An effective Image-based Improved Deep Neural Network Framework for Coal and Gangue Detection

In coal power plants, gangue is a by-product that is not wanted, even though coal is the energy source utilized the most worldwide. As a consequence of this, the removal of gangue from coal during the pre-processing stage of the mining technique is an essential step. With the development of new computer technologies, a method may be discovered that will allow for an improvement in the efficiency of gangue separation. This study presents a method for identifying gangue using improved deep neural networks. The method involves building a coal-gangue filtration system based on the distinctions between gangue and coal in terms of its rough surface and grayscale feature. In addition, this tactic will make it possible for robots, rather than people, to identify the gangue. Suppose the technology described in this article is applied in the coal industry. In that case, it will lead to the computerization of the extraction of coal gangue, an increase in the raw coal processing efficiency, and an improvement in the quality of the coal. According to the findings of experiments, the strategy enhances the accuracy results of coal and gangue materials by as much as 98.87% compared to the currently used approaches.