With the increasing usage frequency of airborne photoelectric radar, its detection distance seriously deviates from the factory indicators. Thus, the working principle of airborne photoelectric radar and the influences of the radiation intensity of the target, atmospheric condition and optical system on the maximum detection distance are analyzed. According to the detection principle of photoelectric radar, the design idea of a new type of portable test system is put forward. The blackbody and parallel light pipes are used to simulate the infrared radiation intensity of targets at infinity and the attenuation feet, stepping motor and transmission gear are combined to simulate different atmospheric conditions. Thus its prototype is ultimately determined and the infrared attenuation measurement test is finished. In addition, the application method and test results of the novel portable photoelectric radar test system are described. As for this test system, it is easy to carry, its test efficiency is high, it is suitable for extremely harsh environmental conditions, and it can be extended to various types of infrared detection system performance tests.

To study the regions of spreading of truncated beams propagating through non-Kolmogorov turbulence, the expressions for turbulence distance and Rayleigh range are derived, the transport path is divided into three regions, and the relevant numerical analysis of the three regions is given. Results show that the lengths of the first region and the second region decreases with increase of coherence and truncation parameter, and the starting point of the third region reduces with increase of coherence and truncation parameter. The smaller coherence and truncation parameter are, the more likely it is to ignore the effect of turbulence on the beam propagation in Rayleigh range. In the three regions, the turbulent distance increases with the change of the generalized exponential. The corresponding physical explanations are given for the main results.

An adaptive compressed sampling approach for color images is proposed based on extended wavelet tree theory and multitask Bayesian model. Exploiting the relationship of parent-children coefficients and sibling coefficients in extended wavelet trees, the images in the red, green, blue channels of the color images are adaptively compressed. Exploiting the correlation of the three channels of color images and multitask Bayesian model, the sampled high frequency wavelet coefficients of three channels are dealt, respectively, and then the color images are reconstructed and fused. The research results show that when the sampling times are 600 and the sampling rate is 14.6%, the peak signal to noise ratio values of the colored reconstructed images obtained by proposed method are all above 27 dB, while the mean value of color difference is the least, the color difference values tend to be stable, and the color consistency and stability of the images can be kept well.

The solid-core polarization-maintaining photonic crystal fiber is very suitable for the application of the fiber optic gyro because of its unique advantages in birefringence, temperature, and radiation resistance. However, the loss is large, which affects the improvement of the performance of the photonic crystal fiber gyro. The scattering caused by the surface roughness of the air hole inner wall is one of the causes of the loss. For the scattering loss of the solid-core polarization-maintaining photonic crystal fiber, a fiber scattering model is established, and the simulation result of the scattering loss is 0.179 dB/km. A fully automatic test device is built with a measurement sensitivity of 1 pW, a scattering angle measurement range of 15° to 165°, and a fiber rotation angle resolution of 1°,the measurement of a three-dimensional scattering ball is realized. And the measured scattering loss of the fiber is 0.23 dB/km, which is in agreement with the theoretical simulation result.

Based on the correlation method and the Levenberg-Marquardt (LM) algorithm, this study proposes a correlation-LM algorithm for fitting the central frequency of the Brillouin spectrum. The proposed algorithm exploits the correlation method because it does not require an initial value and is insensitive to noise. Therefore, the correlation method is used to determine the iterative initial value. Inputting the initial value into the LM algorithm for iteration yields less-accurate results. Accurate information of the Brillouin scattering spectrum is finally obtained. Thus, overcoming the problem of iteration using the LM method is necessary as it has the disadvantage of being extremely sensitive to the initial value. Therefore, this study also proposes a new bimodal Brillouin spectrum fitting method and the Brillouin peak selection principle with double peaks. This algorithm can effectively improve the fitting accuracy and the fitting efficiency, making the single-peak fitting error within 1 MHz, the double-peak fitting error within 2 MHz, and the fitting degree to be 0.9988. To test the practicability of the algorithm, it is used to detect the transformer winding deformation. Two strains are applied to the modified winding. The correlation-LM algorithm could extract the central frequency accurately and decouple the strain information. The parallel test error is within 3 MHz, the fitting degree reaches 0.9964, and the strain is accurately located, thereby proving the superiority and practicality of this method.

