提出一种基于模糊纹理谱的浮选泡沫图像的纹理特征提取方法.该方法根据人眼感知特点定义了一种基于非线性函数的隶属度函数。依据相邻像素与中心像素的灰度变化大小关系统计其模糊纹理单元,计算模糊纹理谱,提取有效描述浮选泡沫纹理的新特征参数一纹理复杂度.并以工况相对稳定情况下的泡沫图像为例,对所提出的方法进行了验证分析,结果表明其有效性,并获得了浮选泡沫的最佳纹理复杂度区间,对浮选的优化操作提供重要的指导意义.
Froth image texture contains crucial information of flotation performance. Based on the fuzzy texture spectrum, an effective texture characterization approach is proposed for accurate description. From a human perception viewpoint, the membership function of the fuzzy texture spectrum is improved by introducing nonlinear function. Using the designed membership function, the fuzzy texture units are calculated according to the intensity variation between the central pixel and its adjacent pixels. Then a novel feature is extracted to describe the froth texture complexity. Given stable operating condition, some validity analyses of the proposed method are carried out. Experimental results prove their effectiveness. Optimum texture complexity range is detected so as to provide operational guidance for flotation process control.