The battlefield threat posed by of unmanned aerial vehicles (UAVs), especially small unmanned aircraft systems (UAVs), are increasing day by day, and have even become a pivotal element capable of altering the course of warfare, gaining asymmetric advantages on the battlefield, which brings an urgent need of counter-small Unmanned Aircraft System (C-SUAs). Major military powers around the world have intensified their investments in counter-small UAV operations, initially developing a certain level of counter-small UAV confrontation capabilities. Based on the analysis of UAVs and the requirements of counter-small unmanned aircraft warfare, the system framework, the technologies and the development status of C-SUAS are summarized in this paper, and the trend of development of C-SUAS is discussed as well. The findings are intended to provide guidance for the development and construction of C-SUAS in the future and to support counter-small unmanned aircraft warfare.

Aiming at the problem that the 3D LiDAR SLAM (Simultaneous Localization and Mapping) algorithm lacks loop-closure detection and are prone to odometry drift when the radar moves too fast, a SLAM method that integrates BA (Bundle Adjustment)/NDT (Normal Distribution Transformation) and 3D LiDAR LOAM is proposed. Firstly, the lidar BA algorithm is improved by minimizing the distance between feature points and edges or planes. Then, the LOAM algorithm is employed, where the front-end estimates coarse poses through frame-to-frame odometry, and the back-end refines the poses using BA. Finally, NDT is utilized to estimate the line and plane feature matrix, performing rotation-invariant transformations and loop closure detection until the latest frame aligns with the initial map, completing the global pose correction. The test is carried out on the public data set KITTI show that the proposed method achieves a root mean square error (RMSE) of 16 cm, reducing the error by about 33% compared to the LOAM algorithm, and the overall running speed is 2.23 times faster than LOAM. To further validate the method in real-world environments, a handheld modeling device is designed, using the pose data from the cartographer algorithm as a reference. The test results demonstrate that the proposed method achieves an RMSE of 20.23 cm, reducing the error by about 75% compared to LOAM, effectively improving positioning accuracy.

Aiming at the problem that the accuracy of terrain simulation of initial irregular triangulation in traditional progressive triangulation filtering algorithm depends on the selection of initial seed points, an improved progressive TIN filtering algorithm is proposed. Firstly, more initial ground seed points are obtained through a moving rectangular grid, and then the local terrain of the region is simulated by using the cubic surface fitting function. Erroneous seed points are eliminated by judging the height difference between the seed points and the fitting surface. Additionally, a multi-scale strategy is adopted to gradually reduce the number of seed points used for surface fitting at each level, achieving precise simulation of the local terrain. This process constrains the influence of noise points on surface fitting, thereby achieving the goal of removing noise points and obtaining the final seed points for constructing the initial TIN. The experimental results show that the improved algorithm can obtain more seed points and enhance the accuracy of seed point extraction. As a result, the final initial irregular triangulation network more closely approximates the real terrain, thereby improving the filtering accuracy.

With the advancement of artificial intelligence technology and the progress in robot technology, the issue of path planning has garnered increasing attention. Reinforcement learning is extensively employed for goal-oriented path planning of mobile robots due to its data-independent nature and strong generalization capabilities. Despite some achievements, several challenges persist, including a scarcity of multi-objective path planning algorithms, low utilization rate of effective experiences, difficulties in model convergence, and sparse environment rewards. To address these issues, the Soft Actor-Critic (SAC) algorithm is applied to multi-objective path planning for the first time, and a flexible motion evaluation-based method that incorporates priority experience replay and expert experience is proposed. The proposed approach enhances sampling efficiency by prioritizing empirical playback while optimizing the reward function enables timely and rational feedback from the environment after each action to overcome local optima problems. Additionally, imitation learning based on expert experience improves training efficiency in reinforcement learning. Finally, simulations are conducted on the ROS platform for multi-objective path planning where results demonstrate that compared to the multi-objective SAC algorithm, the proposed algorithm accelerates convergence in both simple and complex environments with obstacles while generating shorter, smoother paths free from collisions.

In this paper, the damage characteristics of CYMG series color CCD detectors under 1064 nm continuous laser are investigated. Combining the damage morphology, damage depth of the multi-layer structure of the color CCD, and the output image after laser irradiation, the damage mechanism of the color CCD is analyzed. Based on experimental observations, the damage process of the color CCD is divided into three stages: point damage, line damage, and complete damage. The study indicates that in the point damage stage, the color change in the output image of the color CCD is primarily caused by damage to color separation filter induced by the 1064 nm laser. In the line damage stage, longitudinal white bright line damage occurs. This is due to the fact that part of the laser light is transmitted and diffracted into the transfer channel, causing signal charge overflow from the transmission potential well. Damage between the clock lines and various electrodes results in some pixels being unable to transfer the signal charge packets through clock drivers, producing lateral dark lines. During the complete damage stage, damage to the N-type silicon substrate prevents signal charges from completing conversion, storage, and transfer processes, leading to the total loss of imaging capability in the color CCD.

