Advanced Hybrid Information Processing

In order to improve the intelligent management and information scheduling ability of bridge long-term health monitoring, the real-time data monitoring and automatic collection design of bridge long-term health monitoring are carried out with big data analysis method. A classification method of bridge long-term health monitoring data based on fuzzy correlation feature detection and grid area clustering is proposed. The information fusion and fuzzy chromatography analysis method are used to realize the information fusion of the real-time data of bridge long-term health monitoring, and the adaptive feature extraction of related data is carried out. Excavate the positive correlation characteristic quantity of bridge long-term health monitoring real-time monitoring data flow, carry on the fuzzy clustering and information prediction of bridge long-term health monitoring data flow, and improve the accuracy of bridge long-term health monitoring real-time data monitoring. The simulation results show that the intelligent classification of bridge long-term health monitoring based on this method has high accuracy and low error rate, which improves the real-time performance of bridge monitoring.

In order to improve the construction quality detection ability of building concrete, the construction quality detection method of building concrete based on big data is put forward, and a construction quality detection model of building concrete based on feature extraction of association rules is proposed. The nonlinear time series analysis method is used to model the construction quality information flow of building concrete, and the quantitative feature information flow of construction quality of building concrete is reconstructed by quantitative regression analysis. The statistical characteristic quantity of quantitative characteristics of construction quality of building concrete is extracted by statistical feature analysis method, and the spectral density analysis and feature detection of quantitative characteristics of construction quality of building concrete are carried out in the moving average window. According to the abnormal spectrum distribution of high-order statistics, the construction quality inspection of big data building concrete is realized. The simulation results show that the accuracy of using this method to detect the construction quality of building concrete is high.

In order to solve the problem that the visual simulation image of virtual experimental element is missing in the reconstruction module of virtual experimental element, and the reconstruction accuracy of virtual experimental element is not good. A visual display method of virtual experimental element visual simulation image based on big data technology and virtual visual reconstruction is proposed. Firstly, the information transmission model of virtual experimental element visual simulation image is constructed. Then the 5-level wavelet decomposition method is used to decompose and fuse the visual simulation images of virtual experimental elements, and big data fusion technology is used to reconstruct the visual information of virtual experimental elements. The visual simulation image visualization of virtual experimental elements is realized. The simulation results show that this method has good visual display performance and high feature fusion degree in the reconstruction and modeling of virtual experimental element visual simulation image, and has high value in the application of virtual experimental element visual display and digital reconstruction.

In the process of modeling distributed power grid connection, the parameter control effect is not good, and the modeling is not stable. A distributed power energy grid-connected control method based on large data is put forward. Distributed power energy-energy grid-connected model is established, a DC/AC inverter model and a current inner-loop controller model are analyzed. And an equivalent circuit model analysis of the terminal voltage of the grid-connected inverter of the controller is analyzed. The fuzzy PID control algorithm is introduced to identify the unknown parameters in the distributed power energy grid-connected control, the fitness function of the inner ring controller and the fitness function of the outer ring controller are obtained. Expert database is initialized and updated, Until the maximum number of iterations or convergence accuracy is reached. The simulation results show that the proposed method can effectively improve the performance of the distributed power grid-connected control.

To improve the ability of blind identification and scheduling of sparse multipath channels in wireless communication networks under the background of Internet of things, a blind identification algorithm for sparse multipath wireless communication based on random sampling interval equalization and BPSK modulation is proposed. The sparse multipath channel model of wireless communication network under the background of Internet of things is constructed, and the multipath characteristics of sparse multipath channel of wireless communication network are analyzed. The BPSK modulation method is used to filter the inter-symbol interference of sparse multipath channel of wireless communication network. Based on the adaptive random sampling interval equalization technique, blind channel identification is designed, and the tap delay line model is used to suppress the multi-path of sparse multipath channel in wireless communication network. The simulation results show that the blind identification of sparse multipath channels in wireless communication networks is well balanced and the bit error rate (BER) is reduced.

