Magnetic field measurement technology has been applied in various domains extensively, including mineral exploration, electronics manufacturing, national defense and so on. Consequently, it play a crucial role in promoting development in science, technology, and societal. Presently, electric magnetic field sensors suffer from complexity in their insulation structure, high cost, susceptibility to interference, and difficulties in digitalization. In contrast, optical fiber magnetic field sensors offer superior advantages, such as excellent insulation properties, compact structure, digital compatibility, robust immunity to interference, and high reusability. Therefore, optical fiber magnetic field sensors present a highly viable solution to address these challenges. The working principle, current research status, and existing challenges of optical fiber magnetic field sensors based on the magnetostrictive effect, Faraday effect, and magnetic fluid materials are reviewed. Furthermore, the future development directions and prospects of optical fiber magnetic field sensors are forecasted.

The working temperature of semiconductor laser in TDLAS (tunable laser absorption spectroscopy) technology directly affects the accuracy of gas concentration detection. The experimental results show that the multilevel integral separation PID control algorithm has advantages in improving the response speed of the temperature control system, improving the accuracy and stability of the system, and reducing the overshoot. The temperature control system has a temperature control accuracy of better than ± 0.03 ℃ in the temperature control range of 10 ~ 40 ℃, and reaches the set temperature within 148 s. The temperature control overshoot is less than 2.5%. The design meets the requirements of TDLAS gas concentration detection and lays a foundation for high-precision detection of gas concentration.

Aiming to meet the actual demand of airborne 130 mJ laser director, a pulsed driver with for LDA with the characteristics of high reliability, wide work temperature range and miniaturization is designed. A overall scheme is a combination of energy storage unit with ultra-low equivalent series resistance, low power preceding stage charging circuit and backward stage discharging circuit with high power. In detailed circuits design, reverse direction protection and slow start circuits adopt a way with double PMOS. BOOST charging circuits’ controlling IC is CYT5022JVS with the characteristics of adjustable switching frequency and cycle-by-cycle current limit. Energy unit includes some shunt-wound Polymer solid lithium electrolytic capacitors with ultra-low equivalent serial resistance (ESR). To reduce heating power, discharge circuits’ switching tubes employ two IRF150P220s in parallel with on -resistance 2.3 m. In order to adapt wide temperature range, charging voltage need to compensate because of capacitance change in high and low temperature. Tests show that pulsed driver can output a pulsed current with 160A peak value, 230 s pulse width and 25 Hz frequency. The airborne laser director driven by this pulsed driver can output 130 mJ average pulsed energy, which can satisfy the needs of airborne optoelectronic system.

In order to achieve accurate measurement of laser power at different wavelengths and directions in multi -parameter laser warning, an optical power detection circuit was designed. This circuit mainly consists of two parts: a photoelectric detection module and an A/D acquisition module. The photoelectric detection module consists of a photodiode and an operational amplifier OPA2846, while the A/D acquisition module is composed of an AD9220 analog-to -digital conversion circuit and an FPGA interface circuit. The incoming continuous laser is converted into a voltage signal through the photoelectric detection module. The voltage signal is converted into a digital signal through the A/D acquisition module and transmitted to the FPGA through an interface circuit. The FPGA then corrects and displays the optical power based on the wavelength and direction of the incoming laser obtained by the laser warning system, and quantitatively marks it with an optical power meter to meet the linear response range of the photodiode, It can be concluded that there is a linear relationship between the optical power and the collected voltage value. Under the condition of not using an external optical attenuator, this circuit achieves the measurement of optical power at different incident angles in the spectral range of 400 nm ~ 1 700 nm and the power range of 0~60 mW. Compared with the results measured by the power meter, the error is less than 0.5 mW.

In order to achieve real-time instance segmentation and ranging of power transmission lines, a realtime power transmission lines point cloud segmentation (MPPS) method based on light detection and ranging (Li-DAR) mobile scanning is proposed. This method first designs a point cloud fast stitching enhancement algorithm that employs a Kalman filtering strategy with sliding spatial window for dynamic point cloud registration. Then, a 3D point cloud semantic classification network tailored for power transmission targets is designed. The network handles large spatial targets through uniform sampling and local feature aggregation (LFA). Finally, an overhead projection and a fast euclidean clustering algorithm are applied for power transmission line target segmentation and ranging. Experiments demonstrate that this method achieves a classification accuracy of 94.7% and an average intersection over union of 81.6% on a 3D point cloud dataset of power transmission lines obtained from LiDAR mobile scans, validating its ability to achieve real-time instance segmentation and distance measurement for power transmission targets.

