This PDF file contains the front matter associated with SPIE Proceedings Volume 13080, including the Title Page, Copyright information, Table of Contents, Introduction and Conference Committee information.

A non-destructive security inspection system for building curtain walls based on millimeter wave imaging was designed to address the difficulty in detecting internal components of building curtain walls. This system is based on the principle of active millimeter wave imaging, which can scan the components inside the curtain wall and image them in the upper computer. It mainly includes a terahertz wave irradiation source and receiver, as well as a point-to-point imaging system consisting of a pair of 90 ° off axis parabolic mirrors and a semi-transparent and semi-reflective mirror. In order to simplify the system structure, a scheme of transmitting echoes along the same path is proposed. This scheme has a simple structure and can effectively solve the problem of the echo signal being affected by the reflected waves generated by the upper and lower surfaces of the curtain wall and the oscillating reflected waves generated between the lower surface of the curtain wall and the metal components during the process of millimeter wave penetration through the curtain wall. It can also achieve point-to-point clear imaging with an imaging accuracy of approximately 3 mm.

Chemical Oxygen Demand (COD) serves as a pivotal parameter for assessing water quality in wastewater. Traditional chemical detection methods are time-consuming and prone to secondary pollution of the environment. Near-infrared spectroscopy technology as an alternative to traditional chemical detection methods is a viable approach. However, the experimental data shows data overfitting and high interference within the sample dataset parameters，which has poor applicability in traditional quantitative and qualitative analysis methods. In this paper, a one-dimensional Convolutional Neural Network (1D-CNN) method is used to establish a relationship model between near-infrared features of water samples and the COD, which has the advantages of moderate depth, small parameter amount, and fast network training speed. It is extremely suitable for rapid detection of samples which can effectively solve the above problems. Experimental results show that the proposed model is better than random forest classification, SGD Classifier, and support vector Machine regression(SVM).

Infrared camera is a photoelectric device that obtains temperature by detecting thermal radiation. With the need of multi-sensor fusion application, the accurate acquisition of camera parameters has become an important topic. Based on Zhang Zhengyou's calibration method, infrared thermal radiation is used to enhance the reflectivity of the calibration plate. The asymmetric circle is used as the calibration pattern and the circle center is used as the calibration feature. The pre-processing method of calibration image is designed by contrast experiment. The preparation of calibration image is completed through two steps of image contrast enhancement and image binarization under uneven illumination. The error rate of infrared camera calibration reaches 0.04%. This method avoids complex active infrared calibration plate making, improves the problem of difficult extraction of corner features in infrared images, and solves the infrared calibration problem with simple and easy to obtain tools, and obtains sufficient accuracy

Chirality is a pervasive phenomenon in nature. The examination of chirality in biomolecules provides a means to detect their intricate structures. Certain pharmaceutical molecules necessitate precise chirality detection, given that variations in chirality can lead to disparate therapeutic outcomes. Unfortunately, the circular dichroism signal used to discern chirality is notably feeble. Consequently, the implementation of surface-enhancement techniques becomes imperative to amplify this signal and facilitate accurate chirality detection. Induced circular dichroism is a phenomenon generated by the near-field interaction between chiral molecules and nano-structures. Induced circular dichroism enhances detection sensitivity of molecular chirality, and to enhance induced circular dichroism, a detailed investigation of underlying mechanism of circular dichroism inducement is needed. In this paper, to enhance induced circular dichroism, we present numerical investigation on circular dichroism inducement mechanism. The induced circular dichroism was observed to be enhanced by asymmetric electric and magnetic dipole resonance modes. In contrast, for symmetric electric and magnetic dipole modes, the induced circular dichroism is minimal.

