Satellite imagery analytics is software that sells the answer instead of the picture: which fields are stressed, what has changed since yesterday, and which ships have switched off their transponders. A transponder is the device that broadcasts a ship's position, and some ships switch it off to hide.
The satellites take far more images than people can look at. So the analytics company takes in the images from Earth observation satellites, and geospatial AI reads them. It marks crops, water, buildings, roads and ships, and compares each new image with older ones to find what has changed. The result is delivered as a map, a number or an alert, often on a geographic information system (GIS).
Here is a typical case. A crop insurer in Madhya Pradesh wants to know how many hectares of soybean were damaged by a dry spell. Sending surveyors to every village would take weeks. The analytics service measures how green each field is, compares it with last year, and reports that 18,000 hectares are stressed. The insurer uses this to start the claims process. This is crop monitoring.
To find a ship that has gone dark, the software looks for a ship in a radar picture where no transponder signal was received at that place and time. A coast guard or a shipping insurer can then check it.
The same method serves other buyers. A utility finds trees near its power lines. A property insurer gets property risk data. A surveyor can combine satellite pictures with drone surveys for finer detail.
The buyer pays for the answer, not for each image, usually as a subscription for an area or a number of questions. That is why the same picture can be sold many times, each time with a different question. The answers are checked against facts on the ground, such as a field visit, because a model can mistake one crop for another.