The reliability of rotary drilling rigs often hinges on the welding quality of their critical component, the mast. X-ray testing is employed to inspect key welds on the mast, but due to objective constraints, the acquired Xray images frequently contain significant noise. Traditional Gaussian filtering methods, plagued by issues such as blurred edges and insufficient versatility, often fall short. To address these deficiencies, this study introduced three innovative Gaussian filtering techniques: adaptive Gaussian filtering, Gaussian pyramid-based and image fusion Gaussian filtering, and guided Gaussian filtering. Analysis using a series of quality assessment metrics demonstrated that Gaussian pyramid-based and image fusion Gaussian filtering, as well as guided Gaussian filtering, offer superior performance. These results provide new avenues for enhancing traditional Gaussian filtering effects and improving the subsequent accuracy of weld seam identification in rotary drilling rigs.

An all-optical quartz-enhanced photoacoustic spectroscopy (QEPAS) system for the highly sensitive detection of acetylene trace gas was proposed. An erbium-doped fiber amplifier (EDFA), a right-angle prism and a fiber Fabry-Perot interferometer (FPI) for out-of-plane one-way vibration detection were employed to improve the detection sensitivity. A self-stabilization technology was used to further improve the stability of the system. The vibration of commercial quartz tuning forks was simulated by the finite element analysis. Parameters such as laser modulation coefficient, structural size of the acoustic microresonator and output power of laser were optimized. A C2H2 minimum detectable limit (MDL) of 2.9 ppb was achieved, with a corresponding normalized noise equivalent absorption coefficient of 2.4×10-9 cm-1·W·Hz-1/2. The 2f signal amplitude shows a good correlation with the C2H2 concentration, with a linear correlation coefficient of 0.999. Finally, a 1-hour continuous experiment verified that the self-stabilization technology can effectively improve the stability of the system.

Lane detection is a key technology in intelligent driving. It is of great significance to detect lane position quickly and accurately to improve the safety of driving vehicles. Therefore, an improved lane detection method based on row direction position classification is proposed. The coordinate attention mechanism is integrated into the feature extraction backbone network to enhance the weight of effective positions in the feature map. Secondly, ELAN module and MP subsampling module are introduced to improve the feature extraction capability of the model. In reasoning, the idea of structure re-parameterization is used to fuse convolution and BN layer to speed up reasoning. In order to verify the performance of the improved model, the improved model was tested on TuSimple and CULane two classical lane data sets, and the detection accuracy was increased by 0.09% and 2.5% respectively compared with the original model, which verified the effectiveness of the improved model.

Aiming at the problem of degradation of inversion accuracy due to noise entrainment in the second harmonic signal, which is the detection signal of gas concentration in the tunable semiconductor laser absorption spectroscopy (TDLAS) technique. Proposing a denoising algorithm based on the combination of the Improved Northern Goshawk Optimization (INGO) algorithm, Variational Mode Decomposition (VMD) and Improved Wavelet Thresholding Method. The INGO algorithm, which minimizes the value of the envelope entropy as an objective function, is used to find the number of modes k and the penalty factor , which are important parameters in the VMD. Filtering the effective signal IMF components by the fact that the product of the energy density and the average period of each intrinsic mode function (IMF) after decomposition is a constant. The joint wavelet improvement thresholding method is used to reduce the effective signal and combined to obtain the denoised signal. Comparative analysis of simulation experiments shows that the INGO-VMD-improved wavelet thresholding method can effectively filter out the noise signal, reduce the amplitude error, improve the inversion accuracy, and the noise reduction effect is superior. The waveform similarity coefficient after noise reduction is 99.86%, the signal-to-noise ratio reaches 25.230 5 dB, and the root mean square error reaches 0.015 23%. The peak error decreases by 0.157 compared to the noise-containing signal.

To realize a biometric system with high security and good user acceptance, a multispectral palm vein image acquisition device based on an open environment was designed, and a palm vein recognition algorithm based on double constraints of principal component analysis (PCA) and least squares regression (LSR) was studied. In the process of least squares regression projection, the algorithm constrains the main element information extracted by principal component analysis, jointly driving the data and weakening the adverse effects of light scattering on recognition performance. It solves the problem of increased intra-class spacing caused by non-contact image acquisition. Experiments were conducted on palm vein databases of the Institute of Automation, Chinese Academy of Sciences, Tongji University, Hong Kong Polytechnic University, and the self-built, and the algorithm's lowest equal error rates were 0.72%, 0.50%, 0.18%, and 0.03%, and the correct recognition rates were 99.80%, 99.77%, 99.90%, and 99.95%, respectively. Compared with other typical methods, the system has advantages and practical application value.

