The article is devoted to analyzing changes in land areas through the method of automated decoding of aerial photographs, as well as reviewing the nature and algorithm for conducting controlled classification using the maximum likelihood method and calculating the accuracy assessment of such a classification with the corresponding formulas. The object of the research is - Primorsky district of St. Petersburg. The result of the study is a thematic map that allows you to assess changes in the composition of land. In conclusion, conclusions are made about changes in land areas that occurred in the period from 2014 to 2018.