A full duplex visible light communication system based on single light source is designed, which realizes the functions of both visible light illumination and communication. The cat eye reverse modulator is used to perform remodulation on the downlink optical signal, and the uplink information is modulated onto the downlink optical signal and reflected back. The experimental result shows that the system with single light source can realize full duplex visible light communication and make the communications between the uplink and the downlink independent, and the communication effect is good.

In the measurement of surface topography of objects based on fringe projection, the fringe phase distortion is the main factor affecting the accuracy of object measurement. In order to accurately determine whether the fringe phase has phase distortion, the fringe pattern is analyzed from the frequency domain by means of Fourier transform in the phase correction algorithm, and a technique for discriminating the phase distortion of the interference fringe based on frequency domain analysis is formed. Using the basic principle of Michelson interference, the parallel beam interference projection device is built, the interference fringes are projected onto the light screen, the image is collected by the CCD camera, and the obtained interference fringe image is transmitted to the computer, and then the image is filtered by wavelet filtering. The noise and interference are transformed into a frequency domain image by Fourier transform, and the energy distribution information of the adjacent position of the fundamental frequency in the frequency domain image is analyzed to determine whether phase distortion occurs. The experimental results show that the method can accurately determine whether the fringe image has phase distortion, and the recognition sensitivity is 60 mrad.

Focusing on the issue that the traditional kernelized correlation filter (KCF) has poor performance in handing heavy occlusion and illumination variations, a long-term KCF tracking algorithm is proposed combined with fast corner detection and bidirectional optical flow method. First, the KCF tracker is used to extract the multi-channel features of the histogram of gradient, color attributes, and gray features at the target location. The output response map is calculated and the peak sidelobe ratio (PSR) of the tracked target is obtained. The PSR and the empirical threshold determine whether the target is occluded by comparison. When the target is occluded, the bidirectional optical flow method is used to redetect the target position of the next frame based on the corner points detected by the fast corner detection, and a new template updating strategy is adopted to deal with the heavy occlusion. Compared with other algorithms, the proposed algorithm is effective and robust to the processing of occlusion and illumination variations.

Aiming at the applications of moiré fringes in fine measurement, a super-resolution algorithm is proposed based on local self-similarity and deblocking post-processing. In this algorithm, with the local self-similarity of moiré fringes, the initial high-resolution images are first obtained through an interpolation of the original low-resolution images. Then the optimal matching low-resolution block corresponding to each high-resolution image block is searched. The prior knowledge is extracted from the high- and low-resolution image blocks and thus the super-resolution reconstruction of a single-frame image is realized. In addition, the blocking artifacts are introduced in the reconstructed results after the blocking operations. As for this problem, a post-processing algorithm for quickly eliminating blocking artifacts is simultaneously proposed. The results show that the combination of the proposed two algorithms can effectively improve the image quality and simultaneously eliminate the blocking artifacts in the reconstruction process of images. The algorithm does not relay on external images and has a low computational complexity, suitable for the super-resolution reconstruction of moiré fringe images.

In recent years, the convolutional neural networks are widely used in the field of object detection. However, these methods based on convolutional neural networks require a large amount of calculations, so that it is difficult for these methods to run on platforms with limited computation. A fast object detection method is proposed based on single shot multibox detector (SSD), namely Faster-SSD. The method realizes the real-time detection and high accuracy with limited computation. The basic network of SSD is replaced with ResNet-34. In the stage of generating the prediction frame, first obtain the prior boxes which satisfy the condition, and then generate the prediction frame of the corresponding category. The variable minimum threshold is proposed to reduce the amount of computation. Finally, the online hard example mining is applied to remove the simple samples. Experimental results show that the Faster-SSD gets 14 frame/s on NVIDIA Jetson TX2.

