Near-infrared (NIR) polymer photodetectors possess flexible and adjustable photoelectric characteristics, have good compatibility with flexible substrates, require a simple preparation process, and are inexpensive. They also have significant application prospects in aviation, military, industrial, and medical fields. NIR polymer photodetector structures include photoconductors, photodiodes, and phototransistors. This study reviews the research progress of NIR polymer photodiodes (NIR PPDs). First, the photoelectric conversion principle of NIR PPDs is introduced. Second, important advances made by researchers in improving the performance of NIR PPDs from the perspectives of new material development and device structure design are discussed. Finally, a summary is presented along with possible challenges and prospects of current research on NIR PPDs.

Long-distance oil and gas pipelines are widely distributed and have complex background environments. Therefore, their optical-fiber pre-warning system experiences a high false-alarm rate in identifying destructive events that threaten pipeline safety in a real-world environment. This makes it challenging for the system to achieve accurate pre-warning results and ensure pipeline safety. This study applies deep learning to a long-distance fiber pre-warning system. Through deep learning, a vehicle-passing signal that mainly affects the pre-warning effect is identified, which effectively reduces the false-alarm rate of the pre-warning system. The intelligent fiber pre-warning system is mainly divided into two parts: the distributed optical-fiber sensing system and the signal-recognition system. In a real-world environment, an intrusion signal around the pipeline is collected by a Φ-OTDR(phase-sensitive optical time domain reflectometry) distributed optical-fiber sensing system. Additionally, a recognition model is established by convolutional long short-term memory and fully connected deep neural networks to detect the vehicle-passing signal. After training and blind testing, the vehicle-passing event recognition model demonstrated a good recognition and positioning effect in a real-world long-distance fiber-monitoring environment and effectively reduced the false positives of the pre-warning system.

A photon-level imaging-detection system is developed based on 64?64 pixel Geiger-mode avalanche photodiodes(Gm-APDS). By performing a low-frequency imaging-detection test, imaging detection of a point target 4.3 km away is realized. The experimental results show that the low-frequency photon-level echo laser imaging-detection technology based on Gm-APDS can quickly image and detect point targets without requiring long-term and multiple cumulative detections. This study lays a good technical foundation for the active detection of long-distance fast-moving targets in air.

In order to realize high efficiency auto focusing function in infrared imaging system, the characteristics of infrared imaging systems are analyzed. The optimization methods of auto-focusing functions in infrared imaging systems are discussed and summarized. Combined with practical engineering applications, a method for estimating the local random fluctuation noise of the image sharpness evaluation function is proposed. The method improves the reliability of automatic focusing processes. By introducing the sensitivity of image sharpness evaluation function as feedback into the climbing process, the convergence speed of the climbing algorithm can be optimized. The optimization methods were applied to engineering applications, and its implementation verified the correctness of the design and the effectiveness of the optimization methods.

By applying a sample test method, phosphate glass filters were tested in the tropical ocean environment of Wanning and the tropical rainforest environment of Xishuangbanna. Through observations and statistical analyses of the corrosion rate, weight-loss variation, and surface morphology under the two experimental environments, the corrosion characteristics of the filter under different environmental conditions were studied. The test results showed that the corrosion law was the same for the filters in the tropical rainforest and tropical ocean. However, the corrosion rate of the filter in the tropical ocean was greater than that in the tropical rainforest. The corrosion rate of the filter without a film was higher than that of the filter coated with MgF2. This study has introduced a method for improving the anti-corrosion ability of filters.

Three different passivation films were used to prepare InSb detectors for testing the current–voltage(I-V) characteristic curves of chips with different perimeter–area ratios. The influence of surface leakage current on the performance of an InSb detector was analyzed by comparing the dark-current density at a bias voltage of ?0.1 V. The test results indicated that SiO2+SiNx passivation could significantly reduce the surface dark current. The capacitance–voltage(C-V) test results also demonstrated that the composite passivation film could significantly reduce the interface fixed charge. A composite passivation-film process was applied for the preparation of a 128?128 15-?m InSb focal plane detector. The optimal value factor of the detector chip was R0A ≥5?104·cm2, which was much higher than that before the test (R0A≈5×103·cm2).

White-light OLED devices can be used as a surface light source and for full color displays, which have become the focus of OLED research and development. However, there are some shortcomings in the realization of white-light OLED devices, such as insufficient luminous performance, difficult process control,and easy color drift. To solve these problems, the optimized ratios of primary colors(red, green, and blue) of organic OLED devices were preliminarily obtained through the realization of white light and calculation of the color coordinate. An experimental white-light OLED device was prepared with a color coordinate of(0.31, 0.35). Further, theoretical calculations and experimental optimizations were carried out to reduce the red-light doping concentration and increase the green-light doping ratio to realize standard white-light (0.33,0.33) OLED devices.项目、云南省技术创新人才培养项目（2017HB111）、云南省“万人计划”产业技术领军人才培养项目等资助。

An infrared temperature-measurement method based on a radial basis function (RBF) neural network is established in the case of unknown target emissivity. First, the strong nonlinear relationship between the target temperature and the peak of the radiance curve and its wavelength is derived. The inputs to the neural network are determined. Then, according to the RBF network, sample data are studied, and a target radiation-temperature-measurement model is established. The model does not require emissivity. A blackbody and steel plate target are used as test targets to prove the proposed method. The maximum relative error of the temperature of the blackbody is 0.016% and that of the steel plate is 1.08%. These results verify the rationality of the established temperature-measurement method.

