Figure 1 - available via license: Creative Commons Attribution 4.0 International
Content may be subject to copyright.
Maceral groups: (a) vitrinite; (b) liptinite and inertinite; (c) liptinite; (d) inertinite and vitrinite. Oil immersion, magnification 500×.
Source publication
![Figure 5. The architecture of U-Net network [71]. The U-Net is an...](publication/356211009/figure/fig3/AS:1090882942058496@1637098169318/The-architecture-of-U-Net-network-71-The-U-Net-is-an-example-of-an-autoencoder_Q320.jpg)
The study of the petrographic structure of medium- and high-rank coals is important from both a cognitive and a utilitarian point of view. The petrographic constituents and their individual characteristics and features are responsible for the properties of coal and the way it behaves in various technological processes. This paper considers the appl...
Contexts in source publication
Context 1
... identification of macerals is based on the microscopic evaluation of grain morphology and color. On this basis, three groups of macerals were distinguished: liptinite, vitrinite, and inertinite ( Figure 1) [9][10][11]. The color of liptinite changes from brown through dark grey to light grey in the microscopic image. ...
Context 2
... identification of macerals is based on the microscopic evaluation of grain morphology and color. On this basis, three groups of macerals were distinguished: liptinite, vitrinite, and inertinite ( Figure 1) [9][10][11]. The color of liptinite changes from brown through dark grey to light grey in the microscopic image. ...
Similar publications
The majority of traditional text sentiment classification techniques rely on machine learning or sentiment dictionaries, but these approaches have the drawback of sparse data and ignore word semantics and word order information. A convolutional neural network- (CNN-) based music emotion classification model is proposed in this paper to address the...
Citations
... At present, the methods of coal and gangue identification are ray method, multi-spectral identification, signal processing and image recognition method, but there are certain limitations, and failed to promote the use of large-scale . In recent years, deep learning technology through the convolutional neural network to automatically obtain and learn the image features, which can quickly extract and detect the feature information in the image of coal and gangue (Iwaszenko et al. 2021;Gao et al. 2024;Liu et al. 2024). Lv et al. (2022) propose an oversized gangue segmentation network model based on multi-task learning theory, which realizes the effective segmentation and identification of oversized gangue. ...
Accurate identification of coal and gangue is essential for clean and efficient use of coal. Existing target detection algorithms are ineffective in detecting small-target and overlapping gangue, and contain complex network structure and large parameter volume, which cannot meet the demand of real-time detection of edge devices. To address the above problems, a lightweight detection and identification approach of coal gangue based on improved YOLOv5s is proposed. The depth-separable convolutions are used to replace ordinary convolutions, and the C3 (Concentrated-Comprehensive Convolution Block) Ghost module is constructed to replace all C3 modules in the YOLOv5s to reduce model computation and parameters. The CA (Coordinate Attention) attention mechanism is introduced to strengthen the attention to the target to be detected, suppress irrelevant background interference, and improve the detection accuracy of the model. The Focal- EIOU (Focal and Efficient Intersection Over Union) loss function was introduced to replace the original CIOU. Extensive experiments substantiated the proposed approach can effectively and quickly detect the small-target and overlapping coal gangue accurately, and the mAP (mean Average Presicion) reaches 97.7%. Compared with the original YOLOv5s, the proposed approach reduces the number of parameters and the amount of computation by 48.5% and 43%, respectively, under the premise of maintaining the same detection accuracy.
... This has inspired research in the application of deep learning in this area. The positive effects of using these methods in hard coal petrography [17][18][19] and also in geology [20,21] are already known. ...
... Microscopic vision, segmentation, and deep learning have been successfully applied to recognize petrographic components and mineral substances in coal [17], Other researchers have used computer vision methods to identify minerals in rocks [20,21] or determine the grain sizes of aggregates [26]. Since the same components (coal, minerals, rocks) are present in wood pellets and are treated as contaminants, attempts have been made to apply similar microscopic and deep learning methods to their recognition. ...
... There have also been very promising attempts at using deep networks for autonomous analysis of microscopic images [18,42]. A UNet [43]-based network was used for maceral group identification in coal sample cross-sections observed with an immersed microscope yielding satisfactory results [17]. Other researchers reported promising results using selected (SegNet [44], DeepLabv3 [45]) deep networks for maceral identification [19]. ...
Pellet fuels are nowadays commonly used as a heat source for food preparation. Unfortunately, they may contain intrusions which might be harmful for humans and the environment. The intrusions can be identified precisely using immersed microscopy analysis. The aim of this study is to investigate the possibility of autonomous identification of selected classes of intrusions using relatively simple deep learning models. The semantic segmentation was chosen as a method for impurity identification in the microscopic image. Three architectures of deep networks based on UNet architecture were examined. The networks contained the same depth as UNet but with a successively limited number of filters. The input image influence on the segmentation results was also examined. The efficiency of the network was assessed using the intersection over union index. The results showed an easily observable impact of the filter used on segmentation efficiency. The influence of the input image resolution is not so clear, and even the lowest (256 × 256 pixels) resolution used gave satisfactory results. The biggest (but still smaller than originally proposed UNet) network yielded segmentation quality good enough for practical applications. The simpler one was also applicable, although the quality of the segmentation decreased considerably. The simplest network gave poor results and is not suitable in applications. The two proposed networks can be used as a support for domain experts in practical applications.
