AI generation method for optimizing scrap steel image based on machine vision

Abstract

The invention relates to the field of metallurgy and the field of image recognition, in particular to an AI generation method for optimizing a scrap steel image based on machine vision, and the method comprises the steps: obtaining image frame data with scrap steel in a scrap steel unloading scene; preprocessing the image frame data with the scrap steel, adding noise interference items, and randomly dividing the image frame data into a data set according to a proportion; establishing a generative adversarial model, and training the generative adversarial model to obtain a trained generative model; inputting video stream image frame data obtained in real time in the unloading scene into the trained generative model; and sending a result output by the trained generative model to a scrap steel image optimization system. The method has the advantages that the network model is based on the deep learning generative adversarial network technology, scrap steel images in different environments and denoised and deblurred snapshot images are automatically generated in a scrap steel unloading site unloading scene, and the improved generative adversarial network model meets the requirement for sample set shortage in extreme situations.

Images