Thunder detection method and system based on image recognition

Abstract

The invention relates to a thunder detection method and system based on image recognition, and the method comprises the steps: collecting a sound signal which contains or does not contain fan sweepingpneumatic noise in a period of time through audio equipment, and achieving the recognition of thunder in the sound signal through the further analysis of the sound signal. The thunder detection system based on image recognition comprises a sound sensor used for collecting sounds of a draught fan, the environment and the like, a machine-end hardware device and application software running on a station-end server, and the machine-end hardware device comprises an edge hardware data collection system. According to the application, through the deep learning model of image classification, the timedomain and frequency domain feature expressions of thunder on the spectrogram can be obtained without artificially exploring and selecting sound features or carrying out noise reduction or preprocessing on the sound, and the generalization of outdoor scenes in different environments is higher.

Description

technical field [0001] The present application relates to a thunder detection method and system based on image recognition, which is suitable for the technical field of thunder detection. Background technique [0002] As one of the core components of the wind collection device of the wind turbine, the blade is mainly responsible for capturing wind energy and driving the generator to rotate, and its performance directly affects the performance of the entire wind turbine. Blade operation and maintenance and frequent accidents will also seriously affect the comprehensive benefits of the wind farm. Among them, the most common blade damage is the damage caused by lightning strikes. Because wind turbines often operate in areas where personnel are difficult to reach in time, and with the increase of installed capacity, due to the limitation of operation and maintenance resources, lightning strike damage to blades is often difficult to detect and repair in time, and long-term opera...

AI Technical Summary

Problems solved by technology

Although this type of method judges through two signals of sound and light, the judgment method is still based on rules, so it is also impossible to distinguish the noise in some environments.

[0005] The solutions listed above do not use any image-based methods, and the sound-based solutions are easily affected by various environmental noises during collection, and are prone to false positives and false positives in actual scenarios.

Images