In order to solve the problem of low face recognition rate caused by the complex lighting conditions in coal mines, a face recognition method applied to the underground coal mines with complex lighting conditions is proposed. First, the face image is decomposed into low-frequency and high-frequency components by wavelet decomposition, and simultaneously the histogram equalization processing is conducted on the low-frequency components to enhance the image contrast. Then, the wavelet denoising model with a fuzzy degree of membership factor is used to filter the high-frequency components and meanwhile a new PAL fuzzy enhancement algorithm is adopted for the fuzzy enhancement of high-frequency components. Under different thresholds, a non-linear transformation is used to get feature images with different scales and different directions, and the anti-fuzzy processing is conducted. Finally, the processed low-frequency and high-frequency components are reconstructed based on wavelets. The experimental results show that the proposed face recognition method can be used to effectively improve the overall effect of face images and enhance the detail information of images under the complex lighting conditions in underground coal mines. Moreover, the average recognition rate can reach 94.45%, indicating the face recognition rate under complex lighting conditions in coal mines is significantly enhanced.

The fluorescence imaging technology is proposed for non-destructive detection of subsurface defects of fused silica optical elements. The subsurface defects of fused silica optical elements under different processing techniques are obtained by this method. Based on the damage properties of fused silica optical elements, the relationship between subsurface defects and damage properties of fused silica optical elements is analyzed. The results show that the damage threshold of fused silica optical elements is inversely proportional to the density of fluorescent defects, that is the samples with less subsurface fluorescence defects have high damage threshold. This shows that the technique can effectively evaluate the damage performance of fused silica optical elements. The results of this study have guiding significance for optical element processing technology.

The total variation image inpainting algorithm needs more iterations to remove the text and scratches. It is not ideal for the preservation of edge details, and has a staircase effect. The peak signal to noise ratio of the restored image is relatively low. Aiming at these defects, an improved algorithm is proposed. The logarithmic operation and additional directional gradient operator are added into the regular term of the original algorithm to avoid the staircase effect, reduce the number of iterations, and reduce the smoothing of image at edge details. Simulation results show that the improved algorithm reduces the number of iterations retains the details of the image, well and the restored image has good visual effect.

Image registration combining accelerated segmentation feature algorithm and fast retina keypoint (FREAK) algorithm is proposed. Firstly, the scale space is constructed for the image, and the image feature points are detected by the accelerated segmentation feature optimization algorithm. Keypoints are filtered by Harris algorithm and some strong corners retained are reserved for image registration. Secondly, the strong corners are described by FREAK and eigenvectors are calculated. Keypoints are matched by Hamming distance instead of traditional Euclidean distance. Matches are filtered with random sample consensus algorithm to avoid mismatch due to noise and moving objects. From the two aspects of registration accuracy and registration time, the comparative experiments between scale-invariant feature transform, binary robust independent elementary features, original FREAK and the proposed algorithm are carried out. The experimental results show that the proposed algorithm has the characteristics of fast registration speed, high accuracy and well-stability.

Aiming at the problems of ghost and the noise interference from dynamic background in classic visual background extraction algorithm, an improved visual background extraction algorithm is proposed. The important feature information of pixels in complex environment can be collected by creating the auxiliary sample set. By introducing analysis of the pixel ghost factor and the region complexity, the matching threshold and updating rate of each pixel can be adaptively adjusted. With the pixel flicker analysis based on sliding window, the points which may be misdetected as foreground can be updated to the auxiliary samples according to probability. The comparative experiments in the multi-scene show that the proposed method can reduce the wrong classifications rate to as low as 1.49%, eliminate the ghost quickly, suppress the noise interference from the dynamic background, and ensure the complete recognition of foreground target. The results of the algorithm are more accurate in the complex environment.

