Aiming at the problems of non Gaussian, non linear and weak signal energy of microseismic signal, a denoising method of microseismic signal based on particle filter is proposed. By establishing the state equation of the microseismic signal, extracting the background noise of the original signal, taking the sum of the state equation and the state equation as the observation equation, establishing the state space model by combining the state equation and the observation equation, and approximately estimating the posterior probability density through importance sampling and resampling, so as to solve the denoisedmicroseismic signal and improve the denoising effect of the microseismic signal. The application in simulated microseismic data and real microseismic data shows that compared to the traditional denoising method, this method has better processing effect, removes the noise and retains the effective signal, and the signal-to-noise ratio is effectively improved. Therefore, it has a good application prospect.

In order to further reveal the current core countries, institutions, disciplines, research hotspots and frontiers in the field of face recognition, CiteSpace visual analysis software is used to retrieve 10 000 documents from a total of 15 494 documents in the global face recognition field. Visual processing from the perspectives of countries, disciplines, research institutions, hot keywords, etc. and analysis from two dimensions of technology and application. It is concluded that there are still problems in the application of face recognition technology under the condition of mask occlusion. The improvement of K-SVD(K-Singular Value Decomposition) algorithm discrimination and the fusion of face recognition technology are believed can solve to a certain extent by combining 3D image recognition, focusing on the characteristics of the test set, and integrating with multiple disciplines.

In order to solve the problem of three-dimensional measurement of objects that cannot be contacted closely in special scenes, the UAV(Unmanned Aerial Vehicle) hoisting measurement system is designed and implemented based on UAV and raspberry. The cascaded raspberry is used to realize image transmission and data transmission link. The laser sensor, GNSS(Global Navigation Satellite System)sensor and monitoring lens sensor are utilized to realize information acquisition and combined with electric telescopic rod to realize the vertical adjustment of measuring device. The binocular camera is used to collect information to realize scene three- dimensional modeling and object three-dimensional accurate measurement. The UAV hoisting measurement system is composed of five parts: power system part, image transmission and data transmission integrated link part, positioning and height acquisition part, electric telescopic rod part and object three-dimensional measurement part. The test results show that the UAV hoisting measurement system can quickly, effectively and automatically construct the three-dimensional measurement scene and complete the actual three-dimensional dimension measurement of the target object. The accuracy error is up to 1% , and the expected effect is achieved.

In order to maintain high transmission efficiency when different loads are being charged, a dual-switch buck-boost conventer is proposed at the load end. By adjusting the duty cycle of the converter, the equivalent load resistance can be always equal to the optimal load resistance to achieve maximum efficiency. And the expected output power value is set, the corresponding output voltage according to different load resistance is calculated sending to the controller as a reference voltage, the phase shift angle of the inverter is adjusted through closed-loop phase shift control, and output the corresponding voltage stably realizing constant power control. The simulation shows that the transmission efficiency remains constant at 85% and the output power remains constant at 4 200 W when the load resistance varys between 10 ~ 40Ω.

Aimming at the problem of reasonable planning and configuration for the distributed generation to the distribution network, network loss, node voltage deviation, investment, on the basis of comprehensive cost as the objective function, using analytic hierarchy process, the stability of the economy and multi-objective function are considered as a single objective function, and improved Harris hawk algorithm is used to find the optimum configuration scheme. In order to improve the optimization accuracy and convergence speed of Harris Eagle algorithm, Tent chaos is introduced to improve the uniformity of initial population. The combination of whale algorithm and Harris-Eagle algorithm increases the diversity of the algorithm and improves the accuracy of the search. The t-distribution strategy is introduced to enhance the global optimization ability and solve speed. An example is given to verify the efficiency and stability of the improved Harris Eagle algorithm in solving the distributed power supply optimization problem. Through the simulation experiment, compared with the existing particle swarm optimization algorithm and the basic Harris Eagle algorithm, the improved Harris Eagle algorithm effectively reduces the network loss and voltage deviation of the original distribution network system, and reduces the overall investment cost of the distributed generation. The improved Harris-Eagle algorithm is proved to be efficient and stable in solving distributed generation optimal allocation problem.

