Search

RGB-T Salient object detection (SOD) is to accurately segment salient regions in both visible light images and thermal infrared images. However, most of existing methods for SOD neglects the critical complementarity between multiple modalities images, which is beneficial to further improve the detection accuracy. Therefore, this work introduces the MSEDNet RGB-T SOD method. We utilize an encoder to extract multi-level modalities features from both visible light images and thermal infrared images, which are subsequently categorized into high, medium, and low level. Additionally, we propose three separate feature fusion modules to comprehensively extract complementary information between different modalities during the fusion process. These modules are applied to specific feature levels: the Edge Dilation Sharpening module for low-level features, the Spatial and Channel-Aware module for mid-level features, and the Cross-Residual Fusion module for high-level features. Finally, we introduce an edge fusion loss function for supervised learning, which effectively extracts edge information from different modalities and suppresses background noise. Comparative demonstrate the superiority of the proposed MSEDNet over other state-of-the-art methods. The code and results can be found at the following link: https://github.com/Zhou-wy/MSEDNet. • Proposing three novel modules for effective cross-modal information fusion. • Presenting a novel loss function for precise refinement of local features. • Achieving excellence in RGB-T salient object detection benchmarks with MSEDNet. [ABSTRACT FROM AUTHOR]

A novel integrated model predictive control (MPC) strategy for batch processes is proposed in this paper. Both batch-axis and time-axis information are integrated into a two-dimensional control frame. The control law is obtained through the solution of a MPC optimization with time-varying prediction horizon, which leads to superior tracking performance and robustness against disturbance and uncertainty. Moreover, both model identification and dynamic R-parameter are employed to compensate the model-plant mismatch and make zero-error tracking possible. Next, the convergence analysis and tracking performance of the proposed integrated model predictive learning control system are described and proved strictly. Lastly, the effectiveness of the proposed method is verified by an example. [ABSTRACT FROM AUTHOR]

Abstract: The measurement of power plant boiler drum water level has important significance. In this article, a new approach of estimating the position of steam-water with the image processing technology was proposed. Firstly the quality of images should be improved by using image preprocessing. Then the height of water level gauge should be determined according to the process of binary and edge extraction. This method is proved accurately in estimating drum water level through the results of experiments. And it has applications on power plant. [Copyright &y& Elsevier]

Abstract: A fault diagnosis method of integrate of D-S evidential theory based on neural network was presented for turbine. In this method, preliminary diagnosis of fault information with BP and RBF network constitute the initial diagnosis layer; then, the each result of preliminary diagnosis as a body of evidence are fused under the D-S evidential theory for obtaining diagnostic results. Compared to a single diagnosis method, the simulation results show that the proposed method can effectively diagnose the vibration fault of turbine, can avoid misdiagnosis, and has good application prospects. [Copyright &y& Elsevier]

In this paper, a neuro-fuzzy based identification procedure for Hammerstein output error model with colored noise is presented. Separable signal is used to realize the decoupling of the identification of dynamic linear part from that of static nonlinear part, and then correlation analysis method is adopted to identify the parameters of the linear part. Next, a filter is embedded to form extended Hammerstein model to calculate the noise correlation function by the information of zeros and poles of the extended model. The correlation functions which consist of the noise correlation function are applied to compensate the bias caused by colored noise. As a result, the parameters of the nonlinear part can be identified through recursive least square method. Examples results illustrate that the proposed approach has high identification accuracy and good robustness to the disturbance of colored noise. [ABSTRACT FROM AUTHOR]

• Considering the problems of constraints, nonlinearity and parameter uncertainties, we add a robust iterative learning controller into 2D system. Then integrated control methods are proposed. • The analyses of robust convergence and tracking performance based on 2D system are given in this paper. • The proposed control strategy can effectively improve the robustness of the system. This paper studies the problem of integrated control in the 2-dimensional (2D) system with parameter uncertainties for batch processes. An integrated iterative learning control (ILC) strategy based on quadratic performance for batch processes is proposed. It realizes comprehensive control by combining robust ILC in batch-axis with model predictive control (MPC) in time-axis. The design of quadratic-criterion-based ILC for the system can be converted into a min-max problem. Then a model predictive controller with time-varying prediction horizon is designed based on a quadratic cost function. For an uncertain model, a novel integrated robust ILC scheme based on a nominal model is further proposed. As a result, the control law of the 2D system can be regulated during one batch, which leads to good tracking performance and strong robustness against the disturbance and the uncertainties. Moreover, the analyses of the convergence and tracking performance are given. The proposed methods are applied to batch reactor, and results demonstrate that the system has good robustness and convergence. This paper provides a new way for batch processes control. [ABSTRACT FROM AUTHOR]