Drainage melioration systems are important for agricultural areas with waterlogged zones. Open (surface) and closed (subsurface) drainage objects are distinguished. A promising direction for operational state assessment of drainage systems is the use of remote sensing data. A high potential for solving such problems has the use of unmanned aerial vehicles, which can obtain high-quality images of an agricultural area in various imaging spectra in a short time. In this work, as the first stage of solving a largescale problem of automating image analysis methods for assessing the state of drainage complexes, a specialized knowledge base (KB) of the main drainage defects to be repaired, based on aerial photographs and ground measurements, is proposed. The KB structure developed on the basis of a conceptual model can be quickly implemented in web projects and applications, while allowing architectural
changes to be made. Also, on the basis of many years of field experiments, the most common defects of drainage systems objects were identified.