Studying the smoke transmission characteristics of polarized lasers is crucial for enhancing anti-smoke interference imaging capabilities. The traditional Monte Carlo methods (MC methods) lack the tracking and recording of photon polarization and phase information when simulating beam transmission characteristics. Based on Mie scattering theory and the Electric Field Monte Carlo method (EMC method), different polarized lasers are selected for the simulation modeling of the laser smoke channel transmission characteristics, and the variation patterns of intensity and depolarization are analyzed. The results demonstrate that scattered photons maintain near-Gaussian distribution in the plane under Gaussian beam incidence. Meanwhile, with increasing attenuation distance, beam intensity decreases, diffusion broadens, and the degree of depolarization intensifies, gradually approaching unpolarized light. Comparative analysis with other fully polarized lights reveals that circularly polarized light exhibits the lowest degree of depolarization after transmission, demonstrating superior polarization-maintaining performance. This study holds theoretical guiding significance for the research on LiDAR imaging in the presence of smoke interference.

To solve the problem of the rapid startup of detectors, a systematic study is conducted on the factors affecting cooling of the cooling head structure. The cooling capacity of cooling head required for the temperature reduction of the cold-head structure is calculated, and a thermal network model is established according to the actual heat transfer path of the cooling head. The thermal resistance of cooling head is computed and analyzed, indicating that the thickness of cold shield affects both the cooling capacity needed for temperature reduction and the thermal resistance. These two factors are in conflict with each other, and there exists an optimal value for them. Through finite element simulations, calculations are carried out for different cold-head structures, cold-shield wall thicknesses, and adhesives. As a result, the optimal parameters for the cold-head structure in fast-startup detectors are determined, and targeted optimization designs are proposed.

Contact holes are the regions where electrodes for extracting signals from photosensitive elements of mercury cadmium telluride (MCT) infrared detectors are attached, and the uniformity of their shape and dimensions significantly impacts detector performance. The uniformity achieved by high depth-to-width ratio photolithographic patterning decreases as the distance between the photosensitive elements on the detector chip decreases. In this paper, a method is proposed in which a metal layer is employed as a mask to fabricate contact holes with high uniformity and desirable bottom morphology by thin resist lithography and thin resist etching. The non-uniformity of the contact hole is reduced from 11.54% of the photoresist mask to 5.98% of the metal mask. This approach overcomes the challenge of poor uniformity of the small-pitch HgCdTe chips, which is of reference significance for enhancing the uniformity of the detector signals.

In this paper, the growth conditions of highly crystallographic CdTe passivation layer are optimized to regulate the density of the film. The surface morphology, microstructure and optical properties of CdTe passivation layer are thoroughly analyzed by SEM and ellipsometry, and the optimal dense growth conditions are obtained. The interface electrical properties between CdTe films and HgCdTe with different densities and their effects on the properties of photodiodes are systematically studied by C-V and I-V characterization methods. The results show that the density of CdTe passivation layer is the best when the sputtering power is 200 W, which effectively improve the interface electrical properties between the passivation layer and HgCdTe. The MIS devices prepared by this power have a lower interface fixed charge density of 1.21×1011 cm-2 and a slower interface state density of 1.26×1011 cm-2. Moreover, diode devices prepared by this power can effectively suppress the surface leakage phenomenon of the device.

In this paper, an infrared long-term target tracking method based on a state judgment module is proposed to address the problem of insufficient confidence in tracking in infrared target long-term tracking scenarios and tracking errors caused by misjudgment in similar targets. The Siamese network is employed as the base tracker. Geometric information, discriminative information, and state information output by the tracker are fused to build a state judgment module, including LSTM network stack adaptation design, scene adaptive adjustment design, and trigger re-detection mechanism design. It achieves accurate evaluation of infrared target tracking status and long-term tracking. The proposed method is tested and validated on the infrared long-term tracking dataset LSOTB-TIR-LT, and the results show that the F-score of the proposed method reaches 0.681, representing a 17.6% performance improvement compared with the classic Siam RPN++LT algorithm. It is particularly suitable for application scenarios where infrared targets are occluded, disappeared, interfered by similar targets, or undergo significant changes in scale for continuous tracking.