A solution is proposed for human presence recognition in indoor scenarios using FMCW radar. There are often issues of target loss or interference with small moving objects in the environment when it comes to recognizing stationary individuals. Firstly, a Range Doppler Map (RDM) is constructed using the sampled data from FMCW radar, and the precise distance subset corresponding to the target's location range is calculated through cumulative projection. Then, a high-precision micro-Doppler time-frequency feature map is obtained based on the distance subset using Short-Time Fourier Transform (STFT), from which 10 meaningful feature vectors are extracted. Finally, experimental data combined with Support Vector Machines (SVM) is used to screen different combinations of feature vectors. The experimental results show that there exists an optimal combination of 5 feature vectors, achieving a recognition rate of 97.98% for overall recognition of stationary individuals, environmental disturbances, and walking individuals in indoor environments. This method not only improves the performance of the classification model but also enhances its interpretability.

Photo-Response Non-Uniformity (PRNU) is an inherent feature reflecting the defects of imaging sensors, which can effectively identify the source of cameras shooting digital video. Aiming at the problem of poor recognition effect of network compressed video, a multi-scale neighborhood value shrinkage filtering algorithm based on Stein 's unbiased risk estimation and adaptive edge structure preserving smoothing filtering algorithm was proposed, and a weighted PRNU extraction model was constructed. Firstly, a multi-scale transformation based on the dual-tree complex wavelet was performed on video frames that skip loop filtering. Then multi-scale neighborhood value shrinkage filtering algorithm based on Stein's unbiased risk estimation was used to estimate all high-frequency subbands. After obtaining noise residuals, adaptive edge structure preserving smoothing filtering was used to smooth the complex noise residuals. Then the noise residuals were aggregated using a maximum likelihood estimation method based on quantization parameter weighting to obtain the multiplicative factor of PRNU. Finally, PRNU was obtained through preprocessing. The experimental results on the Vision dataset show that when the video duration is 15 seconds, the AUC values of the proposed model under the moving and rotating reference fingerprints are 0.955 1 and 0.954 9, and the Kappa coefficients are 0.840 3, 0.888 9 and 0.913 2, respectively, which are superior to the existing algorithms.

Fiber optic pulse coherent optical communication systems are subject to multipath interference in information transmission, resulting in poor detection and recognition capabilities for network abnormal intrusions. In order to improve the accurate identification of abnormal intrusions in optical communication systems, a clustering analysis based online identification method for abnormal intrusions in optical communication systems is proposed. Firstly, the cellular sensing distributed detection technology is used to realize the packet collection of the transmission data of the optical communication system. Secondly, the collected data is preprocessed to remove multipath noise interference. Then, the Ensemble learning algorithm is used to extract the characteristics of the abnormal intrusion data of the optical communication system, and calculate the relevant parameters of the abnormal intrusion data of the optical pulse coherent optical communication system. Finally, the extracted characteristics are clustered, Through Spectral clustering analysis, the online identification of abnormal intrusion in fiber optic pulse coherent optical communication system is realized. The simulation results show that the proposed method has high efficiency in identifying abnormal intrusions, takes 5 seconds to identify, maintains recognition accuracy at around 98%, consumes 38 J of energy, is less affected by disturbances, and has strong resistance to harmonic interference.

A curved multi groove assisted 6-core fiber structure has been designed, where each core in the multi core fiber is surrounded by three identical low refractive index grooves, and there is a layer of cladding between the core and each layer of groove. The coupling mode theory of optical fibers was applied to analyze the two transmission characteristics of crosstalk and mode field area between bent six core optical fibers. The results show that the bending radius and groove have a significant impact on the crosstalk and mode field area of the optical fiber, respectively. After adding grooves, the crosstalk between fiber cores significantly decreases. Under different bending radii, the change in crosstalk between fiber cores is about 10 dB. At the same time, the change in mode field area under different layers of grooves is about 150 m2.

The array of grating units consists of a large number of units, and the size and shape of each unit need to be precisely controlled to ensure the shaping effect of the beam. However, in the process of optical communication, the change of light intensity distribution will cause the response of each unit in the grating array to be inconsistent, which will cause the difference of the optical signal received by each unit, and then affect the shaping effect of the beam. Therefore, a beam shaping method for optical communication grating array considering light intensity distribution is proposed. Taking the most commonly used array waveguide grating in optical communication grating arrays as an example, by analyzing the core width, core refractive index, cladding refractive index, waveguide period width and other parameters of the array waveguide structure, combined with the influence of beam wavelength and refractive index on the propagation constant in the waveguide, the normalized eigenequations of the array waveguide are obtained. The results are combined with the uniform Lorentz function to obtain the intensity distribution of the incident and outgoing light, By calculating the mapping functions of the incident and outgoing light of the optical communication grating array beams, the optical parameters can be adjusted using the functions to achieve beam shaping of the optical communication grating array beams. The experimental results show that this method has good shaping effect on the optical communication grating array beam, high energy utilization efficiency of the shaping beam, and small beam non-uniformity.

