一种铜钼矿浮选泡沫图像分割算法

A Segmentation Algorithm of Copper-molybdenum Mine Flotation Froth Image

该文对铜钼矿浮选泡沫表面特征模糊、亮度不均、存在高亮点等现象,提出来自适应形态学的改进分水岭的一种分割方法。首先针对现场泡沫亮度不均的现象,通过多尺度的自适应Retinex算法来使浮选泡沫图像的识别度增强;然后,采用重构开闭运算滤除不同的噪声;最后,将梯度值经过阈值和尺度的非线性变换后的图像作为输入图像进行分水岭分割。结果表明该算法是一种适用于浮选泡沫图像的分割方法,很好的解决了浮选泡沫图像的分割问题。

In view of texture complexity,fuzzy boundaries and uneven illumination of metal flotation froth,an segmentation method for adaptive morphology improved watershed are introduced in the paper.Firstly,aiming at the uneven field illumination,adaptive multi-scale Retinex algorithm is applied to enhance the blurry forth image;Seconedly,opening and closing operation is reconstructed so as to filter various noises;Finally,After the threshold and scale nonlinear transformation of the gradient,the image as input image watershed segmentation.The study provided another segmentation alternative for the flotation process,and provide an idea for the subsequent flotation process automation.