Forest fire identification method and system based on Yolov5x of diffusion model

The invention discloses a Yolov5x forest fire identification method and system based on a diffusion model, and the method comprises the steps: building a New-U-Net model, and carrying out the data enhancement processing of all original image data in the U-Net model according to the built New-U-Net model; the advantages of a genetic algorithm and a neural network in the aspect of optimizing and processing nonlinear data are combined, the genetic algorithm and the neural network are applied to feature selection in a packaging mode, and a neural network classifier is used as an evaluation tool to perform feature extraction on image data subjected to enhancement processing; and according to the enhanced image data after feature extraction, generating an improved Yov5x model suitable for forest fire detection, comparing the improved Yov5x model with the initial Yov5x model by using verification data, and verifying the precision of the improved Yov5x model through a preset evaluation index. The method has the chara