Mine flood perception based on gray level co-occurrence matrix and regression analysis

Aiming at problems of low recognition rate and poor stability and timeliness when image recognition was used in mine flood perception, a mine flood perception method based on gray level co-occurrence matrix and regression analysis was proposed. Gray co-occurrence matrix of sample image is calculated, and contrast, dissimilarity, homogeneity, entropy, correlation and energy of the gray co-occurrence matrix are extracted as eigenvalues to form eigenvectors. The classifier is determined based on the sum of the minimum distance from the eigenvector of the sample image to nonlinear regression equation, and flood is identified by the classifier. The experimental results show that recognition rate of the method on data set composed of anthracite, sandstone and surging water is 96.33% for image with resolution of 256×256, and average time-consuming of single image is 16.288 5 ms.