基于小波变换的浮选泡沫图像纹理特征提取

Flotation Froth Image Texture Feature Extraction Based on Wavelet Transform

随矿物浮选工况的变化,精选泡沫表面出现亮度分布不均、灰度分布差异性大的特点,表现出明显的纹理特性。为定量分析精选泡沫表面纹理特征与浮选工况间的关系,采用小波变换方法提取纹理特征均匀度,实现精选泡沫表面纹理的量化描述,并通过实验分析,获得精选泡沫的最佳纹理区间。工业现场运行结果表明,该方法可以实时监视精选泡沫表面纹理特征变化,有利于浮选生产指标的优化控制。

With the changes in flotation conditions,fine froth show obvious texture features due to the features of uneven distribution of brightness and great difference of gray level distribution.In order to analyze the relation between texture features of bubble surface and flotation conditions,a multi-resolution analysis method based on wavelet transform is presented.This work aims to extract texture features called uniformity,achieve the quantitative description of bubble surface texture,and acquire the best texture range through analysis.Experimental results in the industrial field show that the method can realize the changes of bubble surface texture in real-time monitoring and provide a basis for production index of floatation to achieve optimal control.