Considering the strong correlation between adjacent band images of hyperspectral data in combination with the fast searching ability of the particle swarm optimization algorithm, a sparse decomposition algorithm of hyperspectral images based on spectral correlation is proposed. The hyperspectral images are divided into reference and common band images. Particle swarm optimization is performed on the reference band images to find the optimal atoms and realize their sparse decomposition. The optimal atoms of a common band image consist of two parts. Parts of these atoms are inherited from the optimal atoms of the reference band images, and the number of inheritances is determined by the spectral correlation between the common and reference band images. The remaining atoms are obtained using particle swarm optimization. The experimental results on hyperspectral data show that in cases with the same reconstruction accuracy, the sparse decomposition rate is approximately 18 times higher than the orthogonal matching pursuit algorithm.

To address the difficulty of image analysis in a complex environment, an image segmentation and detection method of an HSV spatial model was studied. First, a UAV was used to collect images. Second, Gaussian convolution was used to detect the gradient of the cracked image after the photovoltaic(PV) panel area was extracted. Finally, morphological image processing and the HSV spatial model were applied to extract the occlusion, and the ratio of the minimum external rectangle area to the PV panel area was calculated. This method can effectively segment and detect PV images in complex backgrounds and provides certain innovative and practical value.

This study aims to solve the problem of reduced detection accuracy caused by a complex target-position scene and an uneven size in the detection of the abnormal heating point in an infrared image of a high-voltage switchgear. According to the YOLO v3 algorithm, the basic network architecture was optimized by including a convolution module and adjusting some hyper-parameters to realize rapid detection and identification of abnormal heating points in high-voltage switchgears. Simultaneously, a dataset for abnormal heating points of infrared images in high-voltage switchgears was established, and appropriate weights were obtained through training. The experimental results indicated that the detection method had a fast recognition speed, high accuracy, and strong generalization ability. The test accuracy reached 91.83%, indicating that the method can be initially applied to the detection of abnormal heating-point targets in high-voltage switchgears.

For the infrared image-based fault diagnosis, the region of interest (ROI) needs to be selected.Due to the characteristics of many interference background and low contrast in infrared image, it isnecessary to remove the background and image segmentation to extract ROI. However, the common twovalue segmentation algorithm has the limitation of over-segmentation in the infrared image segmentation.Therefore, a method of infrared image ROI extraction based on region contrast and random forest isproposed in this paper. Firstly, the region contrast method is used to detect the infrared image significantlyto remove the interference background. Then, image segmentation is conducted by applying OTSUalgorithm in order to extract ROIinitially. Finally, aiming at realizing the optimal extraction of ROI, thethreshold of image segmentation based on the results of random forest classification is iterated andoptimized. Infrared images under 6 different conditions derived from the rotors test-bed are utilized forfault diagnosis, applying the ROI extracted by the proposed method to fault diagnosis, the accuracy of theclassification increased by 3.3 percentage points, which is more accurate than that of the artificial selectedarea.机故障诊断机制及预测预警模型研究（51674277）。

This paper proposes a fast image-segmentation algorithm with a multilevel threshold based on the Tsallis relative entropy and wind-driven optimization algorithm. First, the principle of the Tsallis relative entropy is analyzed, and single threshold segmentation is extended to multilevel threshold segmentation. Then, a Gauss distribution is used to fit the image histogram information after segmentation, and the Tsallis relative entropy is used to determine the best segmentation threshold. To improve the speed of the threshold-segmentation algorithm, a wind-driven optimization algorithm is used to find the optimal solution of the Tsallis relative-entropy function. Finally, the proposed algorithm is compared with exhaustive and particle swarm optimization algorithms. The proposed algorithm is also compared with the Otsu algorithm and the multi threshold-segmentation method based on two-dimensional entropy. The experimental results show that the proposed algorithm can be used for multi-threshold segmentation of images with high speed and high accuracy.

When eddy-current infrared thermal-imaging technology is used to detect metal-material damage defects, the infrared image is susceptible to noise and may also contain useless information, which can result in blurring of damage defects. To address this problem, a defect-detection method based on single-scale Retinex and improved K-means clustering is proposed to perform infrared image-feature enhancement, image segmentation, and edge feature extraction. First, the image is enhanced using single-scale Retinex. Additionally, the defect features are enhanced. Then, an improved K-means clustering algorithm is used to segment the image. Finally, a mathematical morphology algorithm is used to process the image, which removes the useless information in the defective image and uses a Canny operator to detect the defect edge. The experimental results show that the method effectively detects defects of metal-material specimens and extracts complete and clear defect edges of the metal-material specimens.

In recent years, solar blind ultraviolet detection has been widely used in power fault detection. Based on atmospheric temperature profile and MODTRAN, we conduct simulation analysis on the influence of atmospheric temperature on solar blind ultraviolet atmospheric transmission. And finally the relationship between atmospheric temperature and ultraviolet atmospheric transmission was obtained through wavelength analysis and curve fitting. Based on the simulation relational expression, the quantitative analysis of atmospheric temperature and atmospheric transmission can be carried out, which can promote the study of the characteristics of solar blind ultraviolet transmission in the atmosphere and improve the accuracy of corona detection.