In this article, we develop an underwater image enhancement framework based on reinforcement learning. To do this, we model the underwater image enhancement as a Markov decision process (MDP), in which states are represented by image feature maps, actions are represented by image enhancement methods, and rewards are represented by image quality improvements. The MDP trained with reinforcement learning can characterize a sequence of enhanced results for an underwater image. At each step of the MDP, a state transitions from one to another according to an action of image enhancement selected by a deep Q network. The final enhanced image in the sequence is obtained with respect to the biggest overall image quality improvement. In this manner, our reinforcement learning framework effectively organizes a sequence of image enhancement methods in a principled manner. In contrast to the black box processing schemes of deep learning methods, our reinforcement learning framework gives a sequence of specific actions, which are transparent from the implementation perspective. Benefiting from the exploration and exploitation training fashion, our reinforcement learning framework possibly generates enhanced images that are of better quality than reference images. Experimental results validate the effectiveness of our reinforcement learning framework in underwater image enhancement. The code and detailed results are available at https://gitee.com/sunshixin_upc/underwater-image-enhancement-with-reinforcement-learning.

This article develops a general yet effective hyperspectral image (HSI) classification framework that is trained with small data. To this end, two identically structured but differently initialized classifiers, which are referred to as two base classifiers, are trained in an iterative manner. Each iteration consists of three steps, that is: 1) the two base classifiers that are trained separately on guide data; 2) unclassified data that are processed by the two trained base classifiers; and 3) the classification results with high confidence that are explored as new guide data. In the first iteration, the guide data comprising the original small training data are the same for the two base classifiers. From the second iteration, the guide data for the two base classifiers start becoming different. Specifically, in each iteration, the guide data for training one base classifier keep being augmented by high confidence classification results provided by the other base classifier. It is in such an iterative manner that the two classifiers continuously provide different new guide data for each other, and thus increasingly augment labeled data from the original small training set to a reasonably larger amount of samples in a HSI. We refer to such a training strategy as mutual guide. We develop a mutual guide implementation scheme by exploiting extreme learning machines (ELMs) as base classifiers. Extensive experiments on four public HSI datasets, i.e., Indian Pines (IP), Kennedy Space Center (KSC), University of Pavia (UP), and Salinas (SA), validate the classification effectiveness of our mutual guide framework with small training data.

Smart Grid is the trend of next generation power distribution and network management that enable a two -- way interactive communication and operation between consumers and suppliers, so as to achieve intelligent resource management and optimization. The wireless mesh network technology is a promising infrastructure solution to support these smart functionalities, while it has some inherent vulnerabilities and cyber-attack risks to be addressed. As Smart Grid is heavily relying on the underlie communication networks, which makes their security and dependability issues critical to the entire smart grid technology. Several studies have been conducted in the field of Smart Grid security, but few works were focused on the dependability and its associated resource analysis of the control center networks. In this paper, we have investigated the dependability modeling and also resource allocation in redundant communication networks by adopting two mathematical approaches, Reliability Block Diagrams (RBD) and Stochastic Petri Nets (SPNs), to analyze the dependability of control center networks in Smart Grid environment. We have applied our proposed modeling approach in an extensive case study to evaluate the availability of smart gird networks with different redundancy mechanisms. A combination of dependability models and reliability importance are used to analyze the network availability according to the most important components. We also show the variation of network availability in accordance with Mean Time to Failure (MTTF) in different network architectures.

Smart Grid is the trend of next generation electrical power system which makes the power grid intelligent and energy efficient. It requires high level of network reliability to support two-way communication among electrical services, electrical units such as smart meters, and applications. The wireless mesh network infrastructure can provide redundant routes for the smart grid communication network to ensure the network availability. Also due to its high level of flexibility and scalability features that make it become a promising solution for smart grid. However, wireless mesh network infrastructure are vulnerable to some cyber attacks which need to be addressed. In this paper, we proposed and implemented a new trust-based geographical routing protocol, named as Dynamic Trust Elective Geo Routing (DTEGR), which use the trust factor with a dynamic threshold value to generate a trust forwarding list, then it uses distance factor as routing metric to decide the next hop from the trust forwarding list. The simulation studies have confirmed our new DTEGR algorithm able to achieve better routing performance in different network scenarios, and also to achieve high level of reliable data transmission in smart grid communication network.

Smart Grid introduces modern information technologies such as the two-way digital communication networks, the automation control techniques and smart metering into the traditional power grid to provide efficient and intelligent electricity supply services to customers. The wireless mesh network (WMN) technology is a promising infrastructure to underpin and support these smart functionalities flexibly and scalably, and also provide redundant routes to enhance the network availability. However, WMN is vulnerable to some cyber-attacks. The importance of the underlie communication networks in Smart Grid can be analogized as the same criticality as the central nervous system such as brain and the spinal cord in human body, thus these security concerns become most critical and need to be addressed before WMN based Smart Grid becomes a reality. In this paper, we propose a new trust-based routing scheme, named as Trust-based Intelligent Geographical Energy-aware Routing (TIGER), to integrate trust, geographical information, and energy consumption considerations into a novel metric for making routing decision. The extensive simulation studies have confirmed that TIGER algorithms show promises for synthesizing trust, distance and energy factors to effectively detect and avoid malicious attacks in WMN based Smart Grid environment.