Video dehazing aims to restore clear, haze-free frames from hazy videos while maintaining temporal continuity. However, existing deep learning methods for video dehazing do not take into account the gap between real and synthesized data. To address this issue, we propose a new method for achieving superior dehazing performance in real scenes. Specially, we construct a synthesizing hazy video datasets, which simulates the atmospheric light of realistic hazy videos. Additionally, we propose a optical flow deformable fusion module that utilizes optical flow to guide frame alignment and employs neighboring frame similarity for frame fusion. Furthermore, a contrastive learning loss function helps recovered frames closer to clear frames in feature space. Experimental results demonstrate the effectiveness of our proposed method, outperforming state-of-the-art methods in real hazy videos.

Adverse weather conditions can significantly hinder the performance of deep learning-based object detection models. Traditional approaches often rely on image restoration techniques to enhance the quality of degraded images prior to detection. However, these methods frequently struggle to balance image enhancement and detection tasks effectively, often overlooking latent information that could be beneficial for detection. To address these challenges, we propose a novel framework: Knowledge Distillation based on Diffusion Models (KDDM). This framework incorporates a Dehaze Network (DN), which employs large kernel convolution to remove weather-specific artifacts, thereby revealing more latent information. The DN, together with a text detector, forms an end-to-end scene text detection network, acting as the student network. Additionally, the nuanced internal representations of text-to-image diffusion models adeptly capture and integrate higher-order visual semantic concepts. Given the rich textual and visual content inherent in scene text, there is a fundamental connection to text-to-image diffusion models. As such, we utilize diffusion models as a teacher network to distill high-level visual semantic knowledge into the student network. Notably, we introduce an innovative distillation technique using a “Threshold_Mask”, which ensures that the student network focuses on text regions while minimizing interference from irrelevant background elements. Comprehensive experimental evaluations demonstrate that our KDDM framework significantly outperforms baseline models under foggy weather conditions, marking a substantial advancement in the field.

This article presents a resilience assessment approach for hybrid ac/dc cyber-physical power system (CPPS), proposing a comprehensive assessment index called cascading failure recovery index (CFRI) that simultaneously considers the system scale and load level in the cascading failure recovery process. First, correlation characteristic matrix-based modeling framework is developed to capture the characteristics of multidimensional heterogeneous power systems, providing a clear description of the cyber-physical coupling network. Besides, the proposed CFRI incorporates cyber-physical coordinated attacks to assess the robustness of hybrid ac/dc power systems under different attack scenarios. The CFRI takes into account the number of nodes, branches, and load levels, enabling an accurate assessment of the disconnection degree and recovery capability of CPPS in case of cascading failures. Finally, simulation studies are conducted on a IEEE 39-bus power system modified with dc transmission lines to validate the effectiveness of the proposed method.

The industrial internet of things (IIoT) is emerging as a global trend to dramatically enhance the intelligence and efficiency of the industries in recent years. With the emphasis on data communication by IIoT, cyber vulnerabilities are introduced at the same time. As a key subsystem of the industrial automation systems, the supervisory control and data acquisition (SCADA) system is becoming one of the primary targets for cyberattacks in the IIoT paradigm. In this paper, the semi-Markov process (SMP) is employed to model and evaluate the cyberattacks against the SCADA systems considering the insider assistance. Based on the SMP model, the probability distribution of the time-to-compromise the system of the attacks is derived with the Monte Carlo simulation (MCS). Then, a FlipIt game model is developed to investigate the defense and attack strategies of the defender and attacker, and analyze the impacts of the insider assistance. Case studies were carried out to verify the proposed model. The results of the case studies show that the insider assistance will improve the payoff of the attacker and increase the defense action frequency of the system defender. With a high enough defense action frequency, the defender can force the attacker to drop out and eliminate the attack actions.

As one of the national critical infrastructures, the water distribution system supports our daily life and economic growth, the failure of which may lead to catastrophic results. Besides the uncertainty from the system component failures, cyberattacks are vital to the secure system operation and have great impacts on the reliability of the water supply service. Malicious attackers may intrude into the supervisory control and data acquisition (SCADA) system of pump stations in the water distribution networks and interrupt the water supply to the customers. Cyber insurance is emerging as a promising financial tool in system risk management. In this paper, cyber insurance is proposed for the cyber risk management of the water distribution system. A semi-Markov process (SMP) model is devised to model the cyberattacks against pump stations in the water distribution system. Both the impacts of the independent cyber risks in the individual distribution network and the correlated cyber risks shared across different water distribution networks are evaluated and modeled. A sequential Monte Carlo Simulation (MCS) based algorithm is developed to evaluate the system loss. Cyber insurance premiums for the water distribution networks are designed based on the actuarial principles and potential system losses. Case studies are also performed on multiple representative water distribution networks, and the results demonstrate the validity of the proposed cyber insurance model.