In response to the increasingly advanced infrared detection, reconnaissance and precision strike technology, and to effectively enhance the battlefield survivability of personnel and equipment, the fiber impregnation molding method is explored in this paper. TMOS is selected as the silicon source, aqua ammoniae is used as the alkali catalyst, and methylsilicochloro form and n-Hexane are employed for hydrophobic treatment. Hydrophobic SiO2 aerogel composite insulation felts are prepared using an atmospheric pressure drying process, and their performance indicators, including microstructure, thermal insulation performance, hydrophobic properties, infrared spectroscopy and other performance indexes are characterized and studied. In the background of Gobi desert and an urban green space, comparative experiments are conducted using different materials such as camouflage nets and thermal insulation sheets alongside the hydrophobic SiO2 aerogel composite thermal insulation felts to verify their infrared stealth performance. The results show that the thermal insulation performance of hydrophobic SiO2 aerogel composite thermal insulation felt is significantly superior to that of materials like camouflage nets, thermal insulation sheets and other materials, and its thermal insulation performance gradually improves with an increase in the thickness, at the same time, it exhibits good hydrophobic performance, meeting the requirements of field environment.

Under the dual influences of high geological background and human activities, the soil of karst mining area is seriously polluted by heavy metals and is an important source area of heavy metal migration in the region. Rapid monitoring of heavy metal content can provide a basis for the study of regional heavy metal migration and transformation. In this study, the content of As and Sb in soil samples are determined simultaneously, and hyperspectral data are collected. By comparing nine hyperspectral data preprocessing methods, five optimal preprocessing methods are selected, and two hyperspectral inversion models of machine learning (partial least squares regression (GA~~PLSR) and random forest (GA~~RF) are established based on genetic algorithm to screen spectral feature bands. The results are as follows: (1) The multi-scattering correction+first-order differential (MSC+FD) pretreatment method is the optimal spectral pretreatment method for soil As and Sb; (2) The optimal inversion model of soil heavy metal As in karst mining area is partial least squares regression model, the R2 (decision coefficient) of the training set and the test set of this model are greater than 0.9, and the RPD (residual predictive deviation) is greater than 3.0. The best inversion model of soil Sb is the random forest model, the R2 of the training set and the test set are both greater than 0.9, and the RPD of the model is greater than 3.0; (3) The spectral characteristic bands of soil as from the karst mining area were distributed in the range of 446.1~468.4, 534.7~561.2, 685~728.3, 735.7~766.6, 817~839.7 and 852.2~1077 nm. The characteristic bands of soil heavy metal Sb are mainly 343.6~359.6, 671.5~686.5, 746~753.4, 921.2~998.6, 1024.3~1037.7 nm. The results can serve as a basis for rapidly obtaining the content of As and Sb in soil within karst mining areas.

Infrared small target detection has a wide range of applications in the military, security and aerospace fields. In these applications, small infrared targets are typically characterized by small target sizes, low signal-to-noise ratios, and a lack of texture features and shape features, posing significant challenges for target detection. The existing infrared target detection methods rely on complex feature extraction steps or unable to effectively extract information from small targets, making it difficult to achieve ideal detection results. In order to solve the above problems, an infrared small target detection model based on a nested network and an attention mechanism is proposed. Firstly, a nested network structure is utilized to effectively extract the multi-scale features of the image. Then, an attention mechanism is introduced to enhance the target area information and suppress the background information, so as to improve the perception ability of small targets in the infrared image and improve the accuracy and robustness of target detection. Experimental results show that the proposed method has significant advantages in terms of detection rate and positioning accuracy, and is of positive significance for further research in the field of infrared image processing.

Based on the basic theory of infrared radiation, a multi-band spatial azimuth optimization algorithm is proposed in this paper, in order to avoid the influence of the surface emissivity of the object on the infrared measurement. Through multi-band design and mathematical transformation derivation, an infrared multi-band temperature measurement vector group is constructed, enabling the closed-form solution of temperature under the condition of unknown surface emissivity of the measured object. Relevant experiments demonstrate that the calculated temperatures obtained using the multi-band algorithm exhibit a higher degree of agreement with the actual temperatures compared to conventional single-band infrared measurement methods, more effectively approaching the true values with smaller relative errors. The method can obtain the surface temperature of the measured object without the surface emissivity and the space azimuth calibration; reduce the error caused by space position in the infrared measurement process. The accuracy of infrared temperature measurement is guaranteed by the multi-band spatial azimuth optimization algorithm.

