Search

The average grade of iron ore is low in China, and it is significant for the effective utilization of iron ore to obtain grades. In the paper, the high-precision model was constructed based on the spectrum of hematite and the multiple hidden layers extreme learning machine with random values and dense connection, which took the output of all previous layers as the input of each hidden layer. It has achieved high precision during the detection of the total iron content for hematite, which meets the requirements for the construction and development of new green and intelligent mines. [ABSTRACT FROM AUTHOR]

[Display omitted] • Data augmentation based on spectral difference and content difference. • Spectral subtraction based on maximum or minimum values. • Augmentation of all data or training data based on spectral and content differences. • Model construction based on spectral difference and content difference. • Prediction for content of new samples through ensemble averaging. Soil analysis makes for developing precision agriculture and monitoring land quality, while the models available for spectroscopy-based chemometrics are constrained by limited samples from small areas. The paper proposed sample expansion and model construction based on spectral difference and content difference, realizing data augmentation and deep learning applied to original samples with limited numbers. The spectral subtraction based on maximum or minimum values exploited the maximum or minimum values to acquire the spectral difference and content difference, which provided a new data form for model construction. Keeping enhanced samples whose spectral difference and content difference were all zero was useful for improving model performance. Augmentation of all data or training data based on maximum or minimum values-based spectral subtraction, which sorted the contents and made them the maximum or minimum values in sequence, achieved sample expansion by the spectral difference and content difference. The model utilized the random vector functional link (RVFL) network, extreme learning machine (ELM), and one-dimensional convolutional neural network (1D CNN), which could predict the content of new samples through ensemble averaging when predicting content difference. The experimental result showed the model of the spectral subtraction based on maximum or minimum values had a similar performance to that of the original samples. Augmentation of all data improved model performance by only RVFL and ELM. Augmentation of training data verified 1D CNN was better than RVFL and ELM. The paper implements a new data augmentation method and applies CNN to original samples with inadequate numbers, which lays the foundation for an improved model and applying spectral preprocessing. [ABSTRACT FROM AUTHOR]

Through remote sensing inversion, the evolution and treatment of saline and alkaline land can be monitored on a large scale in real time. The spectral data and salt content of samples for saline and alkaline land are obtained, and the inversion model is an improvement to the extreme learning machine, which is based on the activation function and the seagull optimization algorithm (SOA). First, the paper proposes the two-hidden-layer extreme learning machine (TELM) based on the hyperparameter activation function and the seagull optimization algorithm (HS-TELM). Then, opposition-based learning (OBL), the Cauchy distribution, and the inverse S-shaped functions (Anti-softsign, Anti-tanh, Anti-sigmoid) are used to improve the SOA, resulting in an improved seagull optimization algorithm (ISOA). At the same time, a series of improved algorithms based on the ISOA and the TELM, including the TELM based on the improved seagull optimization algorithm (ISOA-TELM), the TELM with different hidden layers containing different activation functions (D-TELM) based on the improved seagull optimization algorithm (ISOA-D-TELM), the TELM based on the hyperparameter activation function and the improved seagull optimization algorithm (ISOA-HS-TELM), and the TELM with different hidden nodes containing different activation functions (ND-TELM) based on the improved seagull optimization algorithm (ISOA-ND-TELM), are proposed. The experimental results show that the accuracy of the distribution map of salt content is 2.203 g/kg after adding the spectral correction model, which yields the expected high-precision inversion. [ABSTRACT FROM AUTHOR]

Iron ore is an important raw material for the steel industry, so it is of great economic significance to determine the grade of the iron ore quickly and accurately. And the TFe content is the main indicator that determines the grade of the iron ore and whether the iron ore can be smelted directly. Unlike manual methods and methods for chemical analysis, the paper uses the selection of band for the near-infrared spectrum based on the pruning method and the two-hidden-layer extreme learning machine based on LU decomposition and seagull optimization algorithm (LU-TELM-SOA) to identify the TFe content. First of all, the paper proposes the selection of band based on the pruning method to retain the sensitive band of the near-infrared spectrum. Aiming at the problems of poor stability and low accuracy of a single LU-TELM (the two-hidden-layer extreme learning machine based on LU decomposition) model, the paper proposes LU-TELM-SOA. The experimental results show that LU-TELM-SOA has the advantages of high accuracy and strong stability. [ABSTRACT FROM AUTHOR]

• A remote sensing retrieval method for saline and alkaline land based on RV-MELM. • We proposed RV-TELM, SRV-MELM, and LRV-MELM. • We used random value to solve the weight and threshold. The management of saline and alkaline land requires accurate distribution data, so it is necessary to construct a high-performance inversion model to get distribution data. The paper gets data of reflectance spectroscopy and salt content based on 62 samples in Zhenlai County. Then the paper builds the inversion model and calibration model, and draws a distribution map of salt content to make governance recommendations. The paper proposes the two-hidden-layer extreme learning machine with random values (RV-TELM) and the multiple hidden layers extreme learning machine with random values (RV-MELM) by using random value. Based on whether to specify the root mean square error (RMSE) and coefficient of determination (R2) of the test set that you want to obtain in the end, RV-MELM can be divided into two categories: RV-MELM with specified root mean square error and coefficient of determination (SRV-MELM) and RV-MELM with the largest number of searches (LRV-MELM). The paper selects LRV-MELM as the inversion model, and SRV-MELM as the calibration model. The experimental results show that the overall accuracy of the final map is 1.5286. [ABSTRACT FROM AUTHOR]