and to historical climate data for the agricultural area to generate data indicative of the predicted future climate condition. Based on the data indicative of the predicted future climate condition, one or more output components may be caused to render output that conveys the predicted future climate condition.

Abstract: Implementations are described herein for predicting a future climate condition in an agricultural area. In various implementations, a teleconnection model may be applied to a dataset of remote climate conditions such as water surface temperatures to identify one or more of the most influential remote climate conditions on the future climate condition in the agricultural area. A trained machine learning model may be applied to the one or more most influential remote climate conditions and to historical climate data for the agricultural area to generate data indicative of the predicted future climate condition. Based on the data indicative of the predicted future climate condition, one or more output components may be caused to render output that conveys the predicted future climate condition.