The invention discloses a
power equipment intelligent inspection
system for multi-scene self-adaption, and relates to the field of
power equipment inspection, a scene voiceprint modeling module generates a scene three-dimensional model through
laser radar scanning and marks equipment identity data, and a
sound level meter is arranged to collect equipment audio signals; clustering after extracting
frequency spectrum features, and eliminating anomalies to construct an initial voiceprint model; the
adaptive learning module obtains real-time temperature by using a temperature sensor, calculates a temperature
influence coefficient and a correction value, corrects the initial model to obtain a voiceprint
baseline model, and updates the voiceprint
baseline model regularly; the inspection positioning module acquires coordinates and postures of inspection equipment, calculates a distance and an
azimuth angle between the inspection equipment and the equipment, and generates predicted synthetic voiceprint data according to a formula; and the abnormity identification module collects real-time voiceprint data, compares the real-time voiceprint data with prediction data, identifies abnormity when the fluctuation range is exceeded, and controls the inspection equipment to move and position abnormal equipment according to the distance. According to the
system, the problems of multi-device sound source interference and temperature influence are solved, and the routing inspection reliability and the abnormity positioning efficiency are improved.