In order to track the development and application of foreign high-energy laser weapon, foreign patent searches in the United States, Germany, the United Kingdom, Israel and other countries are carried out within the scope of the global patent database, and 435 related patents published after 2017 are obtained. Based on the analysis of the composition of the patented technology, it is understood that foreign countries have made breakthroughs in the combat distance, maneuvering target tracking and continuous combat capabilities of high-energy laser weapons. What's more, from the analysis of the regional distribution of patents, it is found that the technological innovations in the United States accounts for 61% of total foreign applications and 54.7% of patents are funded by the government. Moreover, for foreign important projects or government-funded patents, program profiling is conducted and the technical improvement direction of nuclear submarine laser weapons and dazzling laser weapons are deeply excavated. Additionally, through the tracking and analysis of foreign laser weapons patents, the direction of foreign technology development is found, providing information support for China's technological development and risk defense.
The nanosecond high-power laser emission system has an important impact on the detection performance of LIDAR. In this paper, the commonly used low-side gate drive is modified to obtain a high-side gate drive method, making it suitable for common cathode laser emission circuits. And the high-side drive method utilizes a half-bridge driver with two GaNFETs as core devices to control the light emission of the laser diode. The effects of voltage, capacitance, resistance, and inductance on the current waveform flowing through the laser are analyzed in this topology. After experimental testing, the circuit is capable of generating a pulse width of 3.1 ns and a peak power of approximately 65 W at high repetition rates, which can be used as a transmitter module for LIDAR to scan and image scenes.
The laser diode (LD) ring uniformly side-pumped laser oscillator has the characteristics of simple structure, high beam quality and large single pulse energy, etc and it is often used as the oscillator stage of main oscillation power amplification (MOPA) structure of the oscillator in high-energy pulsed solid-state lasers, which is an important direction of the domestic and foreign researchers' research. In this paper, the influence of different arrangement of laser diodes and different absorption coefficients of working materials on the absorbed light field is simulated by Zemax software. Based on the simulation results, a ring side-pumped structure with 20 bar strips in series is designed to improve the uniformity of crystal absorption. Through the ring side-pumped structure, with the unstable resonator design, a laser oscillator is built for experimental study. A 1064 nm laser with a single pulse energy of 496mJ and a pulse width of 12.4 ns is achieved at a repetition rate of 100 Hz, with a beam quality better than 9mm·mrad and an optical-optical conversion efficiency of 18.5%. Moreover, the oscillator can maintain the stability of optical axis at the same time, to realize 20 ns pulse width the within variable gear adjustable. The experimental results show that the ring side-pumped structure is reliable and effective, which provides an effective scheme for selecting the oscillation stage of MOPA structure.
In this paper, a point cloud semantic segmentation network combining multi-scale supervision and SCF-Net is proposed to address the problems of low segmentation accuracy of point cloud in complex scene, the lack of direct supervision in neural network hidden units, and the difficulty in extracting specific point cloud features. A category information generation module is first constructed to record the receptive field categories of hidden unit in the encoder, which is used for the supervised learning of auxiliary classifiers in the decoder. Secondly, the point cloud category prediction task in the decoding stage is decomposed into a series of point cloud receptive field category prediction tasks. By adding auxiliary classifiers to each layer of the decoder, the point cloud receptive field category of the current stage is predicted and the category information generated in the coding stage is used as the label to supervise network learning. The model infers point cloud receptive field categories from coarse to fine, and finally predicts point cloud semantic labels. The experimental results show that the method can effectively extract key information of point cloud and improve the accuracy of semantic segmentation.
Aiming at the problems of graph convolution-based point cloud classification models in extracting feature information from different semantic regions of the point cloud and efficiently utilizing aggregated high-dimensional features, a novel point cloud classification model is proposed, which combines dynamic adaptive graph convolution with multi-layer pooling. Specifically, residual structures is employed to construct deeper convolutions and learn feature information from different semantic regions of point pairs at different levels to generate dynamically adaptive adjusted convolution kernels that update the feature relationships of different point pairs, thus extracting more accurate local features. At the same time, the aggregated high-dimensional features are input into a multi-layer max pooling module to recover the discarded feature information from the first max pooling layer and obtain richer high-dimensional features to improve the accuracy of the classification model. The experimental results show that the proposed model achieves an overall accuracy of 93.3% and an average accuracy of 90.7% on the ModelNet40 dataset, which is significantly better than the current mainstream point cloud classification models, and has strong robustness.