Based on the full convolutional network, an image segmentation model is proposed to obtain the target segmentation results. This model consists of two deep neural network branches with the same structures. As for each branch, a convolution-deconvolution structure is adopted to implement the feature extraction and to recover the target area from the features. These two branches receive different image inputs, and then the final segmentation results are obtained via the weighted fusion of the results from these two branches. This model combines the multi-level scale features of different image sources, and the training model is more robust through data enhancement when the number of training samples is limited. The experiments are carried out on the optical image dataset of Weizmann horse and the remote sensing image dataset of Vaihigen. The comparison with the related literatures is also made. The results show that the proposed model has a higher target segmentation integrity and an optimal segmentation performance. The ideal extraction results of remote sensing image buildings under the conditions of limited training data, various shapes, large scale changes and so on, indicate that the proposed model can be applied to the complex remote sensing image object segmentation.

Myocardial ischemia (MI) is a heart disease that can cause various types of fatal heart attacks. Patients often miss the best treatment time when they develop the symptoms of a heart attack. Early detection of MI is considered to be necessary for curbing the deterioration of heart diseases because it is difficult to observe the symptoms of a heart attack through a medical check-up. Infrared thermal images of 165 healthy patients with different degrees of MI are collected, and all the samples are divided into training set and test set. Further, the geometrical differences between the left and right sides of the precordial area are extracted based on the geometric positioning of the infrared thermal image of a particular human body. Additionally, several convolutional kernels are used to reduce the dimensionality of the temperature difference set.The training set is trained using the back-propagation (BP) neural network based on the cross-validation method, and the network parameters are determined for establishing a classification model. After the 3×3 size Gaussian kernel operator is convolved on the temperature difference set, the classification accuracy of the test set with respect to the BP neural network becomes 95.56%, thereby denoting that the predictions for the new sample are considerably accurate. Further, the proposed method can rapidly and accurately assist during the early detection of MI in a clinical examination and provide a new methodology for the pre-diagnosis of MI.

A laser multilateration network is a large-scale coordinate measurement network based on the multi-station measurement by laser trackers. The locations and number of trackers, i.e. the network placement, strongly influence the performance of this measurement network. Thus a grid-based placement optimization method is proposed based on the genetic algorithm, in which the coverage capability, measurement accuracy and overall cost are used as the multi-objective evaluation functions, while the global grids and local grids are used for the dividing of the placement area. Through the global search and local search, a global optimization can be achieved without any initial placements. The experimental results show that compared with the experience placement, the optimized network placement can meet the measurement requirements easily and reduce the requirement of operation experiences from operators. Thus it possesses a good practicality.

In order to improve the ability of recognizing the details of the measured free-form surface, restrain the jumping error caused by the processing of phase unwrapping, and reduce the sufficient conditions for the correct phase unwrapping, a full-frequency phase unwrapping algorithm based on multi-frequency heterodyne is proposed. First, the standard four-step phase shifting algorithm is used to solve the wrapped phase. Then, using the full-frequency phase unwrapping algorithm, the details contained in the fringe pattern with different pitches are converted by the relationship between the phase and the fringe pitch to improve the accuracy of the unwrapped phase details. Compared with the existing method, less constraints are derived in order to restrain the jumping error caused by the processing of phase unwrapping. Simulation results show that the proposed method has no jumping error after the phase unwrapping, and no additional error correction is needed. Experimental results verify that the three-dimensional reconstruction has higher precision, the reconstructed surface is smoother and the details are clearer. Compared with the existing method, the standard deviation of phase unwrapped error of the proposed method is reduced by 44%.