An anti-co-frequency interference suppression method for wireless spread spectrum communication based on equivalent low-pass time-varying pulse modulation technology is proposed. The anti-co-frequency interference system for wireless spread spectrum communication is designed based on Internet of things technology, and the multi-path channel model for wireless spread spectrum communication is constructed. The Doppler spread technique is used to design the channel equalization of wireless spread spectrum communication system. The equivalent low-pass time-varying pulse modulation method is used to suppress the same-frequency interference and blind source separation. Improve the lossless transmission ability of the Internet of things (IoT) transmission signal in the wireless spread spectrum communication system. The simulation results show that this method is used to design the wireless spread spectrum communication system and the co-frequency interference is effectively suppressed and the bit error rate of communication is lower than that of the traditional method.

In order to improve the power grid load detection ability, an optimal method of load signal control of power based on state difference clustering is proposed, and the big data statistical analysis model of the power grid load is constructed. The clustering analysis and state mining of grid load are carried out by using the distributed detection method of association features, and the regression analysis model of grid load state difference is constructed to realize the state differential clustering of power grid load signal in high-dimensional phase space. Based on the classification and fusion of the extracted characteristic sets of grid load, big data analysis method is used to optimize the intelligent control of power grid load signal. The simulation results show that the proposed method has better accurate classification performance and lower misdivision rate, which improves the output stability of power grid load.

The high-speed image signal of LVDS interface is easy to be interfered by the outside world in the process of transmission, which results in packet loss and distortion of high-speed image communication, and the output error is high. Therefore, the lossless coding of high-speed image signal is needed. Intelligent estimation of bit error rate (BER) for high-speed image transmission is needed. The intelligent estimation model of high-speed image transmission bit error rate based on LVDS interface is proposed. The network structure model of high-speed image signal transmission is constructed to estimate the error code distortion of image transmission and the key frame feature extraction method is used to estimate the error rate of image transmission. The intelligent estimation of bit error rate (BER) of high-speed image transmission is realized in LVDS interface. The simulation results show that the proposed method has low bit error rate (BER) for high-speed image transmission and achieves lossless transmission of images.

To improve the security of network sensitive information transmission and storage, it is necessary to design the anti-tampering monitoring of network sensitive information, and a tamper-proof monitoring technology of network sensitive information in big data environment based on big data dimension feature block is proposed. Big data feature space reconstruction method is used to calculate the grid density of network sensitive information distribution, and the network sensitive information to be tampered-proof monitoring is mapped to the divided high-dimensional phase space through the density threshold. The high dimensional phase space of information distribution is divided into dense unit and sparse unit. The coded key is matched to the corresponding network sensitive information block to realize information encryption and covert communication. The simulation results show that the information steganography performance of network sensitive information transmission and storage using this information tampering monitoring technology is better, and the information security transmission ability is improved.

In order to improve the detection and recognition ability of weak sensing signal, an intelligent detection algorithm of weak sensing signal based on artificial intelligence algorithm is proposed. The weak sensing signal model is constructed, the weak sensing signal is separated and processed adaptively, the scale and delay of the weak sensing signal are estimated adaptively, and the high resolution spectral features are extracted. The extracted spectral feature is studied adaptively and detected intelligently by artificial intelligence algorithm, and the spectral peak search of weak sensing signal is realized. The spectral feature component method is used to realize the interference suppression of weak sensing signals, thereby improving the detection of the method. The simulation results show that the algorithm has high accuracy and anti-interference ability, and improves the detection and recognition ability of weak sensing signal.

In order to improve the intelligence and real-time performance of ultra-broadband wireless communication, it is necessary to control the time delay of intelligent data transmission in ultra-broadband wireless communication. An intelligent data transmission time delay control algorithm for ultra-broadband wireless communication based on artificial intelligence algorithm is proposed. A wireless communication network transmission model based on wireless sensor networking model is constructed, and the position and scale parameters distributed in the process of wireless communication transmission are measured by using different scales. It is found that the time delay control problem is the best replica correlation matched filter detection problem, and the communication delay control is realized to the maximum extent. The simulation results show that the artificial intelligence of ultra-broadband wireless communication delay control is good and the communication quality is high.

In order to improve the monitoring ability of uncertain large data stream, an uncertain large data flow monitoring algorithm based on artificial intelligence is proposed. The collected uncertain big data flow is constructed by low dimensional feature set, and the rough set model of uncertain large data stream distribution is constructed. The fuzzy C-means clustering method is used to analyze the uncertain big data flow by fusion clustering and adaptive grid partition analysis. All the abnormal samples of large data stream are sampled and trained, and the feature quantities of association rules of uncertain large data stream are extracted. Combined with artificial intelligence method, the monitoring of uncertain large data stream is realized. The simulation results show that the method has high accuracy and good ability to resist abnormal traffic interference, and the traffic security monitoring ability of the network is improved.

