摘要
Local area spectral features of copper flotation froth image is the size of the area of local black hydration in the froth image which related to flotation performance. This feature has never been used to the copper flotation process monitoring. A feature extraction method based on MPCA is proposed for the irregularity of size and shape. Firstly, a two-dimensional image data which is unfolded from the RGB froth image can be projected into unrelated principal component space based on MPCA method and the black hydration area is marked in the principal component space. Then, the number and spatial position of the marked area feature pixels in the principal component image can be determined according to the correlation of the feature pixel value in it with the score value in principal component space. The size of local area can be calculated as the local area spectral features of the froth image. The subtle spectral information of original froth image can be captured perfectly by this method Industry experimental results demonstrate the effectiveness of this method.
- 出版日期2014
- 单位中南大学