In order to improve the accuracy of load prediction results, a short-term power load prediction model based on VMD(Variational Mode Decomposition) and fractal theory is designed. Firstly, the date similar to the measured meteorological data is selected as the reference date, and the modal of the main signal (trend term) is extracted from the load curve of the reference date through variational modal decomposition. After extracting the disturbance term ( IMF1 ) and noise term ( IMF2, IMF3, IMF4 ), the extreme points of the trend term and disturbance term are extracted after re-scaling range analysis, and the IFS ( Iterative Function System) is established. The load data corresponding to the trend term and disturbance term are determined by the iterative function curve, and then added to the load data corresponding to the noise term (IMF2,IMF3,IMF4) to obtain the final forecast data. Compared with the traditional fractal model and BP(Back Propagation) neural network model, the mean absolute percentage error (MAPE: MeanAbsolute Percentage Error) decreases by 5% , which proves that vmD-fractal prediction model has better prediction effect. 

The traditional fault location method has lower accuracy due to ignoring the change of line parameters. In order to improve accuracy of line fault location, two-terminal asynchronous location is used and total least square method is optimized by Huber robust estimation to reduce the effect of inaccurate line parameter estimation which is resulted from existed error when adopting total least square method. Matlab simulation shows that improved method is more accurate than traditional method in line parameter estimation and fault location accuracy. And the fault location results which are not affected by the fault location and fault type is also verified by simulation.

In order to better solve the problems such as large search range and slow convergence speed in the reconfiguration, a reconfiguration model is established aiming at reducing the active power loss and balancing load, and a reconfiguration strategy of distribution network based on improved slime mold algorithm is proposed. Firstly, the prim algorithm is combined with the continuous branch switching algorithm to eliminate the infeasible solutions for the power distribution network with complex topology, which is used as the initial population of slime mould algorithm. Secondly, in order to solve the problem of slime mould algorithm‘s poor convergence speed and easy to fall into local optimal solution, the search threshold and Gaussian disturbance is introduced to improve the convergence speed, and the refraction reverse learning strategy is used to jump out of the local optimal solution. The simulation results show that the reconfiguration method can effectively decrease line active power loss, improve voltage level and balance load of complex distribution networks, and has high accuracy and short iteration time.

Aiming at the problems of large starting current and bus voltage drop when three-phase asynchronous motor starts, a soft starting control system for heavy-load starting is designed. First of all, based on the motor control theory, the pulse generation module, three-phase voltage regulation module and PID ( Proportional- Integral-Derivative) control closed loop are built, and the soft start control system is designed to control the terminal current and voltage of the asynchronous motor. Secondly, based on the Simulink software, according to the actual data of the feed pump motor of a coal-fired unit in Heilongjiang Province, the soft start control model is built and simulated. Finally, the field test is carried out, and the test results are similar to the simulation results. The results show that the model can effectively limit the starting current of asynchronous motor and reduce the influence of water pump motor start-up on other equipments at the busbar under the condition of heavy load starting. It provides a reference for the heavy-load start-up of the asynchronous motor. 

In order to effectively evaluate the information transmission characteristics of infrared gas sensing system, to serve the teaching needs of basic concepts such as information entropy in the course of “information theory” and to promote the full integration of teaching and scientific research, the infrared gas sensing system is treated as a channel, the actual concentration is treated as an input information source and the measured concentration is treated as an output information source. The information theoretical model of the infrared gas sensing system is established. The analysis method of information transmission efficiency of infrared gas sensing is proposed. Firstly, aiming at the wavelength modulation spectroscopy technique, the functional simulation of laser modulation, gas absorption, photoelectric conversion and harmonic extraction are carried out. According to the relevant principles of Shannon‘s information theory, the mathematical expressions of transfer matrix, source entropy and mutual information are presented, and the characteristics of information transmission efficiency and the minimum resolvable concentration interval of the sensing system are analyzed. Based on the actual measurement results of long-time stability of the sensor, the variation characteristics of the information transmission efficiency and the minimum resolvable concentration interval of the sensor system with the integration time of the system are analyzed. The presented information measurement method integrates the basic knowledge of information theory and expands the scientific research and application case of infrared gas sensing system based on wavelength modulation spectroscopy, which is innovative. The practical application shows that the information measurement method of infrared gas sensing system meets the experimental teaching needs of information theory and the scientific research needs of wavelength modulation spectroscopy, and has achieved good teaching and scientific research application results.