Parallel mining of image data is based on the extraction of internal rules and detail features of image. Combined with image edge detection to realize parallel mining of image data, a parallel mining algorithm of image data based on AI is proposed. Firstly, the multidimensional parallel eigenvalues of image data are extracted by the gray feature extraction algorithm of massive images, and then the template matching and information fusion of massive image data are carried out by using Map/Reduce model. According to the matching results, the parallel mining results of image data are obtained. Finally, the simulation experiment of image data parallel mining is realized by using Matlab software. The results show that compared with other image data parallel mining algorithms, this algorithm reduces the parallel mining time of image data and improves the speed of image data parallel mining, especially for large-scale image data parallel mining.

The jointed-feature detector for the floating small target in sea clutter is addressed in the paper. For the traditional energy-based detectors, it is difficult to detect the low signal-to-clutter ratio floating small target in time domain due to the affection of sea clutter motion. Therefore, a feature detector in the Fractional Fourier transform (FRFT) domain is proposed. The Hurst exponent and fractal dimension variance are extracted as the features in the jointed-feature detector in FRFT domain. The decision region is determined by convex hull training algorithm on the given false alarm probability. The experimental results of 10 groups of IPIX radar data show that the jointed-feature detector is superior to the compared one, and it provides a new detection scheme for radar target detection.

In order to improve the precision of fatigue life prediction of ballastless track, a method for predicting fatigue life of ballastless track based on big data is proposed. The big data model is constructed to analyze the fatigue life cycle of ballastless track. Big data mining and feature extraction are used to extract the fatigue life cycle of ballastless track. Combining with the particle swarm optimization method, the feature classification of the failure state trend of ballastless track construction is carried out, and the information fusion is carried out according to the characteristic parameters of the failure state of ballastless track construction. The expert system model for predicting fatigue life of ballastless track construction is established and the fatigue life of ballastless track is predicted by association rule mining method. The simulation results show that the precision of fatigue life prediction of ballastless track is high, and the strength and life cycle of ballastless track are analyzed.

In order to improve the damage detection and identification ability of high-speed railway turnout structure, a machine learning-based damage identification method for high-speed railway turnout structure is proposed, and the computer vision image analysis method is used to detect the damage of high-speed railway turnout structure. The super-linear segmentation and feature recognition of the damaged parts of high-speed railway turnout structures are realized by means of active contour detection, and the feature segmentation and localization of high-speed railway turnout structures are carried out in the damaged areas. According to the result of feature matching, the machine learning algorithm is used to identify the damage of high-speed railway turnout structure. The simulation results show that the accuracy of the proposed method for damage identification of high-speed railway turnout structure is high, and the ability of damage detection and identification of high-speed railway turnout structure is stronger than that of high-speed railway turnout structure.

In order to improve the data integrity encryption ability of cloud storage users, a cloud storage user data integrity encryption technology based on random linear coding is proposed. Firstly, the cloud storage user data integrity encryption key of cloud storage user data integrity object is constructed, then the encryption and decryption coding design are carried out, and the random linear coding is used to optimize the digital encryption key to improve the anti-deciphering level. Finally, the simulation results show that the cloud storage user data integrity encryption technology has high random distribution of encryption, the deciphering rate of encrypted data is effectively controlled, and the performance of the cloud storage user data encryption technology is better than that of the traditional method. It effectively ensures the integrity of cloud storage user data.

There is color difference in the image collected under the background of night light. Machine learning and fusion tracking compensation method are used to detect and process the maximum color difference of the image, so as to improve the imaging quality of the image. A maximum color difference detection algorithm for nightlight background color difference image based on machine learning and fusion tracking compensation is proposed. Firstly, image feature acquisition and color difference feature blending preprocessing are carried out, image machine learning and fusion tracking compensation are carried out, and image color difference detection algorithm is used for image color difference smoothing and adaptive blending. The background color difference image of night light is automatically divided into target space by feature clustering, and the maximum color difference detection of the detail features of the image is carried out to the greatest extent. The simulation results show that the algorithm has high accuracy and good color difference resolution.

