Working condition recognition based on an improved NGLDM and interval data-based classifier for the antimony roughing process. (February 2016)

• Record Type: Journal Article
• Authors: Peng, Xia; Peng, Tao; Zhao, Lin; Song, Yanpo; Gui, Weihua
• Is Part Of: Minerals engineering. Volume 86(2016)
• Page Start: 1
• Page End: 9
• Publication Date: 2016-02
• Subjects: Working condition recognition -- Improved NGLDM -- Composite texture -- Interval data -- SVM -- Antimony roughing; Mines and mineral resources -- Periodicals; Ressources minérales -- Périodiques; Mines and mineral resources; Periodicals; Electronic journals; 622
• DOI: 10.1016/j.mineng.2015.11.001
• Languages: English
• ISSNs: 0892-6875

Abstract

Highlights: An improved NGLDM is presented for an image. Defined a new texture feature of image namely composite texture. Combined with the ore grade, the numerical intervals representing the extracted composite textures are adopted for off-line classification and on-line recognition by a support vector machine (SVM) classifier for froth images. The improved NGLDM could reflect the texture of froth surface more completely. The composite texture has better accuracy in classification than traditional one. Abstract: Texture, as one of the most important features of the froth surface, is considered closely related to flotation working conditions and the production index. A working condition recognition method based on an improved neighboring gray level dependence matrix (NGLDM) and interval data classifier is proposed for the antimony roughing process. First, an improved NGLDM is presented for an image. Next, a new composite texture is defined that associates the bubble characteristics of size, shape, and roughness with a froth image. Finally, combined with the ore grade, the numerical intervals representing the extracted composite textures are adopted for off-line classification and on-line recognition by a support vector machine (SVM) classifier for froth images under different working conditions. Experiments show that the new composite texture feature extractor using the improved NLGDM has greater stability, separability and classification accuracy than the normal texture feature extractor using NGLDM does. Therefore, the interval data-based SVM classifier is feasible and effective for working condition recognition in the antimony roughing process.