In this paper, the cooling calculation and analysis of the flat plate structure containing internal heat source is carried out, and a relatively complete calculation model consisting of continuous phase control equation, discrete phase equation and mass, momentum and energy conservation equations of liquid film is established, which realizes the calculation simulation that can be applied to the cooling heat transfer of water spray in large space. Calculations show that the trend of calculated results in this paper and the experimental data are consistent. The maximum difference between the average water film temperature and the measured temperature on the infrared imager is 1.9°C, and the maximum difference between the calculated temperature of the flat plate and the 9-point average temperature of the thermocouple is 0.9°C. The water spray cooling of the flat plate is calculated to reduce the temperature of the experiment plate by 16.5℃ in 34 seconds, and the infrared radiance of 3~5μm is reduced by 45.5%, the infrared radiance of 8~12μm is reduced by 24.2%. The water spray cooling has a significant effect on the infrared characteristics of the plate.

The MLVDS interface can support up to 500Mbps data rate and a wide common mode voltage range, and has strong ESD protection characteristics, supporting hot swapping function. MLVDS solves electromagnetic interference (EMI) problems by controlling the voltage swing rate and output amplitude of output data. In addition, MLVDS also retains LVDS low-voltage differential signal characteristics, which can further reduce electromagnetic interference and drive multiple transceivers to achieve interconnection applications, Featuring high transmission rate, low power consumption, and low noise. The main control system of the Ka simulated phased array antenna wave control system controls multiple sub wave control components through the MLVDS bus. The MLVDS bus between the main control and sub wave control adopts a one-to-eight structure, and the MLVDS transmission rate is required to be 50Mbps, with reliable and stable transmission. From the perspective of reliability and derating design, the MLVDS bus of the system is required to transmit stably at a transmission rate of 100Mbps. Based on this, establish a testing platform to verify the reliable transmission of MLVDS one to eight bus at a transmission rate of 100Mbps.

With the development of scientific research, the stability of synchrotron radiation has been paid more attention. The liquid vibration will change the liquid flow state, cause the vibration of the pipe surface, and lead to the crystal jitter. Aiming at the stability requirements of the high-stability monochromator of the partial beam line of SSRF, ANSYS workbench software was used to analyze and optimize the structure, and a cooling pipe system with more stable structure was designed. This paper also analyzes the effect of cooling system vibration on crystal. The test results of the prototype show that the resolution of the device can reach 1 urad and the repetition accuracy is less than 1.071 urad. All the indexes meet the needs of the monochromator.

In this paper, we proposed a novel isolation-based hyperspectral anomaly detector using nearest neighbor ensemble (iNNE) based on the premise that anomaly pixels are more prone to isolation compared to the background. The approach serves as an effective anomaly detector relying on nearest neighbors and isolation. iNNE demonstrates a notable enhancement in computational efficiency compared to established nearest neighbor-based methods, such as the Local Outlier Factor (LOF), especially when applied to datasets with thousands of dimensions or millions of instances. This improved efficiency is attributed to the method's linear time complexity and constant space complexity. In contrast to existing tree-based isolation methods like iForest algorithm, the proposed approach effectively addresses the challenges of detecting local anomaly instances and anomalies in high-dimensional data. The experimental outcomes on four authentic hyperspectral datasets have illustrated that the proposed iNNE not only outperforms other state-of-the-art anomaly detectors in terms of performance but also proves to be quite competitive in terms of computational efficiency.

Antenna in package (AiP) is a technology that can realize the fabrication of low-cost, high-performance and small size antenna, and has a wide range of applications in mobile communication and other fields. The development trend of wave control system is large-scale and distributed system, miniaturization and integration of single entity, high reliability and low cost. Based on Ka phased array AIP antenna technology, this paper introduces the wave control principle and software and hardware design of the wave control system in detail, and realizes the miniaturization and easy expansion of the control system. According to the specific requirements of the system for beam scheduling, the phase shifter of each channel is controlled accurately and quickly, and the corresponding phase shifter is sent to the corresponding channel phase shifter in real time to realize the electronic scanning of the antenna beam. The design of the broadcast control system is miniaturized, integrated, high reliability and low cost. This solution adopts a highly integrated and scalable sub-array design scheme, the number of beams reaches 4, and the integration density reaches 6.25mm2/ unit/beam. The beam switching time is less than 100us, which meets the application scenario of high-speed beam switching.