The Internet of Things usually consists of a large number of sensor nodes, which are distributed over a wide range of areas. Faced with a vast network topology, how to efficiently calculate and select the optimal communication line is a challenge. The optimal communication line selection involves multiple optimization objectives, such as transmission distance, energy consumption, etc. Taking into account the above factors, a method for selecting the optimal communication line of laser sensor nodes in the Internet of Things is proposed. Select inflection nodes in sensor nodes to reduce the number of transmission times of sensor nodes. Taking the minimum energy consumption and shortest path as path selection objectives, a routing algorithm is introduced to obtain the optimal communication line from multiple paths while considering node energy balance. After experimental verification, the proposed method can effectively find the shortest transmission path with low energy consumption and high data integrity in data transmission, indicating that the application effect of this method is good.

Visible light communication networks are one of the main components in the current communication field, but their physical devices (LED lighting devices) have nonlinear characteristics, resulting in nonlinear distortion of visible light communication network signals, which affects the quality of user communication. Therefore, an adaptive correction method for nonlinear distortion of visible light communication network signals is proposed. Through in-depth analysis, it can be concluded that signal nonlinear distortion is divided into nonlinear conversion distortion and nonlinear limiting distortion. Based on this, a signal nonlinear distortion correction circuit (DG pre distortion correction circuit and DP pre distortion correction circuit) is designed, and an adaptive algorithm is introduced to develop an adaptive correction architecture for signal nonlinear distortion. The adaptive correction formula for signal nonlinear distortion is determined, thus achieving effective correction of signal nonlinear distortion. The experimental data shows that there is no nonlinear distortion phenomenon in the adaptive correction results of signal nonlinear distortion obtained by applying the proposed method, and the minimum error rate of communication signal reception is 3%, fully confirming that the proposed method has better application performance.

Optical networks have been widely used due to their advantages such as large capacity, high transmission rate, transparent business, and low loss. With the increase of user scale and transmission needs, the frequency of insufficient traffic in optical networks has increased, which has constrained the development of optical networks. Therefore, a cross source scheduling method for optical network traffic big data based on clustering is proposed. Firstly, cluster clustering algorithm is used to process big data of optical network traffic. Secondly, the XGBoost model is used to predict the next moment of optical network traffic. Build a mathematical model for cross source traffic scheduling with the goal of minimizing costs, and determine the constraints of the constructed model. Finally, using genetic algorithm as a tool, obtain the optimal solution for cross source scheduling of traffic, and execute the optimal solution to achieve cross source scheduling of optical network traffic big data. The experimental results show that the clustering results of optical network traffic big data obtained by the proposed method are consistent with the expected clustering results of optical network traffic big data. The minimum cost of cross source traffic scheduling is 160 000 yuan, indicating that the proposed method has better performance in cross source traffic scheduling.

In order to improve the security and accuracy of information transmission for fiber optic sensor nodes, a fiber optic sensor node information secure transmission technology based on NB IoT technology is proposed. Construct a NB IoT channel structure model for information transmission of fiber optic sensor nodes, use carrier activation modes (SAPs) to achieve carrier modulation and anti-interference suppression during the information transmission process of fiber optic sensor nodes, construct a hybrid mapping scheme to achieve secure mapping encryption design during the information transmission process of fiber optic sensor nodes, and introduce index modulation technology into the orthogonal frequency division multiplexing system for fiber optic sensor node transmission, Combining parity check to achieve channel allocation and transmission bit rate localization detection for fiber optic sensor nodes. The simulation results show that the proposed method has better security and encryption performance in the information transmission of optical fiber sensor nodes, the maximum transmission prediction error is only -2.0, and the transmission bit error rate is the highest 0.20×10-10, the average is only 0.17×10-10, and when the anti-interference ability is tested, the signal-tonoise ratio of this method reaches more than 49.8 dB, which improves the transmission stability and anti-interference ability of optical fiber sensor nodes.