Point cloud registration is one of the key issues in the process of three-dimensional modeling. Fast and high precision registration is the focus of point cloud registration. A method of point cloud registration based on independent component analysis (ICA) is proposed. The independent components, mixing matrix and unmixing matrix are obtained after ICA of two point cloud. Because of the fuzzy problem of ICA, the independent components of the two point cloud may have the difference in the order and the symbol. The optimal transformation matrix of two independent components can be obtained under the optimization criterion of minimum F norm. Further, the accurate registration of point clouds is achieved based on the relationship between point cloud data and independent components. Experimental results show that the proposed algorithm has faster registration speed and higher registration accuracy.

The number of laser echoes can be obviously increased with multi-receiving telescopes to simultaneously receive signals, and the detection ability of target laser signals can be enhanced. However, the precision of time and frequency synchronization directly influences the errors of laser ranging data by means of multi-receiving telescopes. Thus the time and frequency synchronization method for the laser ranging system with multi-receiving telescopes is described based on fiber-time-frequency transfer. The performance of a fiber-time-frequency transfer device is tested and the results show that its time synchronization precision is 62 ps and the linear slope is 4 ps a day, which meet the requirements of laser ranging to satellites. The experiments of laser ranging to satellites are performed with single-transmitting telescope and dual-receiving telescopes based on the laser ranging system with dual-receiving telescopes and the fiber-time-frequency transfer device. Compared with that of the precise orbits of satellites, the error of ranging data derived from the dual-receiving telescopes is less than 6 cm, indicating that it can be used for the precise orbit determination of satellites. Moreover, the experimental results confirm the application feasibility of the fiber-time-frequency transfer method in laser ranging with multi-receiving telescopes.

The single-scan track and block experiment of 200 μm large layer-thickness gas atomized 316L powder by the selective laser melting (SLM) technique is conducted. Through the analysis of densities, microstructures, tensile properties and defect mechanism of forming parts, the process is optimized. The results show that the density is up to 99.99% when the exposure time is 120-160 μs, the overlap rate is 50%-60% and laser power is 400 W. The unmelted defects occurred during the SLM process can be removed by the adjustment of process parameters. Although the micropores and balling defects cannot be completely eradicated, a short exposure time can be used to reduce the balling size. The equiaxed crystals and columnar crystals can be observed in the microstructures of samples. The yield strength, tensile strength and elongation rate of these samples are 530 MPa, 635 MPa and 31%, respectively, indicating that there exists a nice tensile property.

There are about 16 spectral lines emitted from discharge initiated non-chain pulsed HF laser, covering from 2.65 μm to 3 μm, when the laser is working in full spectrum output. The absorption of H2O molecules at about 2.7 μm in air has a great influence on the laser pulse energy. The absorption attenuation rules of wide spectral band pulsed HF laser in air are obtained by using discharge initiated non-chain pulsed HF laser. The experimental results show that the absorption coefficient is about 0.066 m -1 just at the output window of the laser device in full spectrum output, and then decreases to around 0.01 m-1 after 40 m propagation. The absorption attenuation coefficient reduces when the propagation distance increases, and tends to be constant. By suppressing the laser wavelength below 2.8 μm of non-chain pulsed HF laser, the attenuation coefficient is reduced to about 1/3 of the value obtained before at the output window, i.e., about 0.022 m -1, and the attenuation of laser energy by air absorption is reduced to some extent.

The porous Ti-6Al-4V alloy structures with different pore sizes and porosity are fabricated by selective laser melting aiming at the problem of stress shielding existing in the Ti-6Al-4V alloy implants. In addition, the relative densities, molding precision, microstructures, compression properties and elastic moduli of different porous samples are characterized. The results show that the microstructures of the as-built porous structures are composed of fine needle α' martensite phases. As for the porous Ti-6Al-4V alloy structures with deferent relative densities, when the relative density increases from 0.420 to 0.548, the elastic modulus increases from 15.1 GPa to 25.7 GPa, and the compressive strength increases from 223 MPa to 352 MPa. Moreover, the relationship of elastic modulus and compressive strength with relative density satisfies well with the Gibson-Ashby model. In addition, the compression fractures of porous Ti-6Al-4V alloys occur at the junction of struts. The angle between the fracture and the horizontal direction is about 45° and the fracture mode is a brittle one.

