Search

A new image dehazing algorithm based on adaptive sky region is proposed in this paper, which shows good fidelity in sky region and satisfying visual effect in non‐sky region. For robust sky segmentation, we propose a rough‐to‐fine method that can make a balance between efficiency and accuracy. Considering distribution of haze is inconsistent, we divide the input image into three parts and calculate their atmospheric lights respectively. To solve the problem of invalid dark channel prior, we make an improvement for the transmission estimation. Finally, image fusion is taken as a post processing that can solve the problem of partial darkness and ensure a visual pleasing result. The experimental results for both synthetic and natural hazy images demonstrate that our algorithm performs comparable or even better results than the state‐of‐the‐art methods in terms of various measurement indexes, such as the PSNR, SSIM, and so forth. Besides, the proposed algorithm can be also applied in FPGA platform due to the optimized performance. © 2021 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC. [ABSTRACT FROM AUTHOR]

Summary: This article studies the parameter identification problem for bilinear state space models with time‐varying time delays. Considering the correlation of time delays, the Markov chain switching mechanism is adopted to model the delay sequence. Based on the observer canonical form, the bilinear state space model is transformed into a regressive form. A bilinear state observer is designed to estimate the state variables. Under the variational Bayesian scheme, the system parameters, discrete delays, and the Markov transition probabilities are identified by using the measurement data. A numerical example and a continuous stirred tank reactor simulation are employed to validate the effectiveness of the proposed algorithm. [ABSTRACT FROM AUTHOR]

The nonlinear sliding-mode (SLM) controller based on the maximum jerk and time-optimal step response has been developed for the control of servo-solenoid valves, which exhibits its performance advantages when compared to traditional PID controller. However, the experiments shows the responses of the SLM control under only jerk constraint will largely vary with different damping and loads, leading to the large overshoot and breakdown under some working conditions (e.g., large viscous damping coefficient). Thus, an improved SLM controller is developed in this paper to deal with the existing drawbacks, where a novel nonlinear SLM surface is proposed by taking into account both spool’s acceleration and jerk limitation. Firstly, the nominal model of servo-solenoid valve is built by linearizing the dynamics on the null position, and the system identification is carried out to achieve the model parameters and their variation ranges. Subsequently, the valve constraints under the power limitation are analyzed through the frequency response of the identified model, which comes out the maximum available velocity, acceleration and jerk. Finally, the improved SLM control algorithm is proposed, where the novel SLM surface considers the maximum plus jerk, the maximum minus jerk and the maximum acceleration. Experimental studies are conducted and the results show that the improved SLM controller under both acceleration and jerk constraints can achieve the continuous and stable sliding mode state, realize the time-optimal step response of the valve, and exhibit strong disturbance rejection abilities. [ABSTRACT FROM AUTHOR]

With increasing demands of three-dimensional (3D) contents, the conversion of many two-dimensional (2D) contents to 3D contents has become a focus in 3D image processing. The most important and difficult problem in 2D-to-3D conversion is the estimation of the depth map using only a single-view image. Therefore, a novel method for estimating a depth map for stereoscopic conversion is proposed in this paper. A simulated haze image generated by adding a haze veil on the input image is used to represent salient region segmentation. The 3D stereoscopic image is generated on the basis of the depth map estimated by the haze removal algorithm without any heuristic cues or user interaction. A comparative study and quantitative evaluation with some other state of the art conversion methods are carried out, which demonstrate that using the proposed approach, similar or better-quality results may be obtained. [ABSTRACT FROM AUTHOR]

Multi-point shortest path planning problem is a typical problems in discrete optimization. The bat algorithm is a nature-inspired metaheuristic optimization algorithm that is used in a wide range of fields. However, there is one problem with the BA, which is easy to premature. To solve multi-point shortest path planning problem, an improved discrete bat algorithm (IDBA) is proposed in this paper. In this algorithm, the Floyd–Warshall algorithm is first used to transform an incomplete connected graph into a complete graph whose vertex set consists of a start point and necessary points. Then the algorithm simulates the bats' foraging and obstacle avoidance process to find the shortest path in the complete graph to satisfy the constraints. Finally, the path is transferred to the original incomplete graph to get the solution. In order to overcome the premature phenomenon of a discrete bat algorithm, the modified neighborhood operator is proposed. To prove the effectiveness of our method, we compared its performance in 26 instances with the results obtained by three different algorithms: DBA, IBA and GSA-ACS-PSOT. We also performed a sensitivity analysis on the parameters. The results indicate that the improved bat algorithm outperforms all the other alternatives in most cases. • Multi-point shortest path planning problem is solved by using an improved discrete bat algorithm (IDBA). • An incomplete connected graph is first converted into a complete graph. • The bats' foraging and obstacle avoidance process are then simulated to find the shortest path in the complete graph. • The path is finally transferred to the original incomplete graph to get the solution. [ABSTRACT FROM AUTHOR]

Considering industrial process with high-dimensional, intrinsic nonlinearities and possibly abnormal observations, a robust deep learning soft sensor model is developed under the just-in-time learning framework. As an unsupervised deep learning approach, Variational Autoencoder (VAE) has been successfully applied to soft sensing problems owing to its ability to describe the latent representations by probability distributions. In this work, to construct high performance soft sensor model, mutual information (MI) is first introduced for input variable selection. By further incorporating MI as weights on variable of the traditional VAE model, a MI-based output-relevant VAE is developed. For each new sample that arrives, by utilizing Symmetric Kullback-Leibler (SKL) divergence, its relevance with historical samples is determined. Based on the SKL divergence, the input samples that are most relevant to the query sample can be collected. The selected historical input samples and corresponding output samples are employed to build a Gaussian process regression (GPR) local model. Expectation maximation (EM) algorithm is utilized to deal with the nonlinearity and abnormal output observations in GPR local model simultaneously for robustness of the soft sensors. Numerical simulations and a benchmark process are employed to validate the effectiveness of the proposed soft sensor, which demonstrates its superior performance over traditional approaches. • A MI-based output-relevant VAE is presented by incorporating MI as weights on variable of the traditional VAE model. • The selected historical samples through SKL divergence are employed to build a GPR local model. • The abnormal output observations in GPR local model for robustness of the soft sensors is dealt with the EM algorithm. • Numerical simulations and a benchmark process are employed to validate the effectiveness of the proposed soft sensor. [ABSTRACT FROM AUTHOR]