In order to improve the imaging quality of 3D image with visual feature reconstruction, it is necessary to control the color of low contrast image automatically. A color automatic control technology of low contrast image based on 3D color space packet template feature detection is proposed, the automatic color control model of image based on big data analysis is constructed. RGB decomposition technology is used to extract the color components of low contrast images, and color space gray feature fusion algorithm is used to segment fusion of low contrast images to improve the feature pairing performance of color peak points of low contrast images. Combined with the color space block fusion information of low contrast image, the edge features of high oscillatory region are detected, and the color automatic control of low contrast image is realized. The simulation results show that the color automatic control of low contrast image can improve the peak signal-to-noise ratio (PSNR) of image output, improve the automatic color control ability and imaging quality of low contrast image.

In order to improve the ability of data retrieval in cloud computing environment, a reduced dimension classification algorithm of complex attribute big data in cloud computing based on deep neural network learning is proposed. The complex attribute big data under cloud computing is constructed by low dimensional feature set, and the complex attribute big data under cloud computing is analyzed by linear programming and fitting using grid clustering method. Big data samples of all complex attributes are sampled and trained to extract the associated features of big data, which is a complex attribute under cloud computing. The feature extraction results of complex attribute big data under cloud computing are inputted into the deep neural network learner for data classification, and the complex attribute big data dimensionality reduction classification under cloud computing is realized by combining big data fusion clustering method. The simulation results show that the accuracy of big data dimension reduction classification for complex attributes in cloud computing is high and the error rate is small.

Aiming to improve the security of large database in cloud storage space, a hierarchical mining algorithm of spatial big data set association rules based on association dimension feature detection is proposed. The statistical characteristic quantity of large spatial data set is constructed by means of group sample regression analysis, and the sampling and sample recognition of spatial big data set are carried out by using fuzzy rough set mapping method. The association rule distribution model of large spatial datasets is constructed by using the hierarchical mining method of association rules, and the feature quantities of association rules are extracted from large spatial datasets. The correlation dimension feature extraction algorithm is used to optimize the extraction process of large spatial data sets adaptively, so as to realize the hierarchical mining optimization of spatial big data set association rules. The simulation results show that the proposed method has higher accuracy, higher mining accuracy and better feature matching ability, which improves the mining ability of association rules in large database in cloud storage space.

In this paper, the problem of locating and tracking moving target with uniform acceleration by moving multi-stations is studied. Based on the time-difference information and frequency-difference information of target signal arriving at different base stations, a method of locating and tracking aerial moving target based on time-frequency difference is proposed. This method is based on extended kalman filter (EKF) and unscented kalman filter (UKF) filtering algorithms respectively to locate and track moving target, and compares the locating results of the two algorithms. This method can not only locate and track the aerial target, but also estimate the velocity and acceleration information of the target. The simulation results show that the location and tracking results of this method can achieve high positioning accuracy, and the positioning accuracy of UKF is better than that of EKF and better positioning results can be obtained, which has a certain reference value for the engineering realization of multi-station moving target location and tracking in the air.

In this paper, we discuss the problem of sparse recovery in compressed sensing (CS) in the presence of measurement noise, and present a variable iterative synthetic aperture radar (SAR) imaging method based on sparse representation. In this paper, the sparse reconstruction theory is applied to SAR imaging. The SAR imaging problem is equivalent to solving the sparse solution of the underdetermined equation, and the imaging result of the target scene is obtained. Compared with the previous algorithms using $$ l_{1} $$-norm or $$ l_{2} $$-norm as cost function model, this paper combines $$ l_{p} $$-norm $$ (0 < p < 1) $$ and $$ l_{2} $$-norm as cost function model to obtain more powerful performance. In addition, a smoothing strategy has been adopted to obtain the convergence method under the non-convex case of $$ l_{p} $$-norm term. In the framework of this iterative algorithm, the proposed algorithm is compared with some traditional imaging algorithms through simulation experiments. Finally, the simulation results show that the proposed algorithm improves the SAR signal recovery performance to a certain extent and has a certain anti-noise ability. In addition, the improvement is more evident when the SAR signal is block sparse.