In this paper, the repair process of mechanical parts is studied by the method of "laser-electric composite wire cladding". Multi-pass single layer cladding experiments are conducted on 45# steel surface using 630 stainless steel wire to investigate the effects of different feed distances on the macroscopic morphology, microstructure, microhardness and wear resistance of the clad layer. The results show that with the increase of feeding distance, the bonding surface of the clad layer and the substrate became uneven, the surface of the clad layer has small particles of impurities, the thickness of the clad layer decreases, the microstructure is chaotic isometric crystal, the average hardness of the clad layer fluctuated above and below 200 HV. Moreover, the coefficient of friction shows a tendency of decreasing and then increasing, and there is a positive correlation between the amount of wear and the coefficient of friction. The wear patterns of the wear marks are observed and the wear forms are mainly fatigue wear and adhesive wear and the best wear resistance is achieved when the feed distance is 1.2 mm under the current experimental conditions.
Accurate detection of gas concentration is in urgent demand in the fields of atmospheric environmental protection, industrial production control, and exhaust emission monitoring. Tunable diode laser absorption spectroscopy (TDLAS) is an important method for achieving accurate detection of gas concentration. However, temperature variations bring significant errors to the accurate measurement of concentration, making temperature correction of the results essential. In this paper, the mechanism of temperature influence of gas concentration detection by TDLAS technology is described and the temperature influence correction methods for gas concentration measurement based on TDLAS technology is focused on analyzing and summarizing, and the trend of its development is envisioned.
In order to detect and track road targets in complex urban intersection environments, a multi-target detection and tracking algorithm based on roadside LiDAR is proposed. Firstly, the background subtraction method is used to filter out the background point cloud. Then, the curved-voxel clustering algorithm is used to detect the target to obtain 3D bounding box information with fusing 5 frame point clouds. Subsequently, a double-validation gate and life cycle management strategy with adaptive threshold are put forward, which effectively improves the accuracy of object matching and reduces object missing and false detection. Finally, the fusion algorithm of Interacting Multiple Model-Unscented Kalman Filter and Joint Probability Data Association was used to track road targets. The experimental results show that the algorithm meets the real-time requirements while ensuring detection and tracking performance, and has an engineering application value.
In recent years, convolution and graph operations have been widely used in research to capture feature information from point clouds, leading to good performance in semantic segmentation tasks. However, these methods have limitations in representing local information of point clouds, and a significant amount of feature information is lost by employing symmetric pooling operations. To address these issues, the DualRes-Net network is proposed. The network incorporates a Position Encoding Module to encode local coordinate features, enabling the network to focus on point cloud position information and obtain better local feature representations. And differences between center and neighboring points are combined with attention using a Dual-distance Attention Pooling, enhancing the adaptive aggregation ability of attention pooling for local point cloud information. A De-Differentiation Residual structure is used in each stage of the network to extract deep features of point clouds. Since different input types have significant distribution differences, MLP is applied to each type of feature separately to stabilize model training and improve model performance. Finally, in the semantic segmentation experiments in S3DIS Area5, the semantic segmentation performance of the proposed method achieves a mIoU of 63.7%, surpassing many existing networks, and demonstrating the effectiveness of the method.
As a key foundation work for nuclear power plant construction, construction setting out is the basis for other construction activities, which has a direct impact on the quality and progress of the project. At present, the construction setting-out method of point-by-point setting-out with total station is mainly used in the field of nuclear power construction and installation, which has the disadvantages of poor setting-out accuracy, low construction efficiency and large investment in labor costs. In this paper, the application of 3D laser projection technology is put forward to realize the rapid construction lofting of nuclear power plants, which has the characteristics of batch lofting and high projection accuracy. Through the feasibility analysis of construction lofting of nuclear power plants and the on-site application of a nuclear power project, as well as the use of total station for comparative analysis, it is verified that 3D laser projection technology has a certain application prospect in the construction lofting of nuclear power plants.
Indium antimonide (InSb) material is an important material for mid-wave infrared detections and Hall elements due to its special properties. This paper summarizes the current demand for InSb materials and the research progress of mainstream InSb material manufacturers at home and abroad, and highlights the latest research progress of the 11th Research Institute of CETC in InSb materials. Finally, this paper presents the future development trend of InSb materials.
