Ground truth

from Wikipedia, the free encyclopedia

Ground truth (roughly translated as "soil reality", more appropriately to name field comparison) refers to information recorded directly on the ground in remote sensing and cartography (soil samples, photos, vegetation determination, etc.), which is used to analyze remote sensing data, e.g. B. aerial photographs , satellite images or others can be used.

Ground truth data makes it possible to classify remote sensing data more precisely and to check the results for correctness. If only remote sensing data is used, this data can often not be clearly assigned. For example, there could be problems in distinguishing between maize fields and sunflower fields, because both have similar characteristics on remote sensing data (similar backscatter in different spectral ranges). The recorded information can therefore only be assigned sufficiently correctly if data from a field comparison are available at some points in the area.

Ground truth data can be collected before, during or after a flight campaign. Alternatively, ground truth data can also be used from other sources that are not determined on site.