With the rapid development of Internet of Things (IoT), The IoT terminal are diversified, and the attack points available to attackers are also diversified, and most IoT terminal are more vulnerable to attacks because they are less secure. In order to achieve secure access of IoT terminal, ensure the legality of IoT terminal accessing the network, improve the security of terminal entering the network and reduce the security risks that IoT terminal may be exposed to when accessing the platform. This paper proposes a secure method to access to the IoT, which is to use the blockchain to store and verify the fingerprint information of the terminal, thereby improving the security of the IoT terminal accessing the cloud. And this paper also proposes a method to store the fingerprint blockchain of the terminal in the IoT terminal, and then verify the collected fingerprint information through the data in the blockchain to ensure the credibility of the fingerprint information.

Because of bandwidth and buffer limitation in wireless network, rebuffering events and bitrate drop often cause video impairments, e.g. compression artifacts and video stalling. Hence, these problems often make a loss of the quality of experience (QoE). For making a prediction about the impact of video impairments on QoE, a continuous predictive model for QoE in wireless video streaming is proposed. In this paper, the inputs are composed of three vectors that are the quality of video frame, rebuffering events state and human memory effect, and the output represents the predicted continuous QoE. We build the predictive model by a Hammerstein-Wiener model. Experimental results show that the proposed model can accurately make a prediction about continuous subjective QoE.

In this paper, we consider the range distributed target detection in partially homogeneous clutter which satisfies a different statistical property in adjacent range cells. The group method wherein adjacent cells with slightly varied statistics are in the same group is presented firstly, which can improve the accuracy of modeling clutter. We assume that all texture of the compound Gaussian clutter satisfies an inverse Gamma distribution but scale and shape parameters in those groups differ from one another. The group generalized likelihood ratio test (G-GLRT) developed here concerns the cells group effects on deducing the GLRT. Considering a knowledge-aided (KA) model that tracking into account the partially homogeneous training samples, we develop a KA-G-GLRT for range-spread target detection and verify the constant false alarm rate (CFAR) with respect to the estimated covariance matrix of speckle. Experimental results are presented to illustrate the performance and effectiveness of the KA-G-GLRT in real clutter data.

The distributed alternating direction method of multipliers is an effective method to solve large-scale machine learning. At present, most distributed ADMM algorithms need to transfer the entire model parameter in the communication, which leads to high communication cost, especially when the features of model parameter is very large. In this paper, an asynchronous distributed ADMM algorithm (GA-ADMM) based on general form consensus is proposed. First, the GA-ADMM algorithm filters the information transmitted between nodes by analyzing the characteristics of high-dimensional sparse data set: only associated features, rather than all features of the model, need to be transmitted between workers and the master, thus greatly reducing the communication cost. Second, the bounded asynchronous communication protocol is used to further improve the performance of the algorithm. The convergence of the algorithm is also analyzed theoretically when the objective function is non-convex. Finally, the algorithm is tested on the cluster supercomputer “Ziqiang 4000”. The experiments show that the GA-ADMM algorithm converges when appropriate parameters are selected, the GA-ADMM algorithm requires less system time to reach convergence than the AD-ADMM algorithm, and the accuracy of these two algorithms is approximate.

Spectrum sensing is one of the key technologies in cognitive radio systems. Efficient spectrum sensing can improve the communication network throughput and reduce the possibility of frequency collision. Hidden Markov Model (HMM) is a common spectrum sensing algorithm, which can enhance the energy detection (ED) algorithm by using historical observation information under unsupervised conditions. However, this algorithm assumes the regularity of the primary user occupying the spectrum to obey the Markov property. If the assumption is inconsistent with the facts, the performance of the algorithm will deteriorate. So, we propose a spectrum sensing algorithm based on Hidden Semi-Markov Model (HSMM) in this paper. It can solve the shortcoming of HMM because it has a high-order timing representation capability. Numerical simulations show that this model can effectively improve the detection performance of ED. It improves the SNR tolerance of 4 dB, or shortens the sensing time to a quarter of the time that the traditional ED method takes. In addition, the proposed algorithm is applicable to more scenarios than HMM. When the Markov property of the spectrum state fails, the proposed algorithm still performs better than HMM.