Calculation of atmosphere transmittance is an important prerequisite for effective infrared characterization of airborne targets. In this paper, for the law of infrared absorption transmittance of lower atmosphere changing with altitude, a numerical simulation model is established. Based on the fact that water vapor and CO2 absorptive capacity in the lower atmosphere are affected by altitude, the important effect of atmospheric density change with altitude is added, and the absorption transmittance of the lower atmosphere is analyzed comprehensively for both horizontal and inclined path. The simulation experiments are conducted to compare the results of the numerical simulation model and MODTRAN5, and the reliability and validity of the model is verified
Shipborne laser weapon is an effective, fast and low-cost defense against drone swarms. In this paper, aiming at the countermeasures of ship drone swarm, the damage mechanism of shipborne laser weapons to drones of different materials is analyzed. Firstly, combined with the laser damage threshold of drone materials, a laser weapon countermeasure drone swarm model is established. Then, the effective striking distance of drone swarm under different laser power conditions is simulated, and the combat effectiveness of laser weapons against drone swarm is evaluated and analyzed finally, which can provide support for the subsequent laser weapons coordinated fire against UAVs.
Airy beams have important application prospects due to their non-diffraction, self-accelerating and self-healing characteristics. In this paper, a new method for modulating Airy beam energy modulation is proposed based on the programmable characteristics of spatial light modulator. A theoretical in-depth analysis of the energy modulation of the Airy beam by introducing a controllable modulation phase into the phase diagram is presented, and the relationship between modulation parameters and energy modulation range is quantified. Simulation and experimental results show that the beam energy distribution can be flexibly adjusted by changing the modulation phase, and the research in this paper helps to further promote the application of Airy beam.
In this paper, an intelligent optimization algorithm is used to optimize the design of a wide-band achromatic superlens. Firstly, the optical properties of silicon nanocrystals (SI nanocrystals) constructed on silica substrate are studied, and a database is constructed by scanning the radius of the unit structure through numerical simulation. Then, a hybrid algorithm of particle swarm optimization and genetic algorithm (PSO~~GA) is used to find the optimal phase matrix in the phase database, that is, the optimal cell structure corresponding to each position. Finally, the superlens is constructed, and verified through simulation experiments that it can realize the focusing in infrared band (1000~1250 nm), and the achromatic effect is better than that of the superlens structure obtained by the traditional particle swarm optimization algorithm. Therefore, the design of this superlens provides a solution for the automatic design of planar optical devices.
With the rapid development of aviation unmanned system technology, distributed airborne image stitching technology has become a high-profile research field. In this paper, an improved algorithm based on the APAP image stitching algorithm is proposed to solve the problems of large parallax and complex spatial geometric transformation in distributed airborne image stitching. The algorithm employs deformation processing, linear homography smooth extrapolation to global transformation, and mesh division method, which effectively eliminates blurred ghosting, reduces projection distortion at edges, and improves the operating efficiency of the algorithm. In experiments under multiple scenarios, the proposed algorithm exhibits smaller alignment errors and higher image quality metrics, including root mean square error, peak signal-to-noise ratio, structural similarity, and image entropy, among others. Moreover, when performing large-scale image stitching, the improved algorithm can achieve large-scale stitching of 154 images to obtain a 10 k×10 k high-resolution panoramic image, with a stitching time of 138 s. Therefore, the improved algorithm has important practical application value and can be used in the practical application of distributed airborne image stitching.
In this paper, an improved infrared small target detection model, infrared-YOLOv5s, based on YOLOv5s is proposed to address the problems of low resolution, complex background and lack of detailed features of infrared images. In feature extraction stage, SPD-Conv is used for down-sampling, which divides the feature map into feature sub-maps and concatenate them by channel to avoid the loss of features caused by down-sampling in the process of multi-scale feature extraction. And an improved atrous spatial pyramid pooling module is designed to improve feature extraction capabilities by fusing features with different receptive fields. Then, in feature fusion stage, a deep-to-shallow attention module is introduced to embed deep semantic features into shallow spatial features to enhance the expression of shallow features. Moreover, in prediction stage, the prediction layers, feature extraction layers and feature fusion layers for large target detection in the network are cut down to reduce the model size and improve real-time performance at the same time. The effectiveness of each module is verified by ablation experiments, and experimental results show that the proposed model achieves 95.4% mAP0.5 of on SIRST dataset, which is 2.3% higher than that of original YOLOv5s. The model size is reduced by 72.9% to 4.5 MB, and the inference speed on Nvidia Xavier reaches 28 f/s, which is conducive to the actual deployment and application. Therefore, the effectiveness of the proposed model is further verified by transfer experiments using Infrared-PV dataset, and the proposed model can meet the real-time requirements while improving the performance of small target detection in infrared images, and is suitable for the task of real-time small target detection in infrared images.