A quasi continuous wave Nd∶YAG planar waveguide laser amplifier with a wavelength of 1319 nm is reported. The seed source is an oscillator based on a side-pumped rod Nd:YAG laser model, and the gain medium of the amplifier is a planar waveguide Nd∶YAG. The size of the YAG planar waveguide is 0.6 mm×10 mm×60 mm, the central area 0.1 mm×10 mm×50 mm is the doped region, and the surrounding area is undoped YAG. The pump source of the amplifier is a diode laser array, which is coupled from the rear end into the planar waveguide gain medium. The seed laser is coupled into the gain medium via the end facet, and a single-pass amplification is introduced. In order to achieve the good heat dissipation, two micro channel heat sinks are welded onto the Nd∶YAG planar waveguide. The end facets are coated high transmission at 1319, 1064 and 808 nm. When the driving current of the diode laser arrays is 110 A and the pulse repetition rate is 200 Hz, the output of 36 mJ 1319 nm laser is obtained, and the optical to optical efficiency is 8.3%.

In order to investigate the effects of a water jet with a certain speed and angle on the depth, width and heat affected zone of the processed groove in the water jet-assisted laser machining of silicon carbide. The Fluent software is used to simulate the effect of water jets at different speeds and different angles of incidence on the surface impact of the material. The impact of impact force on processing results is theoretically analyzed. At the same time, the relevant experiments are carried out for verification. The results show that when the water jet velocity is the same, the impact force on the surface of the groove gradually increases as the incident angle of the jet increases. Compared with the waterless laser processing, the water jet assisted laser processing has a great influence on the depth of the obtained groove, the effect of the width is small, and the heat affected zone is significantly reduced. There are less slag and re-condensation layer at the bottom and notch of the groove, thus the processing quality is better.

The blade is an important part of aero engine. It is easy to be damaged due to the bad working environment. It is economical to repair the damaged blade, and the model reconstruction is one of the key technologies of aero-engine blade repair. A blade model reconstruction method based on speckle vision measurement is proposed. The speckle vision system is used to collect the speckle on blade surface, and the local point cloud data are obtained by speckle stereo matching. Then, the entire point cloud of the blade is obtained through the point cloud splicing, at the same time, the boundary points are extracted according to the curvature of the blade point cloud, and the boundary points of the blade point cloud are fitted by the cubic B-spline curve to obtain the envelope curve of the blade. The envelope curve and point cloud are used to reconstruct the blade model. Finally, the experimental verification of the method is given, and the feasibility of the method is proved.

Aiming at the problem of tracking failure due to fast motion and scale variation during object tracking, an object scale adaptation tracking based on full-convolutional siamese networks is proposed. First, a full-convolutional symmetric network is constructed using MatConvNet framework, and the multidimensional feature maps of template images and experimental images are obtained by using the trained networks. Through the cross-correlation operation, the point with the highest confidence score is selected as the center of the tracked target. Then, through multi-scale sampling at the center, the error samples that are less than half the template variance are filtered out. The probability histograms of target templates and samples are built. The Hellinger distance between the template and the samples is calculated, and the appropriate scale is selected as the scale of the target tracking window. Experiments on the OTB-13 dataset are carried out. Compared with other tracking algorithms, the tracking success rate of proposed method is 0.832, and the precision is 0.899, which are higher than that of other algorithms, and the average tracking speed is achieved 42.3 frame/s, meeting the needs of real-time object tracking. Selecting the tracking sequences with fast motion or scale change attributes for further testing, the tracking performance of proposed method is still higher than other algorithms.

