281 results about "Correlation matching" patented technology

The invention relates to the technical field of industrial production intelligent process control, in particular to an AI-based electrolytic manganese process production quality intelligent prediction method and system, and the system comprises a drift calibration module, a structure matching module, a process parameter inversion module, a path correction module and a deviation verification module. According to the method, dynamic compensation of deviation between input parameter drift and output parameters is achieved through the response relation between input and output of all links, the actual influence degree of fluctuation of all the parameters on the target control parameters and product quality is effectively recognized according to correlation matching of the periodic change of the production process and the fluctuation frequency of the target control parameters, and the accuracy of the product quality is improved. Potential interference of key process parameters on stability of each process is locked in time, diffusion starting time and space of impurities are clearly traced through path judgment of residual liquid concentration trend and raw material adding position, and migration path parameters are further adjusted based on ion retention behaviors of boundary detection points.

The invention discloses a time-frequency synchronization tracking and judging method based on GFSK (Gaussian Frequency Shift Keying). The method comprises the following steps: S1, positioning the initial position of a synchronization word by detecting a lead code; s2, extracting a fixed synchronization word, completing frame synchronization based on correlation matching, and positioning a starting point of an effective data segment; s3, a range cycle counter collects a central sampling point and two adjacent sampling points to form a candidate sampling point sequence; s4, dynamically and finely adjusting the range cycle counter to form a time sequence calibration sampling point sequence; s5, performing phase difference calculation on the time sequence calibration sampling point sequence, estimating continuous frequency offset and generating a compensation value; s6, applying the compensation value to construct a decision feedback equalizer with delay front and back terms; s7, performing interference suppression on the compensation sampling point sequence by using a decision feedback equalizer with delay front and back items, and outputting a soft decision value; and S8, executing three times of judgment, constructing a sliding reference sequence, and outputting a final bit judgment result. According to the method, the problem of poor GFSK demodulation robustness under Doppler disturbance is solved.

The invention provides a multi-agent linkage scheduling method and system applied to cross-domain information processing, and the method comprises the steps: firstly receiving a cross-domain information processing demand, carrying out the structural analysis, obtaining a demand feature set, calling a multi-agent capability feature library to carry out the correlation matching, generating an agent preliminary matching result set, and obtaining a multi-agent linkage scheduling result set; constructing a multi-agent linkage constraint condition based on the matching degree parameter and the task priority parameter, generating a linkage scheduling strategy, sending the scheduling strategy to a candidate agent to execute a task, collecting state feedback information in real time, and dynamically adjusting the scheduling strategy according to the state feedback information. And generating an adjusted strategy and synchronizing the adjusted strategy to each candidate agent to update task execution parameters, thereby realizing intellectualization, automation and dynamics of cross-domain information processing, and improving information processing quality and efficiency.

The invention provides an elevator emergency rescue data intelligent analysis and early warning method and system, and relates to the technical field of data analysis, and the method comprises the steps: collecting operation parameter data, carrying out preprocessing, carrying out local calculation through an elevator safety evaluation model, outputting a safety evaluation score, determining a safety level, and judging whether to trigger an early warning mechanism or not. If triggered, early warning information is sent, early warning information classification and summarization are carried out, a historical early warning data sequence is obtained, feature learning is carried out, deep features are generated, fault probability distribution is predicted, key omen factors are identified, and an elevator fault early warning report and a preventive maintenance strategy are generated; whether an alarm mechanism is triggered or not is judged, if yes, alarm information is sent, correlation matching, representation learning and semantic label information extraction are carried out, a graph-based Markov decision process is modeled, a fault diagnosis report is obtained through solving, and a rescue action combination is selected to obtain an initial rescue scheme; and an optimal rescue scheme is obtained through optimization and sent to a safety management department.

The invention provides a document-level relation extraction method based on mixed prompt tuning, which comprises the following steps of: constructing a document-level knowledge graph module by utilizing historical documents, and acquiring and processing entity and relation structure information; a sentence-level semantic retrieval enhancement mechanism is designed, and context information is matched and retrieved based on semantic relevancy; providing a mixed prompt tuning method, and performing efficient parameter fine tuning on the large language model; and finally, aggregating the document-level knowledge graph and sentence-level semantic retrieval to enhance and generate an output result of a SetRAG mechanism, and carrying out document-level relationship prediction. According to the method, the domain knowledge based on the atlas and the semantic knowledge based on retrieval are combined, so that the problems that entities and relation interaction information are difficult to capture and cross-sentence relation reasoning is difficult in existing document-level relation extraction can be effectively solved, and meanwhile, efficient and accurate document-level relation extraction is realized; and the performance of the public test data set is obviously improved.