The large-field visual landing guidance system needs to quickly detect the cooperative targets mounted on the UAV during the autonomous landing of the UAV. The cooperative target exists in the form of light spots on the image, so in order to meet the real-time requirements of the system, a fast spot detection algorithm based on contour features is proposed in this paper. Firstly, according to the characteristics of light spots in the image, the target clipping method is used to extract the light spots in the original image, so as to reduce the amount of computation. Then, through the image preprocessing, the irrelevant information and noise interference in the background are eliminated to enhance the clarity of the spot. Finally, the least square algorithm is used to locate the center of the light spot by ellipse fitting. The experimental algorithm is compared with other spot detection algorithms so as to verify the real-time performance of the system. The results show that the proposed algorithm can reduce the running time to 36ms while ensuring the accuracy.
Pedestrian recognition based on infrared images is an important component of modern security systems. In scenarios with limited computing resources, it is often difficult to balance the detection accuracy and deployment difficulty due to the influence of model size in infrared pedestrian detection algorithms. In response to this issue, a lightweight object detection algorithm based on YOLOv5s is proposed in this paper. Firstly, the MobileNetv3 lightweight feature extraction network is introduced and deep separable convolution is used to reduce the model size, making it easier to deploy to CPU devices. Secondly, the nearest neighbor interpolation upsampling method is replaced with CARAFE (Content-Aware ReAssembly of FEatures) which significantly improves the image reconstruction effect. Finally, EIOU Loss is used as the loss function of the bounding box to improve the regression performance of the model. Additionally, tests are conducted on the sampled LLVIP infrared pedestrian image dataset and the results show that for pedestrian targets in infrared images, the model size is reduced by 80.6% and the number of parameters is reduced by 82.8% while maintaining a high detection accuracy (AP=95.4%); and the inference speed is improved by 43.3% when using a CPU platform for inference, and the performance of detecting multi-scale targets is improved. The above two results validate the effectiveness of the algorithm.
The aim of this paper is to address the difficulty in pedestrian target recognition in intelligent assisted driving systems due to the influence of light and climate on visible light cameras. A pedestrian target detection algorithm is implemented and improved by studying image fusion techniques in combination with deep convolutional neural networks. Firstly, using multi-source sensor image fusion technology, the strategy of fusing visible light cameras and infrared thermal imaging cameras, based on the Faster RCNN algorithm, a pedestrian target detection algorithm based on infrared thermal imaging technology and improved depth convolutional neural network is proposed. Then, the research is carried out in terms of improving network structure, feature fusion, optimising model training, and so on, and the research is carried out on pedestrian detection and localisation tracking in complex environments. Finally, the experimental results show that this algorithm improves detection efficiency and accuracy for human target detection in complex climate environments, and increases the safety of intelligent assisted driving vehicles.
In this paper, a blue or green laser detection method with large dynamic and energy adaptability is proposed to solve the problem of optical signal acquisition failure caused by the power saturation in short distance and the power shortage in long distance in underwater optical wireless communication links. A gain-controllable detection module consisting of an electronic control liquid crystal light valve, an adjustable gain photomultiplier tube and an FPGA board is added into the underwater wireless optical communication system, which forms a closed-loop control on the communication link in two dimensions of link power and detector gain to ensure that the photomultiplier tube is always working in an optimal state, realizing the underwater long distance, high speed and large dynamic wireless optical communication.
In response to the non-stationary and nonlinear characteristics of fiber optic temperature signals, as well as the limitations of using fiber optic temperature difference to identify the shallow burial position of submarine cables during the temperature equilibrium time period when the surface temperature of the seabed and the temperature at the depth of the seabed are approximately equal, a shallow burial state identification method is proposed based on optimized VMD mixed domain features and LSTM for identifying the two states of deep and shallow burial of the cables. Firstly, a parameter optimized VMD is used to decompose the fiber temperature signal and extract the component with the highest correlation coefficient between the intrinsic modal components of each order and the original signal. Secondly, the time-domain and frequency-domain features of the original temperature signal are extracted, and a mixed-domain feature set is constructed by combining the time-domain and frequency-domain features as well as the energy and entropy features of the selected IMF, and the CDET is used for sensitive feature selection. Finally, an LSTM structure is designed, the training sets are inputted into the network for training, the test set verifies the effectiveness of the network with the test set, and achieve shallow burial state recognition of submarine cables. Through on-site collection of submarine cable fiber temperature data for verification, the testing accuracy reaches 100%, and the results show that this method can accurately identify the shallow burial state of submarine cables.