Determining localized weather by media classification

A system comprising:

• one or more computers;
• and
• one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
• obtaining images from a camera located at a monitored property;
• determining that the monitored property is located within a particular geographic region;
• obtaining, for each of the images from the camera at the monitored property and based on the particular geographic region, an expected weather forecast for the geographic region at a time the image was captured, and an actual weather condition at the monitored property at the time the image was captured;
• generating a training set that includes each of the images labeled with both (i) an indication of the expected weather forecast for the geographic region at the time the image was captured and (ii) an indication of the actual weather condition at the monitored property at the time the image was captured;
• training a machine-learning model to classify a current weather condition for the monitored property using the training set that includes the images from the camera, the indications of expected weather forecast for the geographic region at the time the images were captured, and the actual weather condition at the monitored property at the time the images were captured;
• obtaining a subsequent image from the camera and a subsequent expected local weather forecast for the monitored property at the time the subsequent image was captured;
• providing the subsequent image from the camera and the subsequent expected local weather for the monitored property at the time the subsequent image was captured as inputs to the machine-learning model;
• and
• receiving, in response to providing the subsequent image from the camera and the subsequent expected local weather for the monitored property at the time the subsequent image was captured as inputs to the machine-learning model, a weather condition from the trained machine-learning model that indicates a particular weather condition at the monitored property based on the subsequent image from the camera and the subsequent expected local weather forecast for the monitored property at the time the subsequent image was captured.

A non-transitory computer-readable medium storing software comprising instructions executable by one or more computers which, upon such execution, cause the one or more computers to perform operations comprising:

• obtaining images from a camera located at a monitored property;
• determining that the monitored property is located within a particular geographic region;
• obtaining, for each of the images from the camera at the monitored property and based on the particular geographic region, an expected weather forecast for the geographic region at a time the image was captured, and an actual weather condition at the monitored property at the time the image was captured;
• generating a training set that includes each of the images labeled with both (i) an indication of the expected weather forecast for the geographic region at the time the image was captured and (ii) an indication of the actual weather condition at the monitored property at the time the image was captured;
• training a machine-learning model to classify a current weather condition for the monitored property using the training set that includes the images from the camera, the indications of expected weather forecast for the geographic region at the time the images were captured, and the actual weather condition at the monitored property at the time the images were captured;
• obtaining a subsequent image from the camera and a subsequent expected local weather forecast for the monitored property at the time the subsequent image was captured;
• providing the subsequent image from the camera and the subsequent expected local weather for the monitored property at the time the subsequent image was captured as inputs to the machine-learning model;
• and
• receiving, in response to providing the subsequent image from the camera and the subsequent expected local weather for the monitored property at the time the subsequent image was captured as inputs to the machine-learning model, a weather condition from the trained machine-learning model that indicates a particular weather condition at the monitored property based on the subsequent image from the camera and the subsequent expected local weather forecast for the monitored property at the time the subsequent image was captured.