The binocular structured light 3D imaging technology can obtain the height distribution of the measured object by projecting structured light and then obtaining 3D information about the object. It can be widely used in industrial production, medical research, and other fields. However, due to the interference of unstable factors such as ambient light and optical lens errors, the widely used three-dimensional imaging technology based on Gray phase shift technology still needs fixing, such as decoding errors and inaccurate matching. This paper proposes an improved algorithm based on existing technology. The algorithm uses mutually exclusive complementary Gray codes for projection, uses a bidirectional averaging decoding method to unwrap the phase, and generates high-quality point cloud data through mask-constrained stereo matching. After experimental analysis and verification, the algorithm in this paper has improved its accuracy by 11.49% compared to traditional algorithms, and the average number of point sets has increased by 14.82%. This proves that the algorithm can effectively suppress optical noise interference and improve the accuracy of 3D imaging.

Chest X-ray (CR) images are crucial for diagnosing chest lesions. To overcome the issue of insufficiently extracting the relationship between disease features and disease dependency in multi-label disease classification tasks, this paper proposes a chest X-ray multi-label disease classification model based on Vision Transformer (CDCViT). Firstly, Efficientnet-B0 is used as the feature extractor to extract the feature map. After mapping the feature maps, patch embedding and position embedding are added and input into the Transformer module. The Transformer network calculates the weight matrix between features to better mine the relationships between disease features. Then, through Mutual Attention Weight Selection (MAWS), feature selection is performed on the feature tokens collected by multiple Encoder modules, selecting the features most conducive to classification. Finally, the classification results are mapped through a fully connected network. In addition, the ASL loss function is used to calculate the differences between the labels for backpropagation to optimize the model parameters. The proposed model is applied to the public dataset ChestX-ray14. The experimental results show that the CDC-ViT model achieves an average AUC of 0.822 8 for 14 chest diseases, which is about 2% higher than the comparison model, indicating that the proposed CDC-ViT model is superior to many existing classification models.

Spaceborne laser spot images are susceptible to the influence of complex light backgrounds during collection, resulting in poor accuracy in extracting laser spot feature parameters. Therefore, a method for extracting feature parameters of spaceborne laser spot images in complex light environments is proposed. Firstly, the Retinex image enhancement algorithm based on illumination correction implements illumination compensation and enhancement processing on the spaceborne laser spot image to eliminate the impact of complex lighting environments; Secondly, based on the combination of secondary threshold segmentation and morphological denoising, the extraction of light spot boundaries is completed; Finally, a combination of first-order grayscale centroid method and ellipse fitting method is used to extract the feature parameters of laser spot images. The simulation results show that the proposed method maintains a feature parameter extraction accuracy of over 95% for spaceborne laser spot images, and has good extraction performance.

In order to solve the problem of color anomalies in digital video images obtained by photoelectric tracking equipment, a design and research on the automatic color matching system for digital video images of photoelectric tracking equipment is proposed. Design a digital video image acquisition module for photoelectric tracking equipment, collect and store digital video images of photoelectric tracking equipment, apply filtering algorithms to remove image noise, use Mean Shift algorithm to segment and extract image color information, debug and process the color boards of corresponding color blocks, and achieve automatic color matching of digital video images. The experimental data shows that under different experimental conditions, the maximum completeness of color information extraction in the design system image is 100%, and the automatic matching results of image colors are closer to the color information in the template image, fully confirming that the design system has better application performance.

Holographic laser projection degraded images contain various factors such as noise and uneven illumination. In order to optimize image quality and improve image clarity, a holographic laser projection degraded image phase recovery technology is proposed. Firstly, the wavelet threshold method is used to remove noise from degraded holographic laser projection images containing noise; Secondly, combining the Retinex algorithm with a Gaussian low-pass filter to estimate the incident component of the holographic laser projection degraded image, using a partitioned local color strip mapping algorithm combined with the incident component to enhance the holographic laser projection degraded image; Finally, the phase restoration method based on the cartoon texture model is used to repair degraded images and achieve holographic laser projection for image restoration. The experimental results show that the proposed method has good denoising and enhancement effects on the image. At the same time, the phase recovery accuracy is higher than 97.0%, and the SSIM value reaches 0.9, the MSE value is lower than 0.9%, and the peak signal-tonoise ratio is higher than 7.14 dB. It has a better phase recovery effect on holographic laser projection degraded images.