... With the development of deep learning technology in recent years, 22,23 deep learning-based target detection is used in the field of gangue detection, [24][25][26] boasting advantages over traditional machine learning methods in terms of no manual feature selection, excellent feature expression capability, and excellent detection accuracy. 27,28 Deep learning technology can automatically acquire and learn image features through a convolution neural network, thereby accelerating the extraction and detection of feature information in coal gangue images. ...
... Compared with the original YOLOv3 and YOLOv4, the mAP is increased by 0.37% and 0.05%, respectively, the fps is increased by 119.04% and 104.7%, and the memory is only 1/24 and 1/25 of the original. Compared with YOLOv3-Tiny, and YOLOv4-Tiny, although the detection speed decreases, the mAP is 2.17% and 1.57% higher, and the memory is reduced by 24 overall performance of this algorithm is better than YOLOv3, YOLOv4, YOLOv3-Tiny, and YOLOv4-Tiny. Furthermore, to intuitively show that the improved SMS-YOLOv3 algorithm can reduce false detection and false detection in special scenes (small targets, overlapping targets, large numbers, etc.), this paper selects four representative pictures for testing and analyzes them with different algorithms. ...
Aiming at the problems of large memory footprint, low detection speed, and low detection accuracy for small and overlapping targets existing in the current coal gangue target detection algorithm, a real‐time detection method for coal gangue based on a multiscale fusion lightweight network (SMS‐YOLOv3) is proposed. Taking MobileNetv3 as a feature extraction network, in which all SE modules are replaced with SKNet, thus improving the ability of image feature extraction and making more effective use of parameters. A shallow detection scale is added to form a detection structure with the fusion of four scales to improve the detection accuracy of small targets. The spatial pyramid pooling is added after the backbone network to convert different feature maps into fixed feature maps, to improve the detection accuracy of the algorithm. CIoU bounding box regression loss and the K‐means++ clustering anchorbox are used to improve the detection accuracy of targets. Experimental equipment was built, and coal gangue datasets of small size, large size, dim light, mutual concealment, and a large number of coal gangue under multiple conditions were constructed. Experiment results demonstrate the effective and fast detection of the proposed algorithm for small targets and overlapping targets of coal gangue accurately, with mAP reaching 98.97%. The algorithm has an mAP improvement of 0.37% and an fps increase of 119.04% compared with the original YOLOv3, with memory only 1/24 of the original.
... Owing to the large intra-class variability in the task of coal maceral identification, such a method faces significant challenges in terms of the generalization problem caused by the complexity and heterogeneity of coal. Owing to the increased computational power provided by deep learning systems, computer vision has facilitated significant progress in research and applications in a wide range of industries [30,31]. In particular, the emergence of the U-Net convolutional network has effectively solved the problem of poor generalization due to complexity and heterogeneity. ...
... Subsequently, it has been widely used in image segmentation tasks in various industries [34]. Researchers have improved its application to maceral image segmentation and analysis with remarkable results [30][31][32]. Transformers constitute a new approach to spatial modeling for input-adaptive, long-range dependencies. Through high-order spatial interactions via self-attention manipulation, the attention heads can learn to perform different tasks [35]. ...
... On the other hand, Iwaszenko and Róg (2021) attempted maceral identification at low rank (0.5% -0.8%) using a deep learning-based algorithm with a U-net architecture with 162 images from 12 samples. With the aid of a professional petrographer, the authors manually generated a ground truth, that is, by segmenting the individual macerals and classifying them, and used it for training the network. ...
... reflectance range due to its industrial interest, even though it is understood that maceral segmentation should be easier in lower ranks, as the difference between the macerals groups is more pronounced. Other works have attempted maceral segmentation in lower rank ranges (Iwaszenko and Róg, 2021)., which suggests that deep learning can be a valuable and effective tool for semantic segmentation in a wide rank range It is, however, understood that proper collotelinite segmentation is a critical step to determine the correct mean random reflectance and, consequently, rank evaluation. While it is possible to post-process those images, refining these results to avoid overreliance on post-processing would be desirable. ...