To address the issue of environmental background radiation interference in the process of aero-engines infrared radiation characteristic test, a systematic analysis of the environmental background radiation interference sources in the engine′s infrared testing process is conducted, and the background interference source suppression method is proposed. Four types of materials with different surface characteristics are utilized as the substrate of the suppression device. The inhibiting effects of different surface characteristics and different shielding angles are verified, and an efficient suppression approach for aero-engines infrared test is determined. By building the infrared test platform of the model engine, the four materials of aluminum plate, acrylic plate, wood plate and rock wool plate are selected as the base materials for the environmental background interference suppression devices to verify the inhibiting effect at the shielding angle of 60°and 90°. Through experiments, it is found that within the mid-wave infrared spectrum of 3~5 m, the best suppression effect is achieved when a wooden board is used as the base material for the suppression device at a shielding angle of 60°, with a suppression rate of up to 93.32%.

To meet the imaging demands of an airborne equipment with an ultra-wide field of view, a wide-angle, low-distortion, large relative aperture infrared optical system operating in the wavelength range of 3m to 5 m, and a field of view of 97°×97° is designed. A reimaging structure is adopted to reduce the diameter of the rear lens group, and on the basis of passive athermal design, the thermal expansion coefficient of the structural material is utilized to compensate for the residual thermal difference, achieving an athermal design over the entire temperature range. Distortion correction is achieved using "non-similar" imaging technology. The design results show that the system′s modulation transfer function (MTF) values are all above 0.4 at a typical spatial frequency of 33lp/mm, approaching the diffraction limit. The imaging quality is fine, with a maximum distortion of less than 0.6%, meeting the design demands.

Concerning the issues that existing infrared and visible image fusion algorithms inadequately extract detailed features from source images and fail to consider the relationship between the fusion network and the detection network, an improved infrared and visible image fusion algorithm based on RepVGG is proposed in this paper. Firstly, the network architecture of PIAfusion is used to replace the convolutional block of the image feature extraction part of the network architecture, and the feature reconstruction convolutional block part of the network architecture with the RepVGG convolutional block. Subsequently, the YOLOv5 detection network is employed to detect the fused images, with YOLOv5 being utilized to build detection losses. Then, the detection loss is leveraged to guide the training of the fusion network through backpropagation, ensuring that the fused images output by the fusion network can be more easily detected by the detection model. Finally, the detection results of the fusion images output by the fusion network in the YOLOv5 detection network are obtained. Compared with the existing fusion methods, the results show that the fusion image obtained by the proposed method has a good effect from the objective indicators and the detection results of YOLOv5.

In response to the challenges of poor resolution, low contrast, and low signal-to-noise ratio in infrared image object detection tasks, a lightweight infrared image object detection algorithm based on YOLOv8 is proposed in this paper. Firstly, a Faster Block module is constructed to replace the Bottleneck module in the Neck section, effectively reducing the number of model parameters and improving the lightweight level of the model. Then, the SE attention mechanism is added to enhance the network′s ability to focus on important features and improve the effectiveness of feature extraction, thereby enhancing the robustness and stability of the model. Meanwhile, a dual-layer routing attention mechanism is introduced to utilize the large amount of redundant information in the feature map and save computation and memory overhead through sparse connections. Finally, the loss function is improved by introducing the complete intersection to union ratio EIoU as the regression loss, which improves the regression accuracy of the model for the target bounding box. The experimental results show that compared with mainstream algorithms such as YOLOv5 and YOLOv8, the improved algorithm in this paper achieves a recall rate of 81%. The model volume is decreased by 7.2% and 21.6% respectively, with only 4.6 MB. Simultaneously, the parameter count and computational complexity are significantly reduced. Compared with mainstream algorithms, the improved algorithm in this paper exhibits significant improvements in detection accuracy, model volume, and computational complexity, thus meeting the detection requirements for infrared targets.

In order to enhance the operational performance of non-sensitive projectiles and achieve target recognition as well as precise attack, a YOLO-1D model based on the idea of YOLO model positioning is put forward in this paper to realize target identification and location of non-sensitive projectiles composite detection signals. Based on the signal variation characteristics of laser and infrared detection signals when a target is present, labels are designed and a dataset is constructed in line with the YOLO concept. Meanwhile, a network structure named Multi-Task Convolutional Neural Network (MT-CNN) based on multi-task learning strategy, and a random forest recognition algorithm relying on artificial feature extraction are constructed to serve as control group. Finally, the target recognition and localization performance of the above models are evaluated and compared with the UAV detection test data. The test results show that the YOLO-1D model achieved the best performance in terms of recognition and localization., and its recognition accuracy rate reaches 95.61% and the positioning accuracy rate reaches 81.10%.

