基于分形维数及细观孔隙结构特征的煤矸石识别研究

为进一步提升基于图像特征的煤矸石分选识别率,将分形维数的分析方法与图像处理和识别技术相结合,选取煤矸石图像的细观孔隙结构特征作为主要研究对象和识别特征。在运用图像处理技术对煤矸石图像进行处理后,最大化地凸显了煤矸石图像细观孔隙结构特征,通过对部分样本的孔隙结构进行初步特征测定和比对,运用分形几何理论对自然形态自相似性特征的描述能力,对煤矸石图像的细观孔隙结构进行分形维数计算,进一步凸显了煤矸石图像在细观孔隙结构特征上的差异,最后将灰度特征、纹理特征和细观孔隙结构特征结合进行识别分类,最终得到较高的识别率,这为结合分形维数的煤矸石识别技术的进一步研究和应用奠定了基础。

Research on coal gangue recognition based on fractal dimension and microscopic pore structure

In order to further improve the recognition rate of coal gangue sorting based on image features, the analysis method of fractal dimension is combined with image processing and recognition technology, and the meso-pore structure characteristics of coal gangue image are selected as the main research object and recognition feature. After using the image processing technology to process the coal gangue image, the microscopic pore structure characteristics of the coal gangue image are highlighted to the maximum. Through the preliminary feature determination and comparison of the pore structure of some samples, the fractal geometry theory is used to compare the natural shape The ability to describe self-similarity features, calculate the fractal dimension of the meso-pore structure of coal gangue images, and further highlight the differences in the meso-pore structure characteristics of coal gangue images. Finally, the gray-scale features, texture features and meso- Pore structure features are combined for recognition and classification, and finally a higher recognition rate is obtained, which lays the foundation for further follow-up in-depth research and application of coal gangue recognition technology combined with fractal dimension.