Detection techniques of micro-motion targets have been explored with increasing attention according to its complex and flexible features. In this paper, concepts and existing achievements of micro-motion and micro-Doppler are summarized horizontally from two aspects: micro-motion analysis foundation and techniques, strategies and implement. Addressing this goal, a general micro-Doppler formula is introduced with four typical micro-motion forms. Moreover, several extraction and imaging methods are demonstrated from four perspectives, i.e. radar quantity, micro-motion complexity, other strategies and potential problems. Subsequently, available application on ballistic target recognition and critical issues of this emerging field are proposed, with a prospect towards the trend of development.

This paper aims to provide an insight into the roles of the different types of airports in China by improved K-means clustering algorithm. The first part of the work analyzed the characteristics of Chinese airline network and pointed out that the key to construct hub-and-spoke airline network is determining the function of each airport. The index system of airport function orientation was established from airport operation index, airport hinterland index and airport growth index. The airports in China were classified into four classes by the K-means clustering algorithm. In order to improve reliability of clustering algorithm, a formula was used to normalize the value of each index, and the airports were clustered by improved K-means clustering algorithm. The algorithm was simulated by the MATLAB and the clustered results show the airports have obvious hierarchy.

This paper focuses on the physical layer security for spatial modulation (SM) based indoor visible light communication (VLC) systems with multi-LED transmitters, a legitimate receiver and multiple eavesdroppers. According to the principle of information theory, a lower bound on the SM-based VLC secrecy outage probability (SOP) is derived by considering the non-negativity, average optical intensity and peak optical intensity constraints. Numerical results show that the lower bound of SOP can be used to evaluate system performance.

Invoice reimbursement is one of indispensable aspects of business in many countries especially in China. Conventional manpower based reimbursement schemes often lead to high cost and inefficiency and robot based reimbursement systems require large space and huge equipment costs. In order to solve these problems, we propose an smart phone aided reimbursement system to realize the intelligent localization and identification in invoice images. First, invoice image is taken by camera of smart phone. Second, the Hough transform is used to detect the linear principle to correct the tilt of the invoice image with different background and different tilt angles. Third, we adopt You Only Look Once-Version 3 (YOLOv3) based target detection network to train the tagged data set, to obtain the training weights, and then realize the intelligent positioning and extraction. Finally, the invoice information is identified using optical character recognition (OCR). Experiment results are given to verify that the localization accuracy can reach 92.5% when the intersection over union (IoU) is set as 0.5 and the identification accuracy can reach up to 97.5% for invoice information.

In reverberant and noisy environments, tracking a speech source in distributed microphone networks is a challenging problem. A speech source tracking method based on distributed particle filter (DPF) and average consensus algorithm (ACA) is proposed in distributed microphone networks. The generalized cross-correlation (GCC) function is used to approximate the time difference of arrival (TDOA) of speech signals received by two microphones at each node. Next, the multiple-hypothesis model based on multiple TDOAs is calculated as the local likelihood function of the DPF. Finally, the ACA is applied to fuse local state estimates from local particle filter (PF) to obtain a global consensus estimate of the speech source at each node. The proposed method can accurately track moving speech source in reverberant and noisy environments with distributed microphone networks, and it is robust against the node failures. Simulation results reveal the validity of the proposed method.

Cognitive radio provides the ability to access the spectrum that is not used by primary users in an opportunistic manner, enabling dynamic spectrum access technology and improving spectrum utilization. The spectrum prediction plays an important role in key technologies such as spectrum sensing, spectrum decision, spectrum sharing and spectrum mobility in cognitive radio. In this paper, aiming at the spectrum prediction problem in cognitive radio, a spectrum prediction technique based on the sequence to sequence (seq-to-seq) network model constructed by the GRU basic network module is proposed. Due to the long and short time memory function of the GRU network structure, its performance is better than the previous Multi-Layer Perception (MLP) network model. This paper also explores in depth the impact of changes in the length of the input sequence on the prediction results. And the proposed seq-to-seq network model also performs well for multi-slot prediction and multi-channel joint prediction.

When calculating the traditional image compression storage algorithm, the key frames of the video image are mainly extracted by the video image feature. In the process of video image acquisition, the influence of factors such as light is detected, and the image features are changed, resulting in a large storage problem. A new anti-noise compression storage algorithm for big data video images is proposed. First, the collected big data video images are divided. The average value of the gray of the image sub-region is obtained, and then the compression process and the stored procedure are given. The actual working effect of the algorithm is verified by comparison with the traditional algorithm. The experimental results show that the improved algorithm is well stored and the error is small. The fast anti-noise compression storage method for the big data video images studied in this paper has a good storage effect, and its application range is wider and more worthy of promotion.

