Segmentation Algorithm of Gangue Image Based on Image Eroding and Region Growing

In actual process of gangue separation,the main target is separating big gangue from coal and gangue.But images captured from running belt usually adulterate a lot of small coal particles or gangue particles,which will influence effect and efficiency of gangue separation if no filtering these particles.For above problems,the paper proposed a segmentation algorithm of gangue image based on image eroding and region growing.The algorithm firstly compresses the original image,then takes the compressed image for histogram equalization,makes image eroding process by setting suitable radius of eroding,and selects suitable seeds and threshold to make region growing.At last,It takes the processed image to compute with the original image,which can get big gangue image with clear edge.The simulation result showed that the algorithm can separate large gangue effectively,and the range of threshold of region growing of eroded image is bigger,which has reference function for images with bad edge.