Microscale VOCs leakage detection method and system, storage medium and equipment

The invention discloses a microscale VOCs leakage detection method and system, a storage medium and equipment. The microscale VOCs leakage detection method comprises the following steps: S10, obtaining infrared images of a VOCs leakage area and a leakage-free area for data preprocessing; S20, inputting the preprocessed data into a backbone network, and performing feature learning by adopting kernel functions of different sizes; S30, filtering a large number of non-VOCs leakage gas background areas by using a coarse classification module, and extracting suspected VOCs leakage images; and S40, classifying the suspected VOCs leakage image by using a fine classification module, and carrying out regression positioning on the VOCs leakage position so as to obtain a detection result. The convolutional neural network is fully utilized to learn the features of the VOCs leakage infrared image, a large amount of background information is effectively filtered through a coarse-to-fine-granularity framework and the coarse