To improve the effectiveness of laser image restoration, a laser image visual communication restoration technology is designed in complex environments. Firstly, design a concave word segmentation noise suppression algorithm and implement laser image denoising processing. Then, in response to the complex imaging environment of laser images, a single pixel observation model of the laser image is constructed after denoising to perform reconstruction processing on the laser image, thereby achieving better visual communication and repair effects. Finally, a visual communication restoration model is designed based on RSM-Net to achieve image restoration function. The test results show that the repair effect of the fracture area and the light spot area of the design technology is very excellent, with a repair satisfaction rate of over 94.2%, which is superior to the comparison method. It can achieve complete repair of the fracture area, with overall clarity and no noise, and the application effect is good.

Due to the strong complexity and high noise of reflected light interference in laser images, the accuracy and efficiency of multi-level feature point recognition in laser images are low. Therefore, a multi-level feature point recognition study for laser images with reflected light interference is proposed. Calculate the slope of adjacent points based on the characteristics of reflected light interfering with laser images, and screen effective laser data. Using data curvature algorithm to change the distance between reflected light interfering laser images, eliminate effective reflected light interfering laser image noise, and complete filtering and smoothing processing. Using Mask R-CNN to detect the effective laser data noise of filtered reflected light interference laser images, the loss function is used to extract feature point information at different levels of Mask R-CNN, thereby completing multi-level feature point recognition of reflected light interference laser images. The test results show that the proposed method has good multi-level feature point recognition results, and can recognize all multi-level feature points. The maximum misjudgment rate for effective laser data filtering is 0.4%, and the multi-level feature point recognition rate is higher than 95.25%. The maximum recognition time is 22 ms.

Laser radar requires accurate calibration before operation, otherwise it can easily cause distortion in the captured images. To ensure the quality of shooting and address the issue of insufficient accuracy in traditional calibration methods, a target self calibration method for unmanned aerial vehicle (UAV) airborne LiDAR imaging is studied. Implement denoising processing for target point cloud images captured by airborne LiDAR. Construct a point cloud matching model using the KCRNet network in deep learning to achieve feature point cloud matching. Based on feature point pairs, construct a calibration parameter optimization model. Using genetic algorithm to solve the model, obtain the optimal calibration parameters, and complete the target self calibration of unmanned aerial vehicle aerial LiDAR images. The results show that after calibration, the gross error rate is relatively smaller, the highest value is only 2.6%, the average calibration accuracy is 99.2%, and the average calibration time is only 5.0s, which indicates that the calibration effect of the method is better and the image quality is higher.

With the development of remote sensing technology, the application of LiDAR remote sensing images in various fields is becoming increasingly widespread. However, there are often various defects in these images, such as noise, distortion, occlusion etc. which can have adverse effects on the analysis and application of the images. Therefore, it is necessary to implement defect segmentation on LiDAR remote sensing images. The aim of this study is to develop a defect segmentation model for LiDAR remote sensing images based on visual communication technology, in order to improve the accuracy and reliability of remote sensing image processing. The application of bilateral filtering functions for discrete point cloud denoising of LiDAR remote sensing images preserves the surface geometric features of the point cloud well while denoising. Based on the image enhancement technology in visual communication technology, LiDAR remote sensing image enhancement processing is implemented to improve the understanding and readability of visual information in remote sensing images. The selected image enhancement technology is the local contrast enhancement variational model. Design a semantic segmentation network that integrates attention mechanism as a remote sensing image defect segmentation model to achieve defect segmentation in LiDAR remote sensing images. The experimental test results show that the mPA of the designed model is relatively high, overall higher than 95%. As the amount of test data increases, there is no significant decrease in the mPA of the designed model. The design model has a high MIoU for different terrain scenes and can maintain a high MIoU for complex terrain scenes, with strong robustness.

Aiming at the optimization of laser cutting process parameters for aluminum alloy, laser cutting experiments were conducted on 5052 aluminum alloy. This article establishes a three factor orthogonal experimental plan based on the influence of different laser power, auxiliary air pressure, and operating speed parameters on laser cutting quality. A 0.3 mm and 0.5 mm thick thin aluminum alloy was selected for laser cutting, and 10%, 20%, and 30% orthogonal operating speeds were designed with 1 MPa and 1.5 MPa argon gas, and laser power of 1 200 W and 1 350 W, respectively. Two quality parameters, roughness and slag length, were observed by a light cutting tester and microscope, and data analysis was conducted to determine the optimal laser cutting process route for 0.3 mm and 0.5 mm thin aluminum plate, providing technical reference for laser cutting of aluminum alloy thin plates.

