Electronic device and method for detecting tool state based on audio

An electronic device comprising:

• at least one processor;
• and
• a storage device coupled to the at least one processor and storing instructions for execution by the at least one processor to cause the at least one processor to:
• acquire a cutting sound of a tool during a cutting process;
• divide the acquired cutting sound into a plurality of recordings of audio according to a preset time interval;
• extract time-frequency features of the plurality of recordings of audio respectively according to multiple feature transformation methods, wherein the multiple feature transformation methods include a short-time Fourier transform, a wavelet transform, and a Gabor transform, the time-frequency features comprise a plurality of time-frequency images;
• form a fusion feature image of the cutting sound according to the extracted time-frequency features;
• calculate a total number of pixels N according to a number of pixels of the plurality of time-frequency images generated by the multiple feature transformation methods;
• determine a closest square root n of the total number of pixels N;
• form the fusion feature image with the number of pixels n*n by sequentially arranging the pixels of the time-frequency image generated by each of the multiple feature transformation methods;
• generate a tool detection model by training a convolutional neural network model using a plurality of fusion feature images;
• take the plurality of fusion feature images corresponding to a plurality of tools in a plurality of known states as a training set;
• generate the tool detection model by training the convolutional neural network model using the training set;
• and
• detect a state of an operating tool according to the tool detection model.