Portable transient electromagnetic system is suitable for unexploded ordnance detection under complex terrain conditions, and the data processing method has great influence on its detection efficiency. In order to improve the data processing speed, the factors affecting the inversion efficiency are optimized. The portable transient electromagnetic system made by the Electromagnetic Sensor Technology Laboratory of Jilin University is used to obtain the target response, and the DE (Differential Evolution) algorithm is used to invert the data based on the magnetic dipole forward model. The inversion speed is greatly improved by optimizing the data preprocessing and DE algorithm parameters. The experimental results show that the position parameters and characteristic responses of Unexploded Ordnance inverted by the optimized DE algorithm are basically the same as those obtained by the conventional DE algorithm, the positioning error is within 6 cm, and the calculation time is 1. 4% of the conventional DE algorithm.

In view of the problems existing in the current national planning system, such as the independent system of various departments‘ planning, content conflict and lack of cohesion, a planning concept of “one city or county, one plan and one blueprint” is introduced based on the sorting out of the spatial planning system, to realize the unified coordination and management of various plans. First the current space planning is described, the existing problems are analyzed, and then an unfied solution, a variety of planning data conflict detection, data management planning, the formation of planning a figure, intuitive display planning data are put forward, based on the graphic joint collaboration with different departments for examination and approval, to ensure project success eventually landing; the implementation of this system solves the business contradiction between planning and management departments and realizes the goal of “integrating multiple regulations”.

In order to summarize the existing applications of blockchain technology, to study the application of blockchain technology in the future, the development process and working principle of blockchain applications are researched, and the development and application scenarios of blockchain are analyzed. It mainly makes a general outlook on the key application scenarios of blockchain technology, including education, medical care, food, Internet of Things, social assistance and targeted poverty alleviation, hoping to provide help for the construction of the application system of blockchain. 

Since the onboard CAN( Controller Area Network) bus is the source of network attacks, onboard data transmitted in plaintext form is easy to be monitored and tampered, and messages without authentication domain and bus arbitration mechanism are vulnerable to network attacks. In order to solve the problem of automobile information security, an asymmetric encryption algorithm based on SM2 is proposed, using multiple point operation strategy, the paper uses public key encryption, private key decryption, carrying out vehicle node identity authentication and data encryption and decryption, according to the characteristics of vehicle CAN bus, the paper uses message of packet protocol to transmits the ciphertext of the SM2 algorithm. The paper transplants the SM2 algorithm to the embedded device with the domestic GD32F103 as the core, simulates the CAN communication between the TBOX( Telematics-BOX) and PEPS( Passive Entry Passive Start) , an important part of the intelligent connected car, and analyze and verify that the SM2 algorithm is effective and feasible in the embedded device through the Kavaser CANKING and Sealaselogic8 platform analysis. It ensures the legitimacy and confidentiality of node data, and it is suitable for single-time identity authentication occasions for in-vehicle devices. The SM2 algorithm provides a new research path for the research on the security authentication method of intelligent networked vehicles.

In order to solve the problem of short service life of the reducer due to inaccurate thermal analysis, taking a certain type of SEW( Süddeutschen Elekromotoren-Werke) reducer as the research object, using the Unity3D virtual simulation engine, the thermal power calculation software of the reducer is developed. A virtual simulation method of the temperature field based on the Unity3D example system is proposed. The basic formula of temperature and color expression is derived, and the virtual simulation cloud diagram of cooling water temperature and the change curve of heat transfer power are obtained. Practice has proved that the virtual simulation method can quickly determine whether the cooling method selected for the reducer is reasonable. The virtual simulation technology also provides a new design idea for the thermal analysis of the reducer. It is more convenient and quicker than traditional theoretical calculation or finite element simulation, and has certain engineering application prospects. 