The inverse problem is a problem as opposed to a positive problem, take a classic example, in mystery novels, the murderer’s crime process can only be traced after the occurrence, and the detection process is a typical inverse problem-solving process. With the development of theory, the Jacobi matrix inverse eigenvalue problem has been widely used in many fields. In this paper, we explore the eigenvalue inverse problem of an anti-tridiagonal matrix with double constraints based on the Jacobi matrix, i.e. an anti-tridiagonal matrix is given one of its eigenpairs, and the submatrix formed by crossing out the first row and the last column of the anti-tridiagonal matrix is given one of its eigenpairs. Then the non-zero elements of the matrix are inverted, and the existence and uniqueness of the solution to the given problem are obtained. In the end, we give two numerical examples to verify the correctness and validity of the solution.

Mainstream image semantic segmentation networks can only extract local features of an image. Their receptive field is limited to the range of convolutional kernel size, which impacts the distinguishability of extracted features by the encoder. This issue leads to inaccurate segmentation, loss of small targets, and other related problems. This paper presents a solution to the aforementioned challenges by proposing the use of a CBAMUNet image semantic segmentation model. The model combines CBAM attention module and UNet En-decoder network. Firstly, the UNet network has been enhanced to eliminate the need for cropping of input images to reduce feature loss. Secondly, the CBAM module has been integrated into its skip connection to enable the UNet network to extract more efficient features and enhance the segmentation effect. And the model uses a new loss function. Data augmentation is also used to improve the generalization ability of the model. Through the comparison experiments with UNet, it is found that CBAMUNet is able to significantly improve the segmentation effect of images, and increase the pixel precision(PA) and the mean intersection over union(mIoU) of the semantic segmentation of images.

Compared to ordinary light field image (LFI), multi-exposure fusion light field image (MEF-LFI) can record more visual information and details of scene. However, MEF-LFI also produces distortions while enhancing LFI, leading to quality degradation. Therefore, it is crucial to develop effective MEF-LFI quality assessment models. This paper proposes a multi-exposure fusion light field image quality assessment method with motion region detection, which considers that the artifact distortion of MEF-LFI synthesized from dynamic scenes usually occurs in motion regions. A motion region detection module is designed for detecting artifact distortion in MEF-LFI. Considering that tone mapping (TM) operations can cause texture distortion in MEF-LFI, the spectral texture distortion feature extraction module and the spatial domain gradient feature extraction module are designed by combining Curvelet transform and Scharr operator, respectively. Due to the distortion of color shift in MEF-LFI, the color feature extraction module is constructed with the characteristics of HSI color model. In addition, considering the unique angular distortion of MEF-LFI, the angular feature extraction module is designed with Log-Gabor operator. Finally, the extracted feature vector is input into the support vector regression model to predicate the quality for MEF-LFI. The experimental results show that the proposed method is superior to the representative quality assessment methods and has better consistency with the human visual perception.

Head-mounted eye-tracking instruments are important devices for sports and human factors research, and their calibration is a more tedious task due to the varied morphology of the human head as well as wear differences. In order to improve the calibration efficiency, we innovatively proposed a Z-shaped moving target instead of a fixed target in our self-research for the calibration of the wearable eye-tracking device, which improved the time needed for calibration from 100 seconds to less than 30 seconds. In addition, in order to reduce the calibration error caused by the user's head movement during the calibration, the world camera image is feature-matched, and the virtual optic target is computed instead of the recognized optic target, thus eliminating the effect of head movement. In the calibration experiments of nine graduate students, the calibration accuracy was less than 20 pixels, and the average calibration time was about 24 seconds.

Image fusion technology can integrate visible and infrared thermal image from different sensors and retain the feature data and complementary information in fused images. Condition monitoring of transmission lines is beneficial to improve the stability for power grid operation. Image sensors that are developed rapidly in transmission line application has advantages such as non-contact measuring, automatic monitoring and unmanned aerial vehicle carrying. With the development of machine learning technology, many deep learning networks are used in visible and infrared thermal image fusion of transmission line. Firstly, the image fusion methods based on convolutional neural network, autoencoder network and generative adversarial network which advantages and applicable conditions of different frameworks are compared. Secondly, the evaluation indicators of fused images are proposed, and the combined evaluation method based on multiple indicators are illustrated. Then, focused with the demand of transmission line operation condition monitoring, different deep learning algorithm frameworks have been verification by visible and infrared thermal images gathered from the transmission line. Finally, these algorithms are compared with evaluation indicators and the research direction of the future developments are proposed.

