Abstract:In this paper, the influence of annealing temperature on the structure and optical properties of silicon films was systemically investigated. Silicon films were deposited by electron beam evaporation and then annealed in N₂ atmosphere within a temperature range from 200 to 500 °C. The films were characterized by X-ray diffraction （XRD）, Raman spectroscopy, electronic-spin resonance （ESR） and optical transmittance measurement, respectively. With annealing temperature increased, the amorphous network order of silicon films was improved on the short and medium range and the defect density decreased remarkably. When sample being annealed at 400°C, the extinction coefficient k decreased from 6.14×10^-3 to a minimum value of 1.02×10^-3 （at 1000 nm）, which was due to the lowest defect density, about one fifth of the as-deposited sample. The results showed that annealing at an appropriate temperature could effectively reduce the optical absorption of silicon films in the near infrared region, which were very critical for the application in optical thin film coating devices.

Abstract:HgCdTe avalanche photodiodes （APDs） of linear mode have attracted much attention for their prospective applications, due to the lowest excess noise close to zero, operating bias in the range of -12~0V, and linear gain tunable as well. In this paper, we investigated the dark current mechanism and Okuto-Crowell model based on the PIN structure of mid-wave infrared HgCdTe e-APD. The characteristics of dark current and the gain were simulated. The thickness and carriers concentration of intrinsic region were discussed, which would impact on the dark current and the gain. The results indicated that the peak intensity of electric field would significantly increase the band to band generation, even up to several orders of magnitude. BBT current could be suppressed by increasing the thickness of the intrinsic region, as well as reducing the carrier concentration. Unfortunately, the gain and BBT current would both increase with electric field. The gain normalized dark current （GNDC） is prospective by means of tradeoff between the dark current and the gain. In our studies, intrinsic region with 3 μm and the concentration less than 5×10¹⁴cm^-3 are optimal. The APD I-V results demonstrated that the theoretical model was identical with the experimental data.

Abstract:The photonic artificial intelligent chip performs calculations at the speed of light, and has the advantages of low power consumption, low delay, and anti-electromagnetic interference. Miniaturization and integration are the key steps to realize this technological innovation. In this paper, lithography is applied to the fabrication of diffraction gratings. A design and implementation method of all-optics diffraction deep learning neural network grating based on 10.6 micron laser is proposed. Since the wavelength of the light source evolved from the millimeter wave to micrometer wave, the characteristic scale of the neuron are reduced to 20 micrometers. Compared with the existing optical computing neural network, the feature size of the deep learning neural network is reduced by 80 times, which laid the foundation for further large-scale integration of photonic computing chips.

Abstract:A high-refractive-index contrast subwavelength grating （HCG） for 850 nm GaAs-based VCSELs was designed. The whole structure is based on GaAs material system，including a grating layer， a stress buffer layer designed to alleviate its stress problem and an AlO_x low refractive index sub-layer oxidized by AlGaAs or AlAs. The reflection characteristics of the HCG are simulated by Rsoft software， and the effects of different grating parameters on reflection spectrum are analyzed. In particular， the effects of stress buffer layer and low refractive index sub-layer on characteristics of gratings are investigated. The 850 nm TM mode HCG shows a very good reflection characteristic. It has a large reflection bandwidth of up to 91 nm with its reflectivity more than 99.9%， and the ratio with the center wavelength reaches 10.7%. Moreover， its reflectivity for TE mode is ensured to be lower than 90%， showing good polarization selectivity. This HCG can replace P-type distributed Bragg reflectors in VCSEL， providing high reflectivity， wide bandwidth， and good stability.

Abstract:A dual-comb system based on phase locking to ultra-narrow linewidth lasers is studied and implemented in this paper. Two Erbium-doped mode-locked lasers with repetition rates of about 200 MHz and repetition rate difference of about 17 kHz were simultaneously locked to two narrow-linewidth continuous wave lasers at 1 542 nm and 1 560 nm, respectively. Hence, dual-comb with high coherence and stability is achieved. After phase locking, the comb linewidth of a single comb is less than 5 Hz（limited by test condition）, and the relative linewidth of two combs is less than 0.35 Hz. The dual comb system has important application value in high precision spectral analysis, time-frequency transmission, and especially absolute distance measurement.

