In the iron reverse flotation production process, the amount of flotation agent and the quality of flotation products are usually judged according to the grade of tailings, so it is essential to measure the grade of tailings froth. This research applies computer vision and image feature extraction technology to the soft sensor of tailings froth grade. An adaptive selection method for the image target region is proposed. The relationship between RGB (Red, Green, Blue), HSI (Hue, Saturation, Intensity), and Lab color space and tailings grade of reverse flotation in iron mine has been analyzed. A new image feature is proposed to characterize the degree of froth mineralization. The RGB and HSI dual color space feature values and froth mineralization degree values are determined as input, and the tailing grade soft sensor model is established by the multilayer feedforward perceptrons and VGG-19 neural network. A tailings grade soft sensor system has been developed and applied in a flotation workshop. The results of industrial tests show that this method is efficient and reliable.
AiMXieYXieS, et al. (2020) Fuzzy association rule-based set-point adaptive optimization and control for the flotation process. Neural Computing and Applications32(17): 14019–14029.
2.
AiMXieYXuD, et al. (2018) Data-driven flotation reagent changing evaluation via union distribution analysis of bubble size and shape. The Canadian Journal of Chemical Engineering96(12): 2616–2626.
3.
ChaiTDingJYuG, et al. (2014) Integrated optimization for the automation systems of mineral processing. IEEE Transactions on Automation Science and Engineering11(4): 965–982.
4.
DoukimCADarghamJAChekimaA (2010) Finding the number of hidden neurons for an MLP neural network using coarse to fine search technique. In: 10th international conference on information science, signal processing and their applications (ISSPA 2010), Kuala Lumpur, Malaysia, 10–13 May, pp. 606–609. New York: IEEE.
5.
FuYAldrichC (2016) Flotation froth image analysis by use of a dynamic feature extraction algorithm. IFAC-PapersOnLine49(20): 84–89.
HuangYChungT (2011) Modelling and analysis of air traffic control systems using hierarchical timed coloured Petri nets. Transactions of the Institute of Measurement and Control33(1): 30–49.
8.
JahedsaravaniAMassinaeiMMarhabanMH (2016) Application of image processing and adaptive neuro-fuzzy system for estimation of the metallurgical parameters of a flotation process. Chemical Engineering Communications203(10): 1395–1402.
9.
JahedsaravaniAMassinaeiMMarhabanMH (2017a) Development of a machine vision system for real-time monitoring and control of batch flotation process. International Journal of Mineral Processing167: 16–26.
10.
JahedsaravaniAMassinaeiMMarhabanMH (2017b) An image segmentation algorithm for measurement of flotation froth bubble size distributions. Measurement111: 29–37.
11.
JianHLihuiCYongfangX (2020) Design of soft sensor for industrial antimony flotation based on deep CNN. In: 2020 Chinese control and decision conference (CCDC), Hefei, China, 22–24 August, pp. 2492–2496. New York: IEEE.
KistnerMJemwaGTAldrichC (2013) Monitoring of mineral processing systems by using textural image analysis. Minerals Engineering52: 169–177.
14.
LinCYiZJiangZ (2018) An alumina rotary kiln monitoring system based on flame image processing. Transactions of the Institute of Measurement and Control41(10): 2764–2771.
15.
LiuJGaoQTangZ, et al. (2020) Online monitoring of flotation froth bubble-size distributions via multiscale deblurring and multistage jumping feature-fused full convolutional networks. IEEE Transactions on Instrumentation and Measurement69(12): 9618–9633.
16.
LiuJJiangCHeJ, et al. (2020) STA-APSNFIS: STA-optimized adaptive pre-sparse neuro-fuzzy inference system for online soft sensor modeling. IEEE Access8: 104870–104883.
17.
McCoyJTAuretL (2019) Machine learning applications in minerals processing: A review. Minerals Engineering132: 95–109.
18.
MassinaeiMDoostmohammadiR (2010) Modeling of bubble surface area flux in an industrial rougher column using artificial neural network and statistical techniques. Minerals Engineering23(2): 83–90.
19.
MassinaeiMJahedsaravaniATaheriE, et al. (2019) Machine vision based monitoring and analysis of a coal column flotation circuit. Powder Technology343: 3330–3341.
20.
Melas-KyriaziL (2021) Do you even need attention?A stack of feed-forward layers does surprisingly well on ImageNet. Available at: https://arxiv.org/abs/2105.02723
21.
NakhaeiFIrannajadM (2018) Reagents types in flotation of iron oxide minerals: A review. Mineral Processing and Extractive Metallurgy Review39(2): 89–124.
22.
NakhaeiFMosaviMRSamA, et al. (2012) Recovery and grade accurate prediction of pilot plant flotation column concentrate: Neural network and statistical techniques. International Journal of Mineral Processing110–111: 140–154.
23.
PanCZhangG (2018) Development of angular velocity vector measurement of a spherical rotor based on color sensors. Transactions of the Institute of Measurement and Control40(13): 3736–3743.
24.
PopliKMariesVAfacanA, et al. (2018) Development of a vision-based online soft sensor for oil sands flotation using support vector regression and its application in the dynamic monitoring of bitumen extraction. The Canadian Journal of Chemical Engineering96(7): 1532–1540.
25.
SimonyanKZissermanA (2014) Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.
26.
SouzaFAAAraújoRMendesJ (2016) Review of soft sensor methods for regression applications. Chemometrics and Intelligent Laboratory Systems152: 69–79.
27.
TangYPengCYinS, et al. (2014) Robust model predictive control under saturations and packet dropouts with application to networked flotation processes. IEEE Transactions on Automation Science and Engineering11(4): 1056–1064.
28.
TolstikhinIHoulsbyNKolesnikovA, et al. (2021) MLP-mixer: An all-MLP architecture for vision. Available at: https://arxiv.org/abs/2105.01601
29.
VinnettLAlvarez-SilvaM (2015) Indirect estimation of bubble size using visual techniques and superficial gas rate. Minerals Engineering81: 5–9.
30.
WangJHanSShenN, et al. (2014) Features extraction of flotation froth images and BP neural network soft-sensor model of concentrate grade optimized by shuffled cuckoo searching algorithm. The Scientific World Journal2014: 1–17.
31.
XieYCaoBHeY, et al. (2016) Reagent dosages control based on bubble size characteristics for flotation process. IET Control Theory & Applications10(12): 1404–1411.
32.
XuDChenXMaA, et al. (2016) Soft measurement method for froth layer thickness based on multi visual features. In: 2016 12th world congress on intelligent control and automation (WCICA), Guilin, China, 12–15 June, pp. 2511–2516. New York: IEEE.
33.
YanHWangFHeD, et al. (2020) An operational adjustment framework for a complex industrial process based on hybrid Bayesian network. IEEE Transactions on Automation Science and Engineering17(4): 1699–1710.
34.
ZhangDGaoX (2021) Soft sensor of flotation froth grade classification based on hybrid deep neural network. International Journal of Production Research59: 4794–4810.
