Abstract: A system of one or more computers can be configured to perform particular operations or actions by virtue of having software, firmware, hardware, or a combination of them installed on the system that in operation causes or cause the system to perform the actions. One general aspect includes an intelligent irrigation system including one or more of a sprinkler and one or more processors configured to: obtain imagery data regarding a plant; determine a current life cycle stage of the plant based on the imagery data, determine a plant crop coefficient for the plant based on the current life cycle stage of the plant, determine irrigation water output for the plant based on the plant crop coefficient, potential evapotranspiration calculated using solar radiation, temperature, wind speed and/or any other factors, and control the sprinkler to irrigate the plant based on the irrigation water output determined for the plant.

Abstract: In various embodiments, information regarding weather sensing devices in an observation network can be collected. Value differences such as root-mean-square errors (RMSEs) can then be calculated for values of a weather variable based on the collected information. A relationship between the value differences and distances among the weather sensing devices can be determined. A pre-determined value difference for the weather variable such as a pre-determined RMSE for the weather variable can be obtained. A distance for positioning two weather sensing devices in the observation network can be determined based on the relationship and the pre-determined value difference. One or more instructions may be issued based on the determined distance for a configuration of the observation network.

Abstract: Embodiments can facilitate performing a localized weather analysis for a portion of a sky. For achieving this, optical sensors can be placed in a capture grid network to capture images of the sky portion. The capture grid network may be logically divided into grids. The images captured by the optical sensors may include meta information indicating time and locations on the capture grid network the images were captured. The images can be combined to obtain wide-view images for the sky portion at various points of time. The wide-view images can be used to perform weather analysis such as cloud analysis.

Abstract: Embodiments can facilitate performing a localized weather analysis for a portion of a sky. For achieving this, optical sensors can be placed in a capture grid network to capture images of the sky portion. The capture grid network may be logically divided into grids. The images captured by the optical sensors may include meta information indicating time and locations on the capture grid network the images were captured. The images can be combined to obtain wide-view images for the sky portion at various points of time. The wide-view images can be used to perform weather analysis such as cloud analysis.

Abstract: Techniques for cloud forecast using configuration engine, segment cloud coverage determination engine, and trained forecast models are described herein. The disclosed techniques include generating a configuration mask based on a type of an optical component for capturing a plurality of sky images and determining cloud coverages of a plurality of corresponding segments by applying the configuration mask and a segment cloud coverage determination algorithm to the plurality of images. The disclosed techniques include training a plurality of forecast models for forecasting cloud coverages of the plurality of corresponding segments using the determined cloud coverages. The plurality of forecast models correlate cloud coverages of the plurality of corresponding segments at a time point with a cloud coverage of any particular segment at a later time point. The plurality of trained forecast models generate data indicative of cloud coverages of the plurality corresponding segments at a future time point.

Abstract: Techniques for local weather forecast using a local weather forecast model are described herein. The disclosed techniques include generating data indicative of future weather conditions for a plurality of locations based at least on future forecast data provided by an existing forecast provider and a correlation between data collected by a plurality of data collection devices and historical data provided by the existing forecast provider. The disclosed techniques also include temporally and spatially interpolating the generated data and obtaining higher-frequency future weather forecast data for the plurality of locations and local weather forecast data for other locations than the plurality of locations.

Abstract: Techniques for cloud detection and cloud coverage determination are described herein. The disclosed techniques include detecting a plurality of color clusters among a plurality of data elements of an image, associating at least one of the plurality of color clusters with a cloud based on a characteristic color of the cloud, and determining a number of data elements within the at least one of the plurality of color clusters associated with the cloud. The disclosed techniques also include detecting a location of an object indicative of a cloud in the image based at least in part on training data and a plurality of features extracted from the image, detecting an edge of the object indicative of the cloud at the location, and determining a number of data elements within the edge of the object indicative of the cloud.

Abstract: Techniques for cloud forecast using configuration engine, segment cloud coverage determination engine, and trained forecast models are described herein. The disclosed techniques include generating a configuration mask based on a type of an optical component for capturing a plurality of sky images and determining cloud coverages of a plurality of corresponding segments by applying the configuration mask and a segment cloud coverage determination algorithm to the plurality of images. The disclosed techniques include training a plurality of forecast models for forecasting cloud coverages of the plurality of corresponding segments using the determined cloud coverages. The plurality of forecast models correlate cloud coverages of the plurality of corresponding segments at a time point with a cloud coverage of any particular segment at a later time point. The plurality of trained forecast models generate data indicative of cloud coverages of the plurality corresponding segments at a future time point.

Abstract: Techniques for cloud detection and cloud coverage determination are described herein. The disclosed techniques include detecting a plurality of color clusters among a plurality of data elements of an image, associating at least one of the plurality of color clusters with a cloud based on a characteristic color of the cloud, and determining a number of data elements within the at least one of the plurality of color clusters associated with the cloud. The disclosed techniques also include detecting a location of an object indicative of a cloud in the image based at least in part on training data and a plurality of features extracted from the image, detecting an edge of the object indicative of the cloud at the location, and determining a number of data elements within the edge of the object indicative of the cloud.

Abstract: Techniques for local weather forecast using a local weather forecast model are described herein. The disclosed techniques include generating data indicative of future weather conditions for a plurality of locations based at least on future forecast data provided by an existing forecast provider and a correlation between data collected by a plurality of data collection devices and historical data provided by the existing forecast provider. The disclosed techniques also include temporally and spatially interpolating the generated data and obtaining higher-frequency future weather forecast data for the plurality of locations and local weather forecast data for other locations than the plurality of locations.