Cold rolls made of Cr12MoV are subjected to fatigue stress and thermal cycling during the rolling process, resulting in defects in the form of spalling and cracks on the roll body. In order to solve this problem, a broad -beam laser melting equipment is used to prepare an alloy coating on the surface of Cr12MoV steel by using Cr12MoV alloy as the cladding powder. Taking the laser power, scanning speed, powder feed rate and overlap ratio as variables and the surface roughness and microhardness of the cladding layer as the optimized target parameters, orthogonal tests are carried out and the optimized combination of the variable parameters is obtained by using the methods of range analysis and variance analysis to analyze the target parameters. The results show that the optimum variable parameters combination is laser power 2 000 W, scanning speed 10 mm/s, powder feed rate 10 g/min and overlap ratio 45%. The cladding test is conducted under this variables combination, the obtained surface roughness is 16 m, and the average microhardness of the coating is 677 HV which reaches 96% of the value of the base material, and the cladding layer has a uniform microstructure without defects such as porosity and cracks.

With the rapid development of the fourth generation reactor, the requirements for wear and corrosion resistance of key equipment in the reactor are becoming increasingly strict. The metal ceramic composite coating combines the good wear and corrosion resistance of ceramics with the good ductility and thermal conductivity of metals, and has a wide application prospect in the field of key reactor equipment. This article uses laser cladding technology to prepare metal ceramic composite coatings of Al2O3 and ZrO2 ceramics with 316L stainless steel. The change rate of Al2O3 and ZrO2 ceramic composition in different composite coatings is 20%. The microstructure, Rockwell hardness, and wear performance of the prepared composite coating material were analyzed using optical microscopy, electric Rockwell hardness tester, and friction and wear testing machine. The experimental results show that the metal ceramic bonding effect in the prepared composite coating is good, and there are no obvious defects inside the material. The hardness of metal ceramic functionally gradient materials is positively correlated with the content of ceramic components. As the content of ceramic components increases, the overall hardness and wear resistance of the material correspondingly improve.

Laser plasma spot center extraction due to the poor effect of image noise reduction, the center extraction under complex background is inaccurate, so a new subpixel extraction method of laser plasma spot center under complex background is designed. In this method, the multi-graph average method and Gaussian filter algorithm are combined to process the laser plasma spot image with noise reduction and complete image pre-processing. Zernike moment method is introduced to extract the laser plasma spot center, and the circle fitting method is adopted to optimize the algorithm to achieve the subpixel extraction of the spot image center. The experimental results show that when the proposed method is used to extract the laser plasma spot center subpixel under complex background, the image of the spot center extracted by the proposed method is more clear, and the signal-to-noise ratio reaches more than 47.0dB, and the absolute error of the extracted center position is only 1mm, this method has high extraction accuracy and better image quality.

In the application process of laser technology, the detection and determination of spot position is a key link, which directly determines the application performance of laser technology. However, due to the limitations of application methods, existing spot position detection methods have problems such as long detection time and low detection accuracy. Therefore, a research on laser spot position detection based on digital image processing technology is proposed. Using image acquisition devices (cameras or cameras) to obtain laser spot images, processing the laser spot images based on digital image processing technology. Based on this, the two-dimensional Otsu algorithm is used to segment the laser spot images. The Canny operator is applied to extract the edge of the spot, and the minimum twice circle fitting method is used to fit and process the edge of the spot. The center position of the laser spot can be calculated, thus achieving accurate detection of the laser spot position. The experimental results show that the minimum detection time for spot position obtained by the proposed method is 0.1 s, and the minimum detection error for spot position is 0.2%, fully confirming that the proposed method has better application performance.

Prolonging the service time of CNC machine tools will lead to the deviation between the machining trajectory and the actual trajectory contour, and there are defects such as poor machining efficiency. In order to obtain the ideal machining effect of CNC machine tools, a control method of machining error of CNC machine tools based on laser tracker is proposed. Firstly, the laser tracker is used to determine the machining trajectory of CNC machine tools, and the cross-coupling controller is introduced to calculate the estimated trajectory error in the machining process of CNC machine tools. Then, the compensation control quantity of machining trajectory contour error is determined according to the error distribution result, and the machining control of CNC machine tools is carried out according to the interpolation value. Finally, the simulation test is carried out. The results show that this method reduces the machining error of CNC machine tools and improves the machining accuracy of CNC machine tools, which has high practical application value.