How to effectively mine students’ behavior data is an important content to improve the level of student information management. The platform of student behavior analysis and prediction based on campus big data is established, and the value of big data produced by students’ campus behavior is analyzed. The behavior data of students’ consumption laws, living habits and learning conditions are collected, modeled, analyzed and excavated around the large data environment, and the student behavior is predicted and warned by the stratified model of students’ behavior characteristics. The experimental results verify the effectiveness of the methods used, and the behavior characteristics can be analyzed according to the behavior characteristics of the students, and the students’ behavior will be guided to the overall health direction in a timely manner.

Conventional model of cooperative behavior mining method, can carry on the analysis, data mining to the conventional collaborative behavior but for specific subject knowledge in innovation ecosystem cooperative behavior, and analysis of the data mining results shooting low deficiencies, therefore puts forward innovation ecosystem in knowledge collaborative behavior main body model of the mining method. Based on knowledge innovation ecosystem in the main body composition analysis of collaborative behavior model, used algebraic representation, data processing design collaborative behavior model, realized the coordinated behavior model of innovation ecosystem knowledge subject data processing; According to the parameter fitting of collaborative behavior of knowledge subject in innovation ecosystem, the mining results were displayed to realize the model mining of collaborative behavior of knowledge subject in innovation ecosystem. The experimental data show that the proposed collaborative behavior model mining method is 41.84% higher than the traditional mining method, which is suitable for the model mining of collaborative behavior of knowledge subjects in the innovation ecosystem.

Conventional electrical circuit leakage fault data acquisition technology of leakage fault information collection, Failure to eliminate noise interference, resulting in failure to achieve real-time acquisition of circuit leakage fault data. There is a problem of low data accuracy and large noise interference, therefore put forward based on signal trigger electrical wiring leakage fault data automatic acquisition methods. Electrical wiring leakage fault detection based on signal trigger automatic acquisition mechanism, structures, acquisition model system, the acquisition model system hardware, electrical wiring to realize automatically leakage failure data acquisition model building; Automatically determine the leakage failure data acquisition software workflow, based on the leakage current fault detection algorithm and software anti-interference design, implementation is based on signal trigger automatic electric circuit leakage failure data collection. The experimental data show that the proposed automatic collection method is 35.24% more accurate than the traditional collection method, which is suitable for automatic collection of leakage fault data of different electrical circuits at different times.

The traditional forecasting system of agricultural products transportation logistics cargo flow relies too much on people’s subjective experience in forecasting, and the forecasting results are not accurate enough. To solve this problem, based on the large data analysis, a new forecasting system of agricultural products transportation logistics cargo flow is studied. The hardware and software parts of the system are designed, the hardware of the system consists of five parts: data collector, data analyzer, matcher, processor and tracer. The internal composition of each construction is described accurately. The working process of software is information input, information analysis, information matching, information processing and information tracking. The software workflow diagram is given. The results of the system are validated by comparing with the traditional cargo volume prediction system. The experimental results show that the system has high intelligence and can accurately predict the volume of goods transported in a short time. It has important guiding significance for the development of agricultural products transportation.

The development of big data technology and data mining technology has brought new opportunities for the scientific and innovative development of ideological and political education in colleges and universities. The evaluation of ideological and political education in colleges and universities in the context of big data was studied in this paper. An evaluation method of college students’ ideological and political education based on data mining was proposed. The proposed method uses K-means clustering method to analyze the data of the “worker’s assessment scale” of the counselor, and can achieve the evaluation of the ideological and political management effect of the counselor. The experimental results show that compared with traditional evaluation methods, the evaluation results of this method are more accurate and objective.

In the process of bio-fermentation, there is a problem of low detection efficiency in the process of recording changes in the concentration of traditional bacterial cells. Therefore, a real-time detection system for the concentration of microbial cells in biological fermentation is designed. In the system hardware design process, the data of microbial concentration changes in the biological fermentation are analyzed to select the system measurement principle. An intermediate conversion circuit is designed based on the measurement principle to complete the system hardware design. The measurement principle is used to derive the software structure of the real-time detection system. Real-time data acquisition and detection are implemented in the software structure to realize system software design. According to the results of simulation experiments, the real-time detection system for the change of bacterial concentration in biological fermentation compares with the traditional detection method, the detection efficiency is improved by 11%, and the operation is stable.