Abstract:A hollow-core anti-resonant fiber is proposed with low confinement loss in infrared band. The cladding tube is a semicircular semi-elliptical splicing structure. The full vector finite element method is used for the simulation. The semi-minor axis of the semi-elliptical tube is equal to the radius of the semicircular tube. A semi-circular tube and a semi-elliptical tube are spliced to form a semi-circular semi-elliptical cladding tube. The negative curvature and the positive curvature are changed by improving the structure parameters of the tube. The purpose of this paper is to study the loss characteristics of the positive and negative curvature of the cladding tube, and to design a low loss hollow core anti-resonant fiber applied in the 1.5~3.0 μm band. The results show that the effect of limiting loss is better when the negative curvature decreases and the positive curvature increases. The lowest confinement loss is 8.22×10^-2 dB/km at the wavelength of 2.1 μm when the semi-circular tube is close to the core and the elliptical tube is far away from the core, as well as the circular radius r_y=25 μm, the semi-major axis of the ellipse r_x=65 μm, the semi-minor axis of the ellipse r_y=25 μm.

Abstract:We propose a beam reconfigurable micro-strip quasi-Yagi-Uda （QYU） antenna operating at terahertz （THz） frequencies. The antenna, on a metal-backed SiO₂ substrate, consists of a metallic micro-strip transmission line, a metallic reflector, a metallic half-round micro patch, and three groups of monolayer-graphene-patch directors. Because the conductivities of the graphene patches can be electrically tuned, the radiation direction of the QYU antenna can be dynamically controlled by adjusting the bias voltages applied to the graphene patches. The basic properties and the tunable behaviors of the antenna are systemically simulated and optimized. It has very fast modulation speed and very low return loss. In particular, the numerical results show that the main radiation lobe angle φ （azimuth angle） of the QYU antenna can be swept from 30° to 150° in θ=60° （zenith angle） plane with different configurations of bias voltages applied to the graphene-patch directors. The proposed antenna is very suitable for THz beam reconfigurable applications, such as phased array radar systems.

Abstract:The use of microwave links in wireless networks to monitor rainfall and water vapor is the latest technology in atmospheric environmental monitoring. This technology can measure meteorological elements such as rainfall and water vapor near the surface and has the advantages of high spatial and temporal resolution and low cost. The water vapor density from June 13, 2017 to July 13, 2017 is calculated and analyzed by using the data of the microwave communication link in the E-band of Gothenburg provided by Ericsson, the data of meteorological station 1 provided by the meteorological station at one end of the link and the data of meteorological station 2 provided by the meteorological website of the Swedish Institute of Meteorology and Hydrology （SMHI）. The results show that the environmental conditions at different locations in the same area have certain differences, the temperature of the same area shows a variation , the correlation between the two is 0.87; We retrieved water vapour density value from the microwave communication link and compared with the measurement from ground weather station in the study area, and we found the results have good consistency. The correlation between the two is 0.89, and the root mean square error is 0.75. We have demonstrated that microwave links can be utilized to provide an extra rich data source for existing weather monitoring networks in addition to existing humidity monitoring methods.

Abstract:A 3.0 THz detector based on plasma-wave theory proposed by Dyakonov and Shur was designed and fabricated in 65 nm standard CMOS process, the detector consists of a patch antenna, a NMOS field effect transistor, a matching network, and a notch filter, it can achieve a room-temperature responsivity （Rv） of 526 V/W and a noise equivalent power （NEP） of 73 pW/Hz^1/2. The terahertz scanning imaging system was built with the detector and stepper motor, and the far-field shape of the terahertz source beam was obtained, the full width at half maximum （FWHM） of the beam is 240 μm; and the image of the polyformaldehyde toothpick and tree leaf were obtained through the scanning imaging system, it shows that CMOS terahertz detectors have potential applications in the imaging field.