Aiming at the problem of poor precision of vehicle intelligent positioning at present, this paper puts forward the research of vehicle intelligent positioning method with real-time data acquisition by lidar. Firstly, the point cloud data around the vehicle are preprocessed by lidar, and the interference point cloud data are denoised by improved C-means method. Secondly, the vehicle surrounding information is clustered under the improved k-means algorithm. Finally, the vehicle intelligent positioning is realized by the global positioning method of laser radar bounded area. The experimental results show that the proposed method has higher precision and shorter time, and is more suitable for practical application.

the operating environment of electromechanical equipment is complex, and current methods cannot obtain high-precision fault identification results of electromechanical equipment. In addition, the fault identification time of electromechanical equipment is long and the real-time performance is poor. In order to obtain more ideal fault identification results of electromechanical equipment, a laser sensor based signal acquisition method for electromechanical equipment fault identification is designed. Firstly, a laser degree sensor is used to collect the working status signals of electromechanical equipment, and the working status signals of electromechanical equipment are preprocessed to extract relevant features for fault identification. Then, the features are used as inputs to the machine learning algorithm, and the types of electromechanical equipment faults are used as outputs. A classifier for electromechanical equipment fault identification is established through training. Finally, the performance of electromechanical equipment fault identification is analyzed through specific simulation experiments. The results show that the method proposed in this paper can identify faults in electromechanical equipment with an accuracy of over 95%. The identification time of electromechanical equipment is controlled within 5 seconds, and the overall identification effect is better than the current typical electromechanical equipment fault identification method.

The problem of urban light pollution is becoming increasingly prominent and has become one of the key areas for urban environmental pollution prevention and control. Drawing on advanced governance experience from abroad, utilizing satellite technologies such as DMSP-OLS operating line scanning system, NPP-VIIRS, Loujia1-01, Jilin1-03B, EROS-B, and CS2000A spectral radiometer for nighttime light environment monitoring, targeted lighting management is achieved based on population density and nighttime light source intensity. Using statistical and image preprocessing techniques to preprocess monitoring data, conducting data analysis using PASW statistical software, MATLAB software, and modeling methods to identify areas of light pollution. Further control of light pollution through source control, reduction of light intrusion, and dark space operations. On this basis, combined with the current situation and problems of light pollution control in China, specific measures will be taken from three aspects: based on reality, promoting technological innovation, and enhancing environmental awareness, in order to effectively control light pollution and create a green and harmonious light environment in collaboration.

In order to solve the problem of light scattering, reflection or transmission caused by surface defects of optical components, which affects the imaging quality of optical systems, a surface defect detection method for optical components based on multispectral technology is studied. Collect multispectral images of surface defects on optical components, Laplace pyramid transformation algorithm is used to fuse multispectral images of surface defects in optical components and frame difference method is used to subtract two multispectral images of optical components and obtain binary image features of surface defects of optical components. Mark the connected regions of the multispectral image fusion results of optical components, and use parameters such as the aspect ratio of the image connected regions to detect the types, quantities, and sizes of surface defects on optical components. The experimental results show that the method can effectively detect different types of surface defects such as pitting, scratches and broken points, and the detection accuracy of surface defects is the highest 98.15%.

In low light conditions, images tend to become blurry and dim, and vehicles and other traffic signs become less recognizable. This makes vehicle detection algorithms more difficult, increasing the likelihood of false and missed detections. For this reason, a real-time automatic vehicle flow detection method is proposed for road intersections under low light conditions. Multi-dimensional attention mechanism and recursive pyramid network are embedded in U network architecture, and combined with regional feature aggregation network, the regions of interest in low-light intersection images are screened. Based on this, the deep reinforcement learning model is used to make the detection frame fit the target vehicle, and the complete detection frame fit the vehicle is obtained through the regression fine adjustment of the multi-layer fully connected network, and the real-time automatic detection of vehicle flow is completed. The experimental results show that the proposed method has accurate and effective detection ability for vehicles, reflective objects and luminous objects in the distance and dark under various weak lighting conditions, and provides a reliable reference for traffic control.

As a kind of intraoperative navigation technology, near infrared fluorescence imaging has become more and more important in the biomedical field. Its non-invasive, high sensitivity and high resolution characteristics make it a useful assistant in biomedical research. Although traditional biooptical imaging techniques such as IR-I can provide rich physiological and pathological information, they still have limitations in some aspects, such as high cost, complex operation and low detection efficiency. Therefore, a more advanced, efficient and convenient imaging technology is urgently needed to meet the needs. It is in this context that the fluorescence imaging technology of IR-II came into being. Because of its unique tissue penetration depth and low background noise, near infrared two region (NILII) fluorescence imaging technology has shown great potential and broad application prospects in the field of biomedical imaging. Based on this, the background principle of NIL-II fluorescence imaging technology and the challenges facing the development of NIL-II fluorescence probes will be presented, and the application of indocyanine green (ICG) as a NIL-II fluorescence probe in the medical field will be introduced, and the practical application of NIL-II fluorescence imaging in tumor diagnosis and treatment will be introduced.