Because the traditional network security situation compound prediction system cannot overcome the defects of SVM algorithm, the accuracy of extraction results is low. For this reason, a large-scale network security situation compound prediction system is designed. Through data normalization process to optimize the SVM algorithm, to optimize the forecasting calculation module, to provide data base system frame structure, system frame structure can be divided into security situational composite sensing module, situational composite evaluation module and situational composite prediction module, synergy is derived using multiple module network security situational values when attacked, to implement network security situation prediction to complete the system design. Simulation application environment design compared the experimental results show that compared with the traditional prediction system of the proposed system under the same data to forecast, the accuracy of predicted results by 65%, and the operation is very stable.

In order to solve the above problem, a method of measuring the back-up path signal decline based on wireless network is designed to solve the problem that the existing signal measurement method can not express the specific decline path of the signal. Through the two steps of signal perception module design and measurement signal transmission mode, the wireless network environment of backup path signal measurement is completed. On this basis, three steps are carried out through the determination of the baseband signal fading frequency, the fading measurement channel estimation and the signal modulation measurement term to complete the construction of the new recession measurement method. The experimental results show that the fading path of the terminal signal is clearly expressed when the backup path signal fading measurement method based on wireless network is applied.

Routine athletes action spatial trajectory data retrieval method can perform the trajectory retrieval to athletes action, but there is the deficiency of low retrieval ability when only the detailed spatial trajectory retrieval of athletes stepping motion is performed, for this reason, the research on spatial trajectory retrieval method of athletes stepping motion data is proposed. Based on the extraction of characteristic information of athlete’s stepping motion, the stepping motion R-Tree and its variant space index are determined, and the construction of the spatial trajectory retrieval model of athlete’s stepping motion data is achieved. Based on the spatial trajectory index design of athlete’s stepping motion data, the spatial trajectory retrieval result is output, and the research on spatial trajectory retrieval method is completed. The experimental data show that the retrieval capability of proposed trajectory retrieval method for athlete’s stepping motion is 53.41% better than that of conventional trajectory retrieval, which is suitable for detailed spatial trajectory retrieval of athlete’s stepping motion.

When using traditional traffic congestion recognition method to judge traffic congestion, there is a lack of accuracy. In view of the above problems, a fuzzy identification method of regional traffic congestion based on GPS is proposed. First, the GPS floating vehicle traffic information collection technology is used to collect the traffic information of the road network, and it is pretreated at the same time. Then the effective data and the electronic map are matched to determine the accurate position of the floating car on the road. Finally, a fuzzy comprehensive discriminant model based on the GPS data is set up, and the road traffic status is entered. The line is accurate. The results show that the accuracy of the method is 44% higher than that of the traditional traffic congestion recognition method, which basically achieves the purpose of this study. Experimental results are better, this article can bring guidance meaning to the future research.

With the popularity of the Internet and the increasing power of video editing software, digital video can easily be tampered with. The detection of the authenticity and integrity of digital video is very important. A video tampering detection method based on multi-scale normalized mutual information is proposed. Firstly, the mutual information is introduced into video tamper detection and the normalized mutual information content of the video frames is extracted. Then, based on the “scale invariance” feature of human vision, the mutual information between frames is analyzed from a multi-scale perspective. The multi-scale normalized mutual information is used to characterize the similarity of content between video frames. Finally, the LOF algorithm is used to calculate the degree of abnormality of the similarity coefficient sequence to achieve three kinds of tampering detection in the time domain: deletion, insertion, and replication. Experimental results show that the proposed method can effectively detect tampered video.

With the rapid development of information technology, information overload has become an important challenge of Internet. In order to alleviate the growing contradiction between users and massive data, the researchers proposed the concept of the cross-harmonic recommender system. By analyzing characteristic of datasets, recommendation algorithms and method for weight calculation, we introduced a fast and general engine for large-scale data processing and implemented the cross-harmonic recommender system based on Spark, aiming at improving accuracy, diversity and efficiency of the recommender system.