The Large Language Model (LLM) as a representative of generative artificial intelligence, demonstrates strong capabilities in natural language comprehension, which was recently put into engineering applications in the field of power emergency. The author proposes a method of extracting information from power emergency plans by leveraging its emergent abilities and prompt learning techniques. By this method, custom-defined contents can be extracted from power emergency plans and linked to the corresponding personnel to generate executable task instructions. The results indicated that this method can accurately extract the custom-defined information from power emergency plans and applys to different LLMs. And the stronger the emergent abilities of the LLM, the more accurate the information extraction is. The method is proofed to effectively assist power emergency personnel in making decisions and expected to be used in various practical scenarios.

With the development of social informationization, more and more people like to use photography to record the details of their lives. However, these styles are only achieved through modern technology and cannot meet people's needs for some oil paintings. Currently, image-to-image style transfer has been widely used in reality and has received high attention in the field of computer vision. In this paper, we use the cycle generative adversarial network (CycleGAN) to perform style transfer on images. We experiment with the network structure of CycleGAN to convert naturally obtained images into images with a certain style. At the same time, this method does not require the source image and the style image to match each other, thus expanding the application scope. In the experiment, we compare the quality of generated samples using the WGAN, WGAN-GP, LSGAN, and original GAN objective functions. Although generative adversarial networks (GANs) have powerful modeling capabilities, they are difficult to train. The research found that WGAN-GP can stabilize the training process and generate more realistic images, followed by WGAN and LSGAN, while GAN often experiences model collapse phenomenon.

The fast development and restricted energy qualities of Unmanned Aerial Vehicles (UAVs) lead to visit changes in the topology of UAV networks, to adapt to this present circumstance, this paper proposes a distributed topology control based on reinforcement learning in unmanned aerial vehicles networks. Neighbor discovery, Election of cluster heads, establishment of clusters, and cluster maintenance are carried out based on parameters such as UAV survival time, connectivity and link stability time. Reinforcement learning method is used to determine the optimal cluster reconfiguration strategy. For every parameter used in CH election, a set of appropriate weights is determined by the maintenance approach to preserve a stable network topology and extend the network lifetime. The strategies are rewarded based on how well they perform on the network. The mechanism proposed in this research was assessed by contrasting it with three topology construction methods, K-means, DCM and MMF. The results show that the method in this paper is applicable to UAV networks with different degrees of improvement in topology maintenance time, packet delivery ratio, end-to-end delay, network throughput, and average node survival time.

The United Nations Population Fund (UNFPA) points out that by 2050, the global population aged 60 and above will increase to 2 billion. With the continuous improvement of people's living standards and the aggravation of population aging, people are paying more attention to their own physical health status, and the demand for health monitoring is also becoming stronger. Real time monitoring of the physiological condition of patients or sub healthy individuals who have already suffered from chronic diseases can play a preventive and alarm role. The home health monitoring system based on the Internet of Things and intelligent robot technology is expected to solve the challenges brought by aging population. This system collects physiological data of the human body through electrocardiogram sensors, heart rate sensors, etc. to detect the health status of family members. It uses temperature and humidity sensors, PM2.5 sensors, combustible gas sensors, and flame sensors to collect household environmental data, allowing people to live in a safe and comfortable environment. This system achieves real-time collection of high-precision household environmental data and physiological data of family members. And based on these data, further analysis and inference can be carried out to comprehensively judge the environment and behavior of the target, which can provide a more comprehensive and comprehensive assessment of human health status. Through a series of experimental designs, it has been proven that the system can accurately collect physiological and household environmental data from the human body, and can automatically alert abnormal physiological and environmental data through analysis and inference, thus achieving realtime health monitoring of the human body.

