Image Restoration for Blast Furnace Burden Surface Based on Dust Multiscattering Model

A clear image of the blast furnace burden surface (BF-BS) is essential for monitoring abnormal furnace conditions, optimizing burden distribution operation, and ensuring stable operation of the blast furnace. Aiming at the problem that the quality of BS image is seriously degraded due to the strong light interference and dust particle scattering, a novel image restoration method for BF-BS is presented in this article. First, a structure-perceivable variational model is designed to extract the dust layer from the original image. Second, a dust multiscattering model is established based on the radiation transfer theory, and then, the image degradation model of the BS is further derived, which quantitatively describes the relationship between the degraded and restored images. On this basis, an image restoration algorithm is proposed to estimate the unknown parameters in the model so as to recover the BS image with high visibility. The experimental results show that the method can effectively improve image quality, enhance overall brightness, remove the influence of dust, and restore the detail and edge information of the BS. The restored image of the BS has important guiding significance for implementing fine regulation of the blast furnace. In addition, this method can be applied to image restoration and object detection in harsh industrial environments.