Camera pose estimation is widely used in computer vision and robotics. Aiming at the stability and real-time performance of camera pose estimation, a camera pose estimation algorithm considering the uncertainty of spatial reference points based on orthogonal iterative algorithm is proposed. The key idea of the algorithm is to obtain the weights of the corresponding feature points in consideration of camera distortion and use an accelerated orthogonal iterative algorithm to regularize the repeated calculations in the iterative process. And the camera pose is obtained by minimizing the weighted re-projection object residual function. Results of simulation data experiments and real image experiment show that the proposed algorithm has higher calculation accuracy, faster speed, and lower time complexity. In the case of deep spatial feature points or the feature points deviating from the optical axis of the camera, the time complexity and accuracy of the algorithm are better than the existing orthogonal iterative algorithms, indicating its feasibility in the real-time estimation of camera pose.

The periodic arrays of edge-to-edge nanorod oligomer on glass substrate are designed to investigate the magnetic resonances and to realize the Fano resonances. With the transverse excitation, that is, the polarized electric field of the incident light is perpendicular to the long axis of the nanorods, the planar symmetric nanorod trimer can obtain a single Fano resonance, while the metal-insulator-metal (MIM) nanorod oligomer can obtain double Fano resonances. The near field electromagnetic distributions and far field extinction spectra of the nanorod arrays are analyzed in visible to the near-infrared regions using finite element method. And the characteristics of its resonance peaks and its realization mechanism are studied. The analysis shows that the near field coupling and superposition of the localized surface plasmon resonance mode of nanorods excite its magnetic surface plasmons (MSPs) to obtain the Fano resonances. In particular, the introduction of MIM nanorods provides more possibilities for the realization of double or even multiple Fano resonances. The Fano resonances of designed nanorod oligomer arrays have the advantages of low resonant loss and high quality with the bandwidth of 30 to 50 nm, which can be potentially used in multi-wavelength biosensor, optical switch and other devices.

A wide range photoelectric tracking sensor is proposed and its working principle is introduced. A test platform is set up and the fields of view of altitude and azimuthal angle as well as its tracking accuracy are experimentally studied. The results show that as for this sensor, its field of view of azimuthal angle is -180°-180° and that of altitude is -90°-90°. Thus the requirement of wide response range for sensors is satisfied. In addition, the accurate tracking accuracy is high and the error is ±1.5°, therefore the requirement of tracking accuracy for the solar photovoltaic power generation is also satisfied. This research provides a basis for the development and use of solar tracking photoelectric sensors.

Perovskite is a promising candidate for photodetector applications owing to its outstanding optical and electrical characteristics, such as a tunable bandgap, a high absorption co-efficiency, and a long and balanced carrier diffusion length. Perovskite crystals are diverse in their morphologies, which include micro and nano crystalline films, single crystalline bulks, and single crystalline nanocrystals. In this study, the photodetectors composed of perovskite crystals with different morphologies based on the principles of photoconductive, photovoltaic, field effect transistor, and photomultiplication are reviewed, which reveal different characteristics in terms of responsivity, detectivity, and response speed. The research progress of perovskite photodetectors is summarized in each of the aspects of flexibility, narrow band response, self-powered photodetection, and patterned array. A future prospect of perovskite photodetectors is also discussed.

Semi-active laser guidance technique greatly improves the hits rate of munitions. The development of laser device, as the core part of laser target designators, is highly significant for the entire weapon system. Development and current status of laser device used for laser target designators is sketched. Current major technical principles and schemes are introduced subsequently, along with the pros and cons of the schemes. Finally, the laser device's future outlook is provided.

Terahertz quantum cascade laser (THz QCL) is a kind of compact and coherent solid continuous wave source with important potential applications, and it is one of the international research hotspots. At present, the research on array coupling is particularly important since it is an effective solution to the problems and difficulties in further increasing output power for THz QCLs. The current structures of THz QCL array coupling are summarized and analyzed. The research status and possible future development directions of THz QCL array are summarized and discussed.