The invention discloses an expert matching method and system based on large language model knowledge extension and fusion. The method comprises the following steps: firstly, collecting expert basic information, and enhancing through a pre-trained big language model to obtain expert extension information; and then, integrating the basic information and the extended information, and analyzing and fusing the basic information and the extended information by using RoBerta and a knowledge graph pre-training model to form expert fusion representation. According to enterprise technical requirements, requirement details are expanded through a large language model, and technical requirement fusion representation is generated in combination with representation analysis of the technical requirements and enterprise information. And finally, carrying out keyword correlation screening on the experts, narrowing a matching range, comparing fusion representation of the experts and requirements, calculating similarity and sorting, and recommending related matching experts. Through the method and the device, skill matching analysis can be effectively performed on external expert information, and expert matching precision and recommendation effect are improved based on semantic analysis and expert matching of technical requirements.

The invention discloses a target tracking method and device, electronic equipment and a storage medium. The invention relates to the technical field of computer vision. The method comprises the following steps: acquiring a detection target frame of a to-be-tracked target in a current image frame; a prediction target frame of a to-be-tracked target in the current image frame is obtained, the prediction target frame is obtained through prediction based on a Kalman filter, and a state vector of the Kalman filter comprises a position change second derivative; performing generalized intersection-to-union ratio correlation matching on the detection target frame of the to-be-tracked target in the current image frame and the prediction target frame of the to-be-tracked target in the current image frame to obtain a matching result; and under the condition that the matching result meets a target tracking success condition, performing target tracking on the next image frame. According to the technical scheme, by introducing the position change second derivative, utilization of the change rate of the target movement speed is achieved, the accuracy of the prediction target frame can be improved, the prediction target frame is closer to a real target, and therefore the target tracking precision is improved.

The invention relates to a customer resource management system and method based on a large model. The method comprises the following steps: carrying out semantic analysis on historical sales data of sales personnel and new customer resources by adopting a natural language processing technology based on deep learning, mining professional experience information of the sales personnel and potential demand information of the customer resources, and carrying out fine-grained correlation matching analysis on the professional experience information and the potential demand information; in this way, whether the new customer resources are matched with the ability and style of the salesman or not is intelligently judged, and therefore accurate pushing and optimal management of the customer resources are achieved. In this way, the efficiency and accuracy of customer resource allocation can be remarkably improved, the invalid communication cost in the sales process is reduced, the customer satisfaction is improved, and enterprises are assisted in achieving continuous increase of sales performance.

The invention discloses an intelligent security check method and system based on image recognition. The method comprises the following steps: acquiring initial multi-mode security check data in a security check system; performing noise reduction processing on the data, establishing a hybrid neural network based on local feature extraction and global context relation captured by a parallel module, and performing dynamic weighted fusion on multi-modal features through a cross attention mechanism; inputting the noise-reduced multi-modal security check data into a hybrid neural network for identification to obtain a security check image suspicious region; and carrying out association matching on the security check image suspicious area by using a Hungary algorithm, and carrying out early warning according to a matching result. Articles with complex shapes and similar densities are effectively distinguished, false detection and missing detection are reduced, and the security check accuracy is improved.

The invention discloses an intelligent networked automobile behavior prediction method based on a fusion recognition algorithm, and relates to the technical field of automatic driving, and the method comprises the steps: carrying out the correlation matching of a sensor object and a tracker object; performing data fusion on the tracker object and the sensor object, and updating the tracker object according to a fusion result; converting the coordinate information of the tracker object into a world coordinate system to obtain a fusion recognition result; performing feature extraction on the fusion recognition result; constructing a training set, a verification set and a test set; constructing an initial prediction model by using the training set; evaluating the initial prediction model by using the verification set; adjusting and optimizing the adjusted prediction model by using the test set; and preprocessing a fusion identification result obtained in real time, and inputting the preprocessed fusion identification result into the final prediction model to obtain a prediction result. According to the invention, the real-time problem of the current behavior prediction system is solved, the accuracy of the prediction result is improved, the traffic safety level is improved, and the stability and comfort of planning control are improved.

