The detection of manmade disasters particularly fire is valuable because it causes many damages in terms of human lives. Research on fire detection using wireless sensor network and video-based methods is a very hot research topic. However, the WSN based detection model need fire happens and a lot of smoke and fire for detection. Similarly, video-based models also have some drawbacks because conventional algorithms need feature vectors and high rule-based models for detection. In this paper, we proposed a fire detection method which is based on powerful machine learning and deep learning algorithms. We used both sensors data as well as images data for fire prevention. Our proposed model has three main deep neural networks i.e. a hybrid model which consists of Adaboost and many MLP neural networks, Adaboost-LBP model and finally convolutional neural network. We used Adaboost-MLP model to predict the fire. After the prediction, we proposed two neural networks i.e. Adaboost-LBP model and convolutional neural network for detection of fire using the videos and images taken from the cameras installed for the surveillance. Adaboost-LBP model is to generate the ROIs from the image where emergencies exist Our proposed model results are quite good, and the accuracy is almost 99%. The false alarming rate is very low and can be reduced more using further training.
250 consecutive female contacts of men with gonorrhoea were examined in Lydia Department of Genito-urinary Medicine, St Thomas' Hospital, London, and 165 (66 per cent.) were found to be positive for gonorrhoea. Over 99 per cent. of the positive results were found by the first two sets of genital investigations. These results are compared with earlier and more recent results from our own and other centres and it is postulated that the higher early diagnosis rate is due to the introduction of a highly selective but non-inhibitory culture medium (VCNT), combined with a new incubator which accurately regulates temperature, humidity, and CO2 content. Reasons for the negative findings in 35 per cent. of the female gonorrhoea contacts are discussed. 相似文献
The Dynamic Cone Penetrometer (DCP) is a device that is used in the construction industry for the assessment of in situ soil compaction quality. Over the past few decades, numerous correlations have been developed between the DCP test results and soil strength and stiffness parameters. This paper proposes a comprehensive set of criteria and recommendations for quality control (QC) of compacted subgrade that take into account the inherent statistical variability of DCP test results. For the development of the QC criteria, a new statistical methodology is used to extract representative test values from the raw field DCP test data. In order to use the proposed QC criteria, soils are first classified into two broad categories (fine-grained and coarse-grained) depending on their fabric and response to compaction efforts. Test results indicate that (i) for fine-grained soils, the DCP test values have good correlation with the plasticity index (PI), which is indicative of the type and amount of clay content of the soil and (ii) for coarse-grained soils, the DCP test values have good correlation with the optimum water content of the soil, which is directly related to its in situ density. DCP blow count correlation equations are presented for both soil categories. Recommendations for field DCP testing and data analysis are also provided to highlight the significance of the statistical distribution of the DCP test results in QC testing of compacted subgrade. 相似文献
Component-based software development is rapidly introducing numerous new paradigms and possibilities to deliver highly customized software in a distributed environment. Among other communication, teamwork, and coordination problems in global software development, the detection of faults is seen as the key challenge. Thus, there is a need to ensure the reliability of component-based applications requirements. Distributed device detection faults applied to tracked components from various sources and failed to keep track of all the large number of components from different locations. In this study, we propose an approach for fault detection from component-based systems requirements using the fuzzy logic approach and historical information during acceptance testing. This approach identified error-prone components selection for test case extraction and for prioritization of test cases to validate components in acceptance testing. For the evaluation, we used empirical study, and results depicted that the proposed approach significantly outperforms in component selection and acceptance testing. The comparison to the conventional procedures, i.e., requirement criteria, and communication coverage criteria without irrelevancy and redundancy successfully outperform other procedures. Consequently, the F-measures of the proposed approach define the accurate selection of components, and faults identification increases in components using the proposed approach were higher (i.e., more than 80 percent) than requirement criteria, and code coverage criteria procedures (i.e., less than 80 percent), respectively. Similarly, the rate of fault detection in the proposed approach increases, i.e., 92.80 compared to existing methods i.e., less than 80 percent. The proposed approach will provide a comprehensive guideline and roadmap for practitioners and researchers. 相似文献
In the past few years, research and development in Wireless Sensor networks (WSNs) have gained momentum due to its numerous applications in agriculture, industrial manufacturing, military surveillance, environmental monitoring, consumer electronics, medical & healthcare, disaster recovery operations etc. Dynamic WSNs offer a robust blend of distributed sensing, computing and communication. Dynamic sensor networks are characterized by large scale deployment, dynamic and unstructured topology, power limitations, less memory and limited computational capabilities. Sensor nodes deployed in real-time environment’s for sensing data have power-limitations which hampers the overall performance of WSNs. So, the only obvious solution is to propose an energy efficient routing protocol to optimize WSN real-time performance. Different specialists have proposed various directing conventions for WSNs dependent on Fuzzy Logic, Genetic Algorithms, Meta-Heuristics, and other improvement strategies. However, every solution suggested till date has its advantages and limitations. In this paper, our primary objective is to utilize Swarm-Intelligence based approach i.e. “Ant Colony Optimization (ACO)”, for routing protocol development. Ant colony optimization (ACO) based approach gives optimal solution in terms of efficient routing path determination, energy efficiency and delivering high performance in terms of packet delivery and throughput. In this paper, we propose a novel energy efficient ACO based multipath routing protocol for WSN i.e. IEEMARP (Improvised Energy Efficient Multipath ACO based Routing Protocol). The proposed protocol works in three phases (Neighbor Discovery via Link Knowledge, Packet Transmission via exponentially weighted moving average method and ACKR packet delivery for assuring end-to-end delivery. To validate the performance of the protocol proposed, extensive simulations were conducted using NS-2.35-allinone simulator on diverse parameters like (PDR), throughput, routing overhead, energy consumption and end-to-end delay. In addition to this, the performance of protocol is compared with traditional routing protocols like Basic ACO, DSDV and DSR and other ACO based WSN protocols like ACEAMR, AntChain, EMCBR, IACR, AntHQSeN, FACOR and ANTALG. Simulation based results, clearly states that as compared to Basic ACO, DSDV and DSR, the performance of WSN network is improvised to around 10% in all performance metrics via IEEMARP routing protocol. And as compared to ACEAMR, AntChain, EMCBR and IACR, IEEMARP performs 20% better in overall functionality and almost 10–12% better as compared to AntHQSeN, FACOR, ANTLAG routing protocols in varied WSN scenarios. It is also observed that IEEMARP protocol is highly efficient in TCP packet transmission from source to destination node.
To the best of our knowledge, the tool of soft set theory is a new efficacious technique to dispose uncertainties and it focuses on the parameterization, while fuzzy set theory emphasizes the truth degree and rough set theory as another tool to handle uncertainties, it places emphasis on granular. However, the real-world problems that under considerations are usual very complicated. Consequently, it is very difficult to solve them by a single mathematical tool. It is worth noting that decision making (briefly, DM) in an imprecise environment has been showing more and more role in real-world applications. Researches on the idiographic applications of the above three uncertain theories as well as their hybrid models in DM have attracted many researchers’ widespread interest. DM methods are not yet proposed based on fusions of the above three uncertain theories. In view of the reason, by compromising the above three uncertain theories, we elaborate some reviews to DM methods based on two classes of hybrid soft models: SRF-sets and SFR-sets. We test all algorithms for DM and computation time on data sets produced by soft sets and FS-sets. The numerical experimentation programs are written for given pseudo codes in MATLAB. At the same time, the comparisons of all algorithms are given. Finally, we expatiate on an overview of techniques based on the involved hybrid soft set models. 相似文献