In order to solve the problem of identifying the mental health status of college students, this paper based on students’ data of consumption, Internet access and mental status assessment result,firstly, Jenks Natural Breaks algorithm is used for feature classification, and then according to the results of feature classification, Apriori algorithm is used for feature association analysis to mine the behavioral features that have a certain correlation with students’ mental health status. Finally, the inertia weight is improved based on Particle swarm optimization algorithm, and the process of identifying, mutating and selecting inferior particles is added to avoid the algorithm falling into local optimal solution, and the firefly perturbation strategy is used to accelerate the convergence of particle swarm optimization to the global optimal solution. A PDNN (Particle Difference Neural Network) neural network model is constructed to predict the mental health state of students. The experimental results on the data set of students’ behavior characteristics show that the proposed model is superior to the traditional machine learning and related deep learning models, it can converge quickly and predict students’ mental health status more effectively and accurately. 

Multimedia Information Encryption Algorithm Based on Big Data Analysis MARDAN Zunon (Department of Information Engineering, Xinjiang Institute of Engineering, Urumqi 830023, China) Abstract: The traditional method of digital multimedia information encryption is based on the knowledge of a single mathematical field, which has some problems such as easy leakage, easy cracking and low security level. As a result, a digital multimedia information encryption method based on big data analysis is proposed. The encryption method of the cloud and encryption hardware is combined to write the digital multimedia information encryption algorithm into the cloud database. Users use cryptographic hardware for identity binding, creating isolated spaces for information interaction. Complete the encryption of digital multimedia information. The experimental results show that the digital multimedia information encryption method based on big data analysis reduces the crack rate of theoretical encrypted files and has good information encryption effect, high safety factor, and strong practicability. 

LPR(License Plate Recognition) in nighttime is one of the challenging works for developing artificial vehicle systems. Traditional license plate recognition algorithms have a low accuracy due to uneven lighting and camera shake exposure at nighttime. This paper proposes a novel license plate recognition algorithm based on improving light streak deblurring method, which introduces a constraint on the linear fuzzy model to estimate the fuzzy kernel. The method is able to estimate the blur kernel and deblurred image more accurately, and output image computed by the improved algorithm has better contrast ratio and resolution. We conduct the experiments on the benchmark license plate database, experimental results show that the proposed method in this paper is able to recognize license plates in low-illumination environments which has better resolution and contrast ratio with little interference resulted by noise, besides that its average kernel similarity and peak signal-to-noise ratio are significantly better than state-of-the-art works. The recognition accuracy improved at least 22. 6% compared the result without de-blurring.

We have the effect of adversarial examples is reduced and the accuracy of the classification model is improved under the threat. Inspired by the mammalian visual modality, we proposed a purification defense method using a novel parallel attention mechanism to mitigate the effect of adversarial examples, called PSCAM- GAN(Parallel Spatial and Channel Attention Mechanism Adversarial Generative Network). The defense model first generates the feature map through the encoder, the parallel attention module is used to extract the object and space information. Under the condition that these features remain unchanged, the weight of the feature map is readjusted generating purification results by decoder. This method can keep the consistency between the purification result and the input while removing malicious perturbation, and effectively reduce the influence of adversarial samples on the model accuracy. The robustness of the model is evaluated through various types of attacks on CIFAR-10 and MNIST dataset. The experiments show that PSCAM-GAN completely surpassed other pre-processing based defense methods. These mean the defense method can effectively improve the robustness of the original models.

