In this paper, we propose a saliency detection model for RGB-D images based on the deep features of RGB images and depth images within a Bayesian framework. By analysing 3D saliency in the case of RGB images and depth images, the class-conditional mutual information is computed for measuring the dependence of deep features extracted using a convolutional neural network; then, the posterior probability of the RGB-D saliency is formulated by applying Bayes’ theorem. By assuming that deep features are Gaussian distributions, a discriminative mixed-membership naive Bayes (DMNB) model is used to calculate the final saliency map. The Gaussian distribution parameters can be estimated in the DMNB model by using a variational inference-based expectation maximization algorithm. The experimental results on RGB-D images from the NLPR dataset and NJU-DS400 dataset show that the proposed model performs better than other existing models.
Recent developments in humidity sensors have heightened the need for reliability. Seeing as many products such as humidity sensors experience multiple dependent competing failure processes (MDCFPs) with self-recovery, this paper proposes a new general reliability model. Previous research into MDCFPs has primarily focused on the processes of degradation and random shocks, which are appropriate for most products. However, the existing reliability models for MDCFPs cannot fully characterize the failure processes of products such as humidity sensors with significant self-recovery, leading to an underestimation of reliability. In this paper, the effect of self-recovery on degradation was analyzed using a conditional probability. A reliability model for soft failure with self-recovery was obtained. Then, combined with the model of hard failure due to random shocks, a general reliability model with self-recovery was established. Finally, reliability tests of the humidity sensors were presented to verify the proposed reliability model. Reliability modeling for products subject to MDCFPs with considering self-recovery can provide a better understanding of the mechanism of failure and offer an alternative method to predict the reliability of products.
With the increase of the penetration rate of renewable energy power generation in the power system, the power grid gradually tends to be weak. Under the scenario of nonlinear load connecting to weak power grid, the grid-side voltage will produce unbalance, distortion and frequency deviation due to the influence of large internal impedance, insufficient inertia, three-phase current unbalance and harmonics of weak power grid. The operation of active power filters (APF) under above conditions can lead to significant degradation in harmonic compensation performance. An adaptive proportional integral (PI) + vector PI (VPI) harmonic compensation strategy for a three-phase APF is proposed for use in complex conditions of weak grid. The strategy consists of an adaptive notch filter based synchronous phase-locked loop (ANF-PLL) and an adaptive PI + VPI current control method. In this strategy, a series of adaptive lattice notch filters are introduced to improve the accuracy of grid phase and frequency estimation under the condition of voltage unbalance, distortion and a large range of frequency deviation. Then the phase and frequency information are used in the Park transformation of harmonic detection and current compensation control and the update of PI + VPI current control law parameters. A VPI current tracking controller with adaptive resonant frequency is constructed, which can improve the gain of the resonant frequency point and ensure that the APF can maintain the best compensation performance under the above weak grid conditions. Finally, the simulation and experimental results indicate that the ANF-PLL shows a better phase and frequency estimation performance under complex conditions of weak grid than the conventional and notch filter based methods, and the harmonic compensation strategy presented in this paper achieves a significant performance improvement under complex conditions of weak grid compared with Fixed VPI/SRF-PLL strategy and adaptive VPI/NF-PLL strategy.
Due to the complex operating environment of LED lamp beads (hereinafter referred to as LED), the test method that only considers the action of multiple stresses alone ignores the mutual influence between stresses, making it difficult to accurately obtain the life indicators of LEDs, resulting in a large deviation in the theoretical life results. In this regard, this paper proposes a multi-stress accelerated degradation evaluation method considering generalized coupling and conducts an accelerated degradation test (ADT) to evaluate the life of LEDs. We identified three stress sources and designed five new high-gradient ADTs. Through experimental data, we found that the three stress sources are strongly coupled on this LED. Then, a generalized coupling maximum likelihood estimation method (MLE) for the entire sample was constructed, and the particle swarm algorithm was used to solve the parameters. Finally, the life of this LED was evaluated based on the experimental data. The results show that the life of the LED considering multiple-stress coupling is within 6.5% of the historical life scatter point, which is more in line with the actual working environment.
1 September 2020 in Journal of Quantitative Spectroscopy and Radiative Transfer