Abstract:As an important thermal detection material, Mn-Co-Ni-O has been widely used in variety of fields as the key thermal detector. NiMn₂O₄ and Mn_1.56Co_0.96Ni_0.48O₄ bulk materials were prepared by grinding and wintering methods. The crystallization of the two blocks was studied by X-ray diffraction experiments, and it was found that the addition of Co cation with larger radii resulted in deterioration of the bulk crystalline. The optical constants and dielectric constants of NiMn₂O₄ and Mn_1.56Co_0.96Ni_0.48O₄ in the ultraviolet-far infrared wide band were obtained by ellipsometry. It was found that significant difference appeared between the optical properties (intensity and peak position of optical constants) of the two materials after the addition of ions. The reflectance spectra of the two materials were obtained by Fourier spectrometer, it was compared with the calculated values of optical constants, so the effect of surface roughness on reflection spectra was evaluated.

Abstract:Wood is usually stored outdoors so that when its hyper-spectral image is picked up, the acquired image is usually disturbed by environmental factors such as illumination, temperature, and humidity. This disturbance may produce the false wood species classification results. To solve this issue, the wood texture feature is extracted in its hyper-spectral image by use of PLS and LBP. This texture feature is then combined with the near infrared spectra of wood hyper-spectral image so that the fused features are sent into SVM and BP neural network classifiers. Experimental results indicate that our scheme can reach to 100% classification accuracy without environmental disturbance. Moreover, to testify our scheme’s robustness in case of illumination variation, a simulation experiment is performed and it indicates that our scheme outperforms the conventional and the state-of-art wood recognition schemes.

Abstract:In the autonomous driving system of automobile, in order to improve the performance of single-wavelength LIDAR in physical property detection classification and state, and draw lessons from the principle that multi-spectral detection has physical property detection ability, this paper studies the band selection of multi-spectral LIDAR, calculates and analyses the spectrum of typical targets in autonomous driving by using principal component analysis method. The characteristics of laser source and detector, the band selection method of multi-spectral LIDAR, the spectral characteristic analysis of typical targets for autonomous driving application scenarios and the availability of commercial LIDAR are synthesized. The central wavelength of the multi-spectral LIDAR suitable for autonomous driving of automobiles is 808 nm, 905 nm, 1 064 nm and 1 310 nm. The validity of the selected wavelength of the multi-spectral LIDAR is verified by testing.

Abstract:The bandedge electronic structure including the optical bandgap, band-tail states, and deep/shallow donor and acceptor levels in Cu₂ZnSnS₄ semiconductor was analyzed by absorption, photocurrent and photoluminescence spectroscopy, and the theoretical reports. It is revealed that the -related defect in Cu₂ZnSnS₄ with abundant defect states is one of the key factors affecting the band-edge electronic structure. High concentration of the neutral defect cluster [] can narrow the band gap substantially, while the partially-passivated （ionic） defect cluster [] is the main deep donor defect. A large number of band-tail states are responsible for the obvious red-shift of the bandedge-related photoluminescence transition energy. These detrimental defects related to can be effectively suppressed by properly reducing the Sn content in the copper-poor and zinc-rich growth condition, which also avoids the narrowing of the optical bandgap of the Cu₂ZnSnS₄ absorption layer.

Abstract:Hierarchical spectrum recognition is pointed out in the aspect of hyper-spectralimage classification using Hyperion imagesof Lanzhou and Yulin in China， by the comparison study of SVM supervised classification.There are 4 problems for hyper-spectralimage classification: spectral information fidelity image fusion， extraction of sensitive band， removal of "salt and pepper effect" which avoids misclassification for "same body with different spectrum" phenomenon. Hierarchical spectrum recognitionputs forward 4 methods for solving these problems，which are WP-GS image fusion， derivative transformation， object oriented segmentation with 4 scales and multispectral SAM. Hierarchical spectrum recognitioncan identify 9 kinds land type exactly， based on sensitive bands extracted from derivative transformationimage. Visual examination and quantifiable evaluation have been executed to verify authenticity. By contrast， SVM supervised classification is tested， with Gram-Schmidt Spectral sharpening/ Savitzky-Golay convolution filtering/PCA transformation.Using the method of hierarchical spectrum recognition proposed in this paper， the above 4 problems for hyper-spectralimage classification are solved， and good classification results is achieved， with overall classification accuracy and kappa coefficient as 89.52%， 0.852 in Lanzhou， 91.12%、0.873 in Yulin. Growth of 18.68% in Lanzhou and 17.80% in Yulin for overall classification accuracy， as well as17.52% in Lanzhou and 16.89% in Yulin for kappa coefficient are achieved compared with SVM supervised classification.Hierarchical spectrum recognition can provide more accurate recognition results are provided for Hyperion images classification in comparison with SVM supervised classification .

