分数阶微分算子结合正则化的煤层包裹体图像边缘识别算法

Edge recognition algorithm of coal seam inclusion image based on fractional differential operator and regularization

为准确识别煤层中包裹体图像边缘,提升截齿自适应截割煤岩能力,通过伪彩化煤层包裹体实际图像,利用Tikhonov正则化方法预处理识别的煤层包裹体图像,将一阶Sobel与二阶Laplacian边缘图像识别算法扩展到分数阶模式,构造其图像边缘识别的分数阶微分模式,实现煤层中包裹体结构的快速识别及边缘特征的提取。结果表明,相比整数阶微分算子,所提算法能够更好地识别煤层包裹体图像的边缘,执行时间分别为0.2221、0.2451、0.1413、0.1732 s,占用内存分别为36.35、39.82、28.99、33.13 MB,其最佳的分数阶次值为0.6。该方法可以有效识别煤层包裹体边缘特征,对噪声具有较强的鲁棒性。

This paper aims to identify the edge of coal seam inclusion image accurately and to improve the adaptive ability to cut coal and rock.The study involves pseudo coloring the actual image of coal seam inclusion and preprocessing the coal seam inclusion image with Tikhonov regularization method,constructing the fractional differential mode of image edge recognition by extending the fractional mode of the first-order Sobel and second-order Laplacian to the fractional differential mode of image edge recognition,and obtaining the rapid recognition of inclusion structure in coal seam and the extraction of edge features.The results show that the proposed algorithm can identify the edge of coal seam inclusion image better compared with the integer order differential operator with the execution time 0.2221,0.2451,0.1413 and 0.1732 s respectively,with the memory occupied 36.35,39.82,28.99 and 33.13 MB respectively,and the optimal fractional order value is 0.6.The proposed method can identify the edge features of coal seam inclusions effectively and has strong robustness to noise.