摘要
针对当前铜矿筛选过程中的浮选工艺,提高该工艺对铜矿的识别率,进而提高铜矿的产量的利用,并最终提高铜矿资源的利用率,是当前思考的重点。结合传统图像分割存在的问题以及小波变换的特点,构建了一个小波变换的二值化泡沫图像处理方式。首先借助小波变换得到多尺度特征,然后利用频率和空间宽度对图形进行重构,并借助二值化对图像的特征进行提取,最后得到子图的等效尺寸特征,最后以该特征作为聚类的数据特征,通过聚类分析方法对特征进行分类,进而验证通过该方法得到的识别率要明显高于传统的识别方法,验证了该方法的可行性。
In view of the current flotation process of copper ore screening process,we are improving the recognition rate of copper ore in the utilization of copper production. We know how to ultimately improve the utilization ratio of copper resources is the key points of current thinking. In this paper,we are combining the existing problems of traditional image segmentation with the characteristics of wavelet transform. We have found that a wavelet transform two value foam image processing method is constructed. Firstly,it is wavelet transform with multi-scale feature.Then,we use the frequency width and space to reconstruct the graphics in binarization for image feature extraction.The equivalent size sub images are obtained. Finally,with the feature as feature data clustering,we classify the feature through the analysis of the clustering method to verify the identification obtained by this method. We think they are higher than the traditional identification methods to verify the feasibility of the method.
出处
《中国锰业》
2018年第1期118-121,共4页
China Manganese Industry
基金
咸阳师范学院2015
2017年科研立项课题(15XSYK005
XSYK17024)
关键词
小波变化
铜矿浮选工艺
图像特征
纹理
多尺度
Wavelet change
Copper flotation process
Image feature
Texture
Multiscale
