The combustion of gangue will release toxic gases, and sorting out the mixed gangue before burning the raw coal is one of the crucial steps in clean coal technology. However, the mutual occlusion caused by the dense distribution of raw coal makes the detection methods susceptible to misdetect the targets. In view of the above problems, a detection network for densely distributed coal and gangue with uneven particle size is proposed based on DeepLabV3+. Firstly, group convolution and 1×1 convolution are used in the encoder to reduce the model parameters, and deformable convolution is introduced to enhance the feature extraction capability for irregularly shaped targets. Then, two shallow feature maps and CA are introduced into the decoder, to enhance the network's focus on the edge detailed information of the adhesion target and the small target location information. Finally, relying on the self-collected multi-condition coal and gangue dataset, the method's effectiveness is demonstrated through comparative validation and test experiments, with an average F1-score of 0.983 and a detection speed of 32.5 FPS. The technique can be adapted to the poor working environment in coal mine production, and realizes the low-cost and high-efficiency identification and localization of coal and gangue.
High-precision position control of hydraulic systems is of great significance in industrial applications. However, unknown external force disturbance, measurement noise, and other uncertain nonlinearities widely exist in electro-hydraulic servo systems, which severely degrade the system performance. To handle this problem, a sliding-mode controller based on K-observer with nonlinear disturbance observer is proposed for electro-hydraulic position systems in this article. The nonlinear disturbance observer is to estimate and eliminate the time-varying external force disturbance. The sliding-mode controller based on K-observer is to reach low gain observation of states and improve the system performance by reducing the chattering of the sliding-mode algorithm. The proposed controller is proved to be stable by the Lyapunov stability criterion. Comparative experiments support that the controller has high precision and strong robustness.
To solve the problems of the difficult feature extraction, poor feature credibility and low recognition accuracy of coal and gangue, this paper utilizes the difference in the dielectric properties of coal and gangue and in combination with a support vector machine (SVM) to propose a recognition method based on the dielectric characteristics of coal and gangue. The influence rule of the edge effect of the electrode plate on the capacitance value is analyzed when the thickness of the electrode plate changes. By changing the frequency and voltage of the excitation source, curves of the dielectric constant of coal and gangue versus frequency and voltage are obtained. Combined with the Kalman filter, the adaptive noise complete set empirical mode decomposition (CEEMDAN) denoising method is improved, which results in a signal with a higher signal-to-noise ratio and lower root mean square error after denoising. The effective value and frequency of the denoised response signal are extracted to construct the feature vector set to form the training set and test set. The data of the training set are input into the SVM to train the intelligent classification model, the test set is used to test the SVM classification effect, and the classification accuracy is 100%. Unlike these of the probabilistic neural network (PNN) intelligent classification model and the learning vector quantization (LVQ) neural network classification model, the recognition and classification accuracies of the three can reach 100%, but the classification speed of SVM is the fastest, only taking 0.007916 s, which fully reflects the feasibility and efficiency of the capacitance method in identifying coal gangue. In this paper, the capacitance method and SVM are applied to identify coal and gangue, and accurate and efficient identification results are obtained, providing a new feasible solution for research on coal gangue identification.
Abstract The positioning accuracy of coal and gangue is related to the discharge accuracy of gangue, which will affect the utilization rate of coal. But the detectability of the small coal and gangue is poor due to the fewer number of pixels and texture information in coal and gangue dual-energy X-ray images. So, the Otsu with crotch structure based on Adaptive Partical Swarm Optimization (APSO) for small targets detection is proposed, called after APSO-C_Otsu. Firstly, the Otsu with crotch structure is used to perform multi-threshold segmentation of coal and gangue dual-energy X-ray images to increase the contrast between small target and background. Meanwhile, the APSO algorithm was used to optimize the Otsu algorithm with crotch structure in order to improve its convergence speed and reduce its calculation amount. Finally, the processed image is binarized, and the location of the target was labeled based on the bwlabel algorithm. The experimental results revealed that the APSO-C_Otsu algorithm could effectively detect the small pixel size (less than 8 × 8 pixels) of coal and gangue with a particle size of 6 ~ 30 mm, and was also applicable to the coal and gangue with the particle size larger than 30 mm, which was of great significance for accurate separation of coal and gangue and the improvement of coal utilization.
An automatic annotation method of coal gangue image data based on X-ray acquisition is proposed.Firstly, the manually screened coal and gangue are sent to the X-ray acquisition device for image sampling.Secondly, according to the characteristics of the pure background of coal and gangue images, a gray threshold is used to detect.Further, the mask is used to purify the detected foreground objects, which obtain a single object image.At the same time, the manually screened coal/gangue is sampled and the category information of the currently detected objects is obtained at the same time, and the monomer sample database is established.Finally, based on the monomer sample library established, the monomers randomly were selected in the sample library and combined to generate sample images with classification and location.That is, the automatic annotation of coal gangue data is completed.The results show that the automatic annotation method of coal gangue image data through digital image processing can effectively annotate the collected coal gangue image data, with a high sample generation rate and annotation accuracy of 99%.
In lump coal and gangue separation based on photoelectric technology, the prerequisite of using a dual-energy X-ray to locate and identify coal and gangue is to obtain the independent target area. However, with the increase in the input of the sorting system, the actual collected images had adhesion and overlapping targets. This paper proposes a pit point detection and segmentation algorithm to solve the problem of overlapping and adhesion targets. The adhesion forms are divided into open and closed-loop adhesion (OLA and CLA). Then, an open- and closed-loop crossing algorithm (OLCA and CLCA) is proposed. We used the conjugate lines to detect the pit and judge the position and distance of the pixel point relative to the conjugate lines. Then, we set the constraint of the distance of the pixel point and the relatively straight line position to complete the pit detection. Finally, the minimum distance search method was used to obtain the dividing line corresponding to the pit to complete the image segmentation. The experiment results demonstrate that the segmentation accuracy of the overlapping target was 90.73%, and the acceptable segmentation accuracy was 94.15%.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)