Abstract:Regional leaf area index （LAI） mapping is important for crop growth monitoring and yield estimation. Due to the lower accuracy and instability of statistical models for regional LAI estimation, we proposed a new deep neural network model, i.e. Small Simple Learning LAI-Net （SSLLAI-Net）, based on small sample training, to achieve stable relationship between hyperspectral reflectance and LAI. The new proposed SSLLAI-Net was constructed with two convolution layers, one pooling layer and three connect layers, for which the inputs and outputs were hyperspectral reflectance and LAI estimation. Moreover, SSLLAI-Net could support small training sets. We applied SSLLAI-Net to an Environmental Mapping and Analysis Program （EnMAP） hyperspectral imagery for regional LAI mapping, in which cereals, maize, rape seed and other crops are selected as our objects. The achieved R² values for estimated LAI of cereals, maize, rape seed and other crops were 0.95, 0.99, 0.98 and 0.90 based on small training sets with 50 samples, while for the inputs with noise, the R² values were 0.95、0.98、0.96 and 0.89, respectively. In all, our new proposed SSLLAI-Net has high precision of regional LAI mapping, stability and noise resistance with hyperspectral remote sensing observations.

Abstract:FY-3D/MERSI-II polar orbiting meteorological satellite has 250 m resolution in 10.8 and 12 μm wavelengths of far-infrared channels, which provide more abundant data sources for fire monitoring. The characteristics of fire monitoring in FY-3D/MERSI-II 10.8 μm far-infrared channel were studied. Although 10.8μm far-infrared channel is far less sensitive than FY-3D/MERSI-II"s 3.8 μm mid-infrared channel in the spectral aspect to high-temperature heat source detection, its spatial resolution is 4 times higher than that of 1 km mid-infrared channel, so it can clearly reflect larger fire points, and its fire detection ability is better than that of 1 km resolution far-infrared channel. There has been a marked improvement. Using mixed pixel linear spectral separation method, the brightness temperature increment caused by the open fire area with an average temperature of 750 K and an area of 400 m² in far infrared channel pixels at 1 km resolution is about 0.47 K, while that caused by the same fire condition in far infrared channel pixels at 250 m resolution is about 7.30 K, which is quite different from the brightness temperature of surrounding background pixels. The fire location accuracy can be improved from kilometer level to hundreds meter level by using 1 km resolution mid-infrared channel to identify the fire range and 250 m resolution far-infrared channel to further determine the open fire area. Based on the method proposed in this paper, an example of forest and grassland fire application is analyzed. The location of fire point determined by 250 m resolution far-infrared channel is in good agreement with the field investigation information, which shows the effectiveness of the proposed method. Several application examples showed that the FY-3D/MERSI far-infrared channel has obvious advantages over the mid-infrared channel in reflecting the location of the intense fire area and the distribution of the open fire line of grassland fire in large-scale fire field. It can reflect the spatial distribution of wildfire more accurately, so it has application value in fire monitoring.

Abstract:Supervised classification is a vital approach to extract impervious surface areas (ISA) from satellite images, but the training samples need to be provided through heavy manual work. To address it, this study proposed an automatic method to generate training samples from high-resolution night light data, considering that nighttime lights generated by human activities is strongly correlated with impervious surface. First, positive and negative samples for ISA were located according to the distribution of nighttime lights. Second, the feature sets were constructed by calculating the spectral and texture feature from the OLI images. Third, an ensemble ELM classifier was selected for ISA classification and extraction. Four large cities were selected as study areas to examine the performance of the proposed method in different environment. The results show that the proposed method can automatically and accurately acquire ISA with an overall accuracy higher than 93% and Kappa coefficient higher than 0.87. Furthermore, comparative experiments by biophysical composition index (BCI)and classification by manual sample were conducted to evaluate its superiority. The results show that our method has better separability for ISA and soil than the BCI. In general, the proposed method is superior to manual methods, except Harbin mostly because some impervious surfaces with weak light intensity are selected as negative samples.