Image augmentation method, electronic device and readable storage medium

An electronic device comprising:

• a storage device;
• at least one processor;
• and
• the storage device storing one or more programs, which when executed by the at least one processor, cause the at least one processor to:
• obtain an image to be augmented and a test image, the test image comprising a gas leakage image;
• construct a variational learner and a discriminator based on a fully convolutional neural network;
• obtain a target image by inputting the gas leakage image into the variational learner;
• obtain a variational autoencoder model by training the variational learner based on a discrimination result of the discriminator on the target image;
• calculate a reconstruction accuracy rate of the variational au ° encoder model based on the test image;
• obtain an augmented model by adjusting the variational autoencoder model based on the gas leakage image, in response that the reconstruction accuracy rate is less than a preset threshold;
• and
• obtain an augmented image by inputting the image to be augmented into the augmented model.

A non-transitory storage medium having instructions stored thereon, when the instructions are executed by a processor of an electronic device, the processor is caused to perform an image augmentation method, wherein the method comprises:

• obtaining an image to be augmented and a test image, the test image comprising a gas leakage image;
• constructing a variational learner and a discriminator based on a fully convolutional neural network;
• obtaining a target image by inputting the gas leakage image into the variational learner;
• obtaining a variational autoencoder model by training the variational learner based on a discrimination result of the discriminator on the target image;
• calculating a reconstruction accuracy rate of the variational autoencoder model based on the test image;
• obtaining an augmented model by adjusting the variational autoencoder model based on the gas leakage image, in response that the reconstruction accuracy rate is less than a preset threshold;
• and
• obtaining an augmented image by inputting the image to be augmented into the augmented model.