According to the automatic navigation operation requirements of Lycium barbarum plantation sprayers, a spray machine automatic navigation control system based on the Beidou navigation system was designed using a tracked Lycium barbarum sprayer as a research platform. The system mainly consists of Beidou positioning system, navigation control system, steering control system, and spray machine. The steering control system can achieve dynamic steering control of the spraying machine based on the expected steering angle sent by the navigation control system. Analyzing the motion model of the spray machine and combining it with a pure tracking model, a linear tracking navigation control system for the Lycium barbarum spray machine was designed. The navigation path planning of the spraying machine was carried out based on the planting mode and ground characteristics of Lycium barbarum berries, and experiments were conducted in Lycium barbarum berry plantations. The experimental results show that under the condition of a forward speed of 3km/s of the spraying machine, the maximum deviation of the linear navigation tracking of the navigation control system is not more than 0.16m, and the average absolute deviation is not more than 0.04m, meeting the requirements of automatic navigation operation for planting spraying.

In order to advance the development of secure and effective self-driving technology, a novel approach for intelligent vehicle target detection and automated navigation has been introduced. To improve the exploitation of semantic information, we enhanced the FPN structure by using the Spatial Adaptive Filter (ASF) module and input it into the FPN structural layer. Then the ROS system is used to realize the function package to complete the configuration of parameters and the real-time construction of the map. On this basis, the elite ant colony algorithm is combined to realize the planning of optimal paths. The experimental outcomes demonstrate that in the improved algorithm scheme, the average accuracy mean under simple, medium, and complex categories is 87.12, 77.80, and 76.02 respectively. The set of four different types of obstacles can realize the path planning of the intelligent vehicle and obtain better results. To conclude, the effectiveness and feasibility of the program is verified by multiple sets of experimental data.

This paper presents a novel pitching moment model designed for solar-powered Unmanned Aerial Vehicles (UAVs) operating within thermal updraft environments, drawing upon the conceptual framework of thermal updraft modeling. The model demonstrates its versatility by finding application in tasks related to environmental thermal updraft positioning and tracking, all facilitated through the use of a state Unscented Kalman Filter (UKF) estimation algorithm. The simulation outcomes convincingly affirm the substantial advantages brought about by the introduction of the solarpowered UAV pitching moment model within the framework of thermal updraft state estimation. When pitted against the use of a standalone thermal updraft model or a solar-powered UAV thermal updraft field rolling moment model, the algorithm exhibits remarkable enhancements in both convergence speed and accuracy when it comes to locating the thermal center. These improvements signify the pertinence and effectiveness of the proposed model, establishing it as a valuable addition to the arsenal of tools available for optimizing solar-powered UAV operations within thermal updraft fields. The novel model's contributions have the potential to significantly advance the capabilities of solar-powered UAVs, enabling them to more efficiently harness the benefits of thermal updrafts in various environmental monitoring and surveillance applications.

Agricultural cultivation is gradually entering a new era of technology and intelligence, particularly in the field of "digital agriculture" innovation, thanks to the rapid development of agricultural technology. Soybean and corn seeding, in particular, has received a lot of attention as a critical link in agricultural production. Traditional soybean and corn seeders have flaws such as a simple structure that can easily result in missed or repeated seeding. With the return of straw to the field, the quality of traditional mechanical seeding has deteriorated even further. To address this issue, this study employs Beidou navigation technology and a fuzzy PID control algorithm to design and test a Beidou navigation-based soybean-corn high-speed precision electric control seeding system. The organic combination of electric control and seeding is realized by parsing the latitude and longitude signals issued by the Beidou navigation device for motor drive. For motor control, the fuzzy PID control algorithm is used, which achieves precise control of motor speed while significantly shortening motor response time. The system can also change crop spacing and fertilization speed to better adapt to different agricultural production needs. According to the results, the average qualification rate of the seeding mentioned in this paper exceeds 93.5%, and the seeding speed has increased to 8km/h.