In order to explore the related properties of porous materials at pore size scale and better apply them to engineering, it is necessary to fully understand the internal pore structure of materials. Because it is difficult to obtain 3D CT images directly, the deep learning method is considered to approximate the real probability distribution of porous materials, and the random number is used for 3D reconstruction. Firstly, the existing binary images are segmented to obtain data sets with appropriate size and sufficient quantity. Then, WGAN (Wasserstein Generative Adversarial Networks) is transformed into a generativeadversarial model that can process three-dimensional data, and the data is generated by using generator andthe random noiseobeying Gaussian normal distribution. Finally, WGAN is trained by generating data and real data. By calculating the two-point correlation function, Minkowski functional and permeability of the generated image, the relevant parameters such as porosity, specific surface area and function curve fitting are investigated, and tested with different data sets. According to the results, the three-dimensional reconstruction model based on WGAN has high accuracy under different input conditions. At the same time, the algorithm also has the characteristics of low time complexity and the generated model can be stored and reused.

In power system, line loss is a common energy loss of electric energy loss. Therefore, it is particularly important to analyze the causes of line loss in distribution network and carry out theoretical calculation to reduce energy consumption. This paper introduces the basic causes of line loss and simply analyzes the common methods of calculating line loss. In many methods, in view of the equivalent resistance method is suitable for the operation of the distribution network electric energy loss calculation characteristics, the principle of equivalent resistance method, application conditions are analyzed, the equivalent resistance method is applied in the 10 kV distribution line loss calculation. Based on the comparison between the result of calculation example and practice, the application of this method in practice is analyzed. The results show that the error of equivalent resistance method is small, and the results have a certain reference significance to establish the evaluation index of line loss.

At present, the storage rate of the super dense heterogeneous network data storage method is too slow, resulting in a small amount of storage, which is difficult to meet the requirements of people. Therefore, a data storage method based on mobile edge computing is proposed. We construct a network structure mode based on mobile edge, establishe a corresponding selection hierarchy model for the ultra dense heterogeneous network by using the analytic hierarchy process of network structure, and classify the ultra dense heterogeneous network data according to the two aspects of network data collection and processing, including data collection network and data processing network. By using the mobile edge computing technology, the network data written will sink to the edge of the network to realize the persistent storage of ultra dense heterogeneous network data. The experimental results show that the super dense heterogeneous network data storage method based on mobile edge computing can effectively improve the storage rate and expand the storage capacity. 

The laboratory is an important place for experimental practice teaching in universities. A large number of experimental equipment and the high mobility of personnel in open laboratories pose certain difficulties for management. In order to realize the intelligent management of laboratory equipment safety and personnel information, a set of intelligent laboratory equipment security and personnel information management system is researched and developed. The system integrates a variety of sensor modules, which can realize the functions of equipment entry and exit detection and alarm, personnel information statistics and so on. It is composed of equipment detection module (RFID: Radio Frequency Identification), video monitoring module, communication module, alarm module, personnel information detection and recording module, etc. And the human-machine interface design is completed based on the host computer software USART(Universal Synchronous Asynchronous Receiver Transmitter) HMI (Human Machine Interface). The test results show that the system can effectively detect the unregistered takeout of instruments and equipment, and ensure the safety of experimental equipment. The system can record the information of personnel entering and leaving the laboratory, and improve the efficiency and safety of laboratory management. The management system and design ideas have good reference significance for laboratory safety management.

Power companies need many kinds of materials, and the fluctuation of material prices is affected by many factors. To predict the current price trend, a cable price prediction model is established which can provide the basis of bidding base price and reasonable procurement opinions for power companies is eslablished. The dynamic time warping method is used to determine the price delay time for the collected material prices, that is, the time when the tax free unit price of materials lags behind the price of raw materials. And the corresponding relationship among unit price excluding tax, raw material price and economic indicators is determined. There are many factors affecting the change of material price. The key characteristics are screened by Pearson coefficient and random forest. According to the selected key characteristics and data, AdaBoost, xgboost and random forest models are established to predict the material price. The average absolute percentage error of the prediction evaluation index is used to evaluate the prediction effect. It is found that the key features screened by random forest and then combined with XGBoost model have the highest accuracy of prediction.