Search

Benefiting from the abrupt phase changes within subwavelength thicknesses, metasurfaces have been widely applied for lightweight and compact optical systems. Simultaneous broadband and high-efficiency characteristics are highly attractive for the practical implementation of metasurfaces. However, current metasurface devices mostly adopt discrete micro/nano structures, which rarely realize both merits simultaneously. In this paper, dielectric metasurfaces composed of quasi-continuous nanostrips are proposed to overcome this limitation. Via quasi-continuous nanostrips metasurface, a normal focusing metalens and a superoscillatory lens overcoming the diffraction limit are designed and experimentally demonstrated. The quasi-continuous metadevices can operate in a broadband wavelength ranging from 450 nm to 1000 nm and keep a high power efficiency. The average efficiency of the fabricated metalens reaches 54.24%, showing a significant improvement compared to the previously reported metalenses with the same thickness. The proposed methodology can be easily extended to design other metadevices with the advantages of broadband and high-efficiency in practical optical systems.

For the purpose of detecting waterborne bacteria, a high-phase-sensitivity SPR sensor with an Ag–TiO2–Franckeite–WS2 hybrid structure is designed using an improved seeker optimization algorithm (ISOA). By optimizing each layer of sensor construction simultaneously, the ISOA guarantees a minimum reflectance of less than 0.01 by Ag (20.36 nm)–TiO2 (6.08 nm)–Franckeite (monolayer)–WS2 (bilayer) after 30 iterations for E. coli. And the optimal phase sensitivity is 2.378 × 106 deg/RIU. Sensor performance and computing efficiency have been greatly enhanced using the ISOA in comparison to the traditional layer-by-layer technique and the SOA method. This will enable sensors to detect a wider range of bacteria with more efficacy. As a result, the ISOA-based design idea could provide SPR biosensors with new applications in environmental monitoring.

Summary: Chemiresistive gas sensors generally surfer from low selectivity, inferior anti-humidity, low response signal or signal-to-noise ratio, severely limiting the precise detection of chemical agents. Herein, we exploit high-performance gas sensors based on topological insulator Bi2Se3 that is distinguished from conventional materials by robust metallic surface states protected by time-reversal symmetry. In the presence of Se vacancies, Bi2Se3 nanosheets exhibit excellent gas sensing capability toward NO2, with a high response of 93% for 50 ppm and an ultralow theoretical limit of detection concentration about 0.06 ppb at room temperature. Remarkably, Bi2Se3 demonstrates ultrahigh anti-humidity interference characteristics, as the response with standard deviation of only 3.63% can be achieved in relative humidity range of 0–80%. These findings are supported by first-principles calculations, with analyses on adsorption energy and charge transfer directly revealing the anti-humidity and selectivity. This work may pave the way for implementation of exotic quantum states for intelligent applications.

A differential evolution particle swarm optimization (DEPSO) is presented for the design of a high-phase-sensitivity surface plasmon resonance (SPR) gas sensor. The gas sensor is based on a bilayer metal film with a hybrid structure of blue phosphorene (BlueP)/transition metal dichalcogenides (TMDCs) and MXene. Initially, a Ag-BlueP/TMDCs-Ag-MXene heterostructure is designed, and its performance is compared with that of the conventional layer-by-layer method and particle swarm optimization (PSO). The results indicate that optimizing the thickness of the layers in the gas sensor promotes phase sensitivity. Specifically, the phase sensitivity of the DEPSO is significantly higher than that of the PSO and the conventional method, while maintaining a lower reflectivity. The maximum phase sensitivity achieved is 1.866 × 106 deg/RIU with three layers of BlueP/WS2 and a monolayer of MXene. The distribution of the electric field is also illustrated, demonstrating that the optimized configuration allows for better detection of various gases. Due to its highly sensitive characteristics, the proposed design method based on the DEPSO can be applied to SPR gas sensors for environmental monitoring.

The machine vision system was employed to inspect the surface defects of bridge cables of cable-stayed bridges. After the acquisition and preprocessing of the defect images, it is necessary to classify and identify the defects of the cables to meet the requirements of non-destructive testing and evaluation. In this paper, feature extraction for defect images was performed using mathematical statistical methods. After that, 10 feature parameters including shape features, grayscale features and texture features of the defect images were obtained and selected for a classification model of support vector machine (SVM). To improve the SVM classification performance, the particle swarm optimization algorithm (PSO) was adopted to obtain the punish factor c and the kernel parameter g of the SVM model, namely the PSO-SVM algorithm. Finally, our PSO-SVM classification model was employed to implement the classification of real surface defect images of the bridge cables. Longitudinal crack, transverse crack, surface corrosion, and pothole defect can be automatically identified and the classification accuracy reached 96.25%. The experimental results showed that the PSO-SVM model can improve the classification performance of the surface defects. Based on the effective classification, we can find the distribution characteristics of the surface defects of the cable. It is very important to analyze the relationship between the type of surface defects and the material of the protective layer, so as to adopt appropriate materials and reasonable maintenance measures. Thus, it has a great significance in the structural health monitoring of bridge cables.

Deep neural networks have received increasing attention in human action recognition. Previous research has established that utilizing 3D convolution is a reasonable approach to learn spatio-temporal representation. Nevertheless, constructing effective 3D ConvNets usually need an expensive pre-training process that performing on a huge-scale video dataset. To avoid this burdensome situation, one major issue is to determine whether the pre-trained parameters of 2D convolution networks can be directly bootstrapped into 3D. In this paper, we devise a 2D-Inflated operation and a parallel 3D ConvNet architecture to solve this problem. The 2D-Inflated operation is used for converting pre-trained 2D ConvNets into 3D ConvNets, which avoiding video data pre-training. We further explore the optimal quantity of 3D ConvNet in the parallel architecture, and the results suggest that 6-nets architecture is an excellent solution for recognition. Another contribution of our study is two practical and valid skills, accumulated gradient descent and video sequence decomposition. Either of those techniques can promote the improvement of performance. The recognition results of UCF101 and HMDB51 reveal that, without the video data pre-training, our 3D ConvNets still can achieve competitive performance to the other generic and recent methods of using 3D ConvNets in the RGB image domain.

Automobile exhaust gases, such as carbon dioxide (CO2), carbon monoxide (CO), and propane (C3H8), cause the greenhouse effect, photochemical smog, and haze, threatening the urban atmosphere and human health. In this study, a non-dispersive infrared (NDIR) multi-gas detection system consisting of a single broadband light source, gas cell, and four-channel pyroelectric detector was developed. The system can be used to economically detect gas concentration in the range of 0–5000 ppm for C3H8, 0–14% for CO, and 0–20% for CO2. According to the experimental data, the concentration inversion model was established using the least squares between the voltage ratio and the concentration. Additionally, the interference coefficient between different gases was tested. Therefore, the interference models between the three gases were established by the least square method. The concentration inversion model was experimentally verified, and it was observed that the full-scale error of the sensor changed less than 3.5%, the detection repeatability error was lower than 4.5%, and the detection stability was less than 2.7%. Therefore, the detection system is economical and energy efficient and it is a promising method for the analysis of automobile exhaust gases.