Coal is widely used in industrial applications such as carbonization and coke production for steel making, combustion, or gasification to generate electricity, and liquefaction to generate petrochemical feedstock. The utility of a coal is dictated by its properties, commonly referred to as its rank, type, and grade. Coal type or maceral composition and coal rank determination by vitrinite (collotelinite) reflectance are traditionally conducted manually by trained petrographers using reflected light optical microscopy and bulk chemical tests. This study developed an automatic method based on machine learning for rank determination on petrographic images of coal that might improve the efficiency of this process and potentially eliminate operator subjectivity. Firstly, a Mask R-CNN-based architecture deep learning approach was developed to identify and segment the collotelinite maceral, which is fundamental for rank analysis, as rank can be assessed by the reflectance of this maceral. Secondly, an image processing method was developed to analyze the collotelinite segmented images and determine coal rank by associating the grey values with the reflectance. For the segmentation, five samples were used to train the network, 174 images were used for training, and 86 were used in the test set, with over 80% success rates. Four of those five samples were used alongside another eight to determine the rank. The samples ranged from 0.97% to 1.8% reflectance. The root mean square error calculated between the proposed method and the reference values of reflectance was 0.0978%.
... Intuitively, the semantic segmentation algorithm [22][23][24][25] assigns each pixel to a specific category that captures the location and size of the mine dump material entirely. However, as shown in Figure 2, the semantic segmentation requires a large number of pixel-level labels. ...
... The results showed that Dense-Hg-Center reduced the error by 1.14% and the standard deviation by 12.84 compared to Hg-Center on the conglomerate data set. On the clay data set, Dense-Hg-Center reduces the error by 0.19% and the standard deviation by 25 Table 3. The experimental results indicated that the mixed training strategy is superior to the separate training, and the different material objects will assist the model in extracting discriminative features between them. ...
The granularity distribution of mine dump materials has received extensive attention as an essential research basis for dump stability and mine land reclamation. Image analysis is widely used as the fastest and most efficient method to obtain the granularity distribution of the dump materials. This article proposes a deep learning-based approach for granularity detection and identification of mine dump material, conglomerate, and clay. Firstly, a Conglomerate and Clay Dataset (CCD) is proposed to study the granularity of the mine dump. A typical study area is selected for field sampling, and the sampled conglomerate and clay is photographed and labeled. In addition, this article proposes a keypoint-based detection algorithm for the conglomerate and clay detection. The algorithm considers the scale variation of conglomerate and clay in orthophoto images and adopts center point detection to avoid the difficulty of localization. On this basis, dense convolution is introduced in feature extraction to reduce the computational redundancy to conduct detection more efficiently. Finally, the corresponding granularity distributions of conglomerate and clay are obtained by geometric calculation in the deep learning-based detection results. The proposed algorithm is validated on the proposed dataset CCD, and the experiments demonstrate the effectiveness of the proposed algorithm and its application to the granularity analysis of mine dump material.
... Jia et al. [19] performed mineral photo recognition based on feature fusion and online hard sample mining. Iwaszenko et al. [20] applied deep learning in In this paper, the sound and vibration signals are collected in real time by the state of the crusher feeding stage. After denoising and processing the collected signal, the CNN is used to classify the feeding feature images. ...
In large coal preparation plants with a capacity of 30 million tons/year, the belt speed can reach 7 m/s and the thickness of the material layer can reach 500 mm. Therefore, in high-throughput and complex environments, the problem exists that harmful feeding materials such as iron and gangue are not easily detected, and thus fault diagnosis in the crushers lags behind. Therefore, it is necessary to extract the equipment operation signals from the noisy production environment and identify the feeding materials. Currently, there is no systematic research on signal processing and image classification of crusher feeding materials, while the convolutional neural network (CNN) is outstanding in computer vision. In this paper, sound and vibration signals of the feeding materials are denoised by spectral subtraction and transformed into feature images by continuous wavelet transforms. Then, an image classification model based on CNN is built for these feature images to study its classification mechanism and performance. The results show that the model classification accuracy is respectively 84.0%, 93.5% and 80.1% in coal–iron–wood classification, coal–iron classification, and coal–wood classification. The good classification performance for coal, iron and wood can satisfy the practical demands to remove the harmful feeding materials, which provides the core technical support for the establishment of operating status monitoring and fault diagnosis system of crushing equipment.
... Advanced imaging technologies are changing geological sampling and analysis. Iwaszenko and Róg [13] provide an image analysis model to segment important geological features of coal. The modeling can speed up the analysis, thereby influencing key mineral processing decisions and earlier capturing valuable time and energy. ...
This is an exciting time for the mining industry, as it is on the cusp of a change in efficiency as it gets better at leveraging data [...]
Segmenting blasted stockpile particles in open-pit mines is essential for improving mining operations. However, complex rock textures often challenge traditional segmentation models. This paper presents an enhanced U-Net model that leverages depth-separable convolution and feature depth concatenation to enhance segmentation performance while reducing model complexity and training time. We evaluate our model on a homemade open pit burst pile ore segmentation dataset and report an average accuracy improvement of 1.53 % over the U-Net model. Our work contributes to the field of mining engineering and shows the potential of deep learning methods to improve mining operations.