In actual welding operations, due to the interference of complex environmental factors such as strong optical noise, spark splashing, and smoke, traditional weld seam detection methods relying on a single tracker often find it difficult to maintain accurate tracking of the weld seam, resulting in a significant decrease or even failure in tracking performance. Therefore, a real-time tracking and detection method for infrared images of weld surface defects based on U-net neural network is proposed. This method accurately identifies the weld seam features in the infrared image of the weld seam surface during the initial stage of welding, and precisely locates the feature points. In order to cope with the strong interference environment during the welding process, two parallel kernel correlation filters are designed to track the weld seam feature points, and the output results of these two trackers are fused through a Kalman filter to ensure real-time, stable, and robust tracking of the weld seam even in complex environments. Real time tracking of weld seam feature point information is used as a key input and fed into the U-net neural network. In the U-net architecture, a branch network is introduced to optimize the feature extraction process and improve the quality of the segmentation map, enhancing the ability to capture details of surface defects on the weld seam. Using the bounding box mechanism to analyze the segmentation map output by U-net, automatic determination of the position and size of defect areas is achieved, and infrared image detection of surface defects in welds is completed. The experimental results show that this method performs well in both weld seam tracking and infrared image detection of weld surface defects, with an evaluation function Q value as low as 21.36, indicating high detection accuracy.

To improve the sensitivity and stability of underwater depth detection, an enhanced packaging structure was designed. A novel design combining a sensitization bridge with a cavity structure is adopted to improve the sensitivity of FBG (Fiber Bragg Grating) sensors. FBG1 is fixed at the connection between the soft silicone top cover and the sensitization bridge, and tightly adheres to the sealant. FBG2 is secured at the bottom of a hard sealed shell to form a cavity type sensitization structure. FBG1 is highly sensitive to water pressure caused by water depth, and the functional relationship between wavelength variation of FBG1 and water depth is derived. FBG2 is used to compensate for FBG1 temperature offset. The correspondence between water depth and structural parameters is analyzed through simulation, and the deformation degree of the sensitized bridge under different external force conditions is compared. In an experimental test conducted in a 1.8 m deep swimming pool, this sensor is compared with a traditional multi-layer structured FBG sensor. The results show that the total wavelength change of the traditional structure is 1.524 nm, and the depth sensitivity is 8.4 pm/cm. After sensitization optimization, the total wavelength change can reach 3.312 nm, the depth sensitivity is increased to 1.84 pm/mm, and the average error is reduced by 28.9%. The feasibility of the sensitized design is thus verified.

Underwater wireless optical communication, a critical technology for achieving high-speed and large-capacity data transmission in aquatic environments, has received extensive attention in recent years. This paper summarizes the working principle of underwater optical communication, analyzes the advantages and disadvantages of underwater optical communication compared with other traditional of underwater communication methods, and the core technologies of underwater wireless optical communication systems are systematically expounded, with a focus on key technologies such as the selection of transmitting light sources, the construction and attenuation of underwater channels, and signal modulation and codec schemes. At the same time, the future development trend of underwater wireless optical communication is summarized, which provides a reference for the future development of underwater wireless optical communication.

To investigate the polarization characteristics of snowflake-like ice crystals under terahertz waves (1.5 THz, 2.524 THz, 3.437 THz), the Discrete Dipole Approximation (DDA) method is employed to calculate their linear polarization degree, circular polarization degree, and the fraction of circularly polarized light intensity. These results are compared with those of hexagonal plate-like ice crystals of the same equivalent radius. The findings indicate that the polarization characteristics of snowflake-like ice crystals are contingent upon the particle′s equivalent radius, shape, and the frequency of the electromagnetic wave. For particles of identical size, the polarization characteristics of snowflake-like ice crystal particles are significantly more affected at a frequency of 3.437 THz. Within the same frequency band, the larger the size of the snowflake-like ice crystal particles, the more pronounced the variation in polarization characteristics with respect to the scattering angle, with 90°emerging as a pivotal boundary for the scattering angle and the alteration in polarization characteristics. When hexagonal plate-shaped ice crystals are used as substitutes for snowflake-shaped ones, the degree of circular polarization would be significantly underestimated at 1.5 THz, whereas the degree of linear polarization would be overestimated at 3.437 THz. These research outcomes are of significant importance for gaining a deeper understanding of the mechanisms underlying the interaction between terahertz waves and ice crystal particles.