The invention discloses a PRNU-based camera source identification method, belongs to the technical field of camera traceability, and aims to solve the technical problem of how to improve the estimation quality of camera fingerprints and improve the accuracy of correlation scores in camera source identification. Comprising the following steps: constructing a camera fingerprint estimation model, introducing a channel attention mechanism into the camera fingerprint estimation model to perform feature extraction and weight distribution on an input image to obtain a weight feature map, extracting a PRNU of the input image through a PRNU extraction network, fusing the weight feature map and the PRNU of the input image, and outputting a camera fingerprint; and a correlation matching model is constructed, and the correlation matching model takes the PRNU of the query image and the camera fingerprint of the target camera as input, and predicts and outputs a correlation score between the query image and the target camera. And for a query image to be traced, outputting a correlation score between the query image and the target camera through the trained correlation matching model.

The invention relates to the technical field of aero-engine detection, in particular to an aero-engine blade tracking and counting method, which comprises the steps of aero-engine blade tracking and aero-engine blade counting: step 1, adopting a blade tracking algorithm based on a TBD framework to realize aero-engine blade tracking, comprising: step 1, target detection; 2, motion modeling and state estimation; 3, association matching; step 2, counting the aero-engine blades based on the trajectory information calculated in the step 1: step 1, identifying newly appearing aero-engine blades; 2, judging the rotating direction of the newly-appearing aero-engine blade and recording the starting direction of the newly-appearing aero-engine blade; and 3, updating the current count value according to the movement direction of the aero-engine blade. According to the invention, the problem of missing detection in the manual blade counting process can be solved, the problem of repeated detection of aero-engine blades by workers due to missing detection is solved, and the working efficiency of the workers is improved.

The invention discloses a role playing model fusion system and method based on retrieval and reflection, and belongs to the technical field of natural language processing. The implementation method comprises the following steps: 1, performing data cleaning on original role historical dialogue data, and performing semantic screening by using semantic similarity to form historical dialogue context data; 2, event knowledge associated with the roles is constructed through association degree matching; 3, constructing a role reflection strategy mode by using historical dialogue data, and generating role reflection data; 4, when a user inputs a question, replying by adopting a multi-source fusion mode based on role thinking; compared with the prior art, in a multi-scene task, a dynamic retrieval role knowledge base is utilized to screen task related background knowledge, and a thinking mode modeling framework based on a role reflection mechanism and a context fusion strategy are constructed to generate high-quality answers conforming to the context; and thus, the fusion ability of the role playing model is improved, and the consistency of role words and language styles is realized.

The invention discloses a base station power supply fault intelligent diagnosis method and system based on anomaly detection, and the method comprises the following steps: S1, collecting and preprocessing multi-modal monitoring data, and generating a multi-dimensional time series data set; s2, inputting the multi-dimensional time sequence data set into an improved ResNet-18 network for feature extraction, and generating a power supply fault feature vector; s3, inputting the power supply fault feature vector into an improved bee colony algorithm for anomaly detection, and outputting a power supply abnormal state mark and a power supply fault mode; s4, outputting a power supply fault diagnosis label through the classification function; s5, calling a power supply fault diagnosis label mapping rule, and outputting a power supply fault diagnosis result; and S6, based on the correlation matching degree, adjusting the convolutional layer weight of the improved ResNet-18 network and the fitness function parameter of the improved bee colony algorithm. According to the method, deep residual network modeling and a swarm intelligence anomaly identification mechanism are fused, and accurate identification, adaptive optimization and closed-loop diagnosis of the power supply fault of the base station are realized.

The invention discloses an automatic frequency stabilization system based on dual-core correlation matching intelligent peak searching. The automatic frequency stabilization system comprises the following steps that S1, a saturated absorption spectrum light path and a circuit are built, debugged and optimized; s2, constructing a double-matching template; s3, triangular wave voltage scanning and clock timing are started, and signals are collected and transmitted; s4, starting a KCF algorithm, applying a template to track model matching, judging a result obtained by matching, finding a peak point and solving a corresponding scanning voltage; and S5, recording a peak point scanning voltage, closing a scanning signal, starting sinusoidal signal modulation, and giving a triangular wave scanning voltage corresponding to a peak value to the sinusoidal signal as an offset. The spectrum line peak value suspected area can be effectively positioned, the double-branch and correlation filtering tracking method can perform precise secondary searching in the suspected area, the locking precision is greatly improved, the long-term stability of 10-12 orders of magnitude is ensured, and meanwhile the good vibration interference resistance can be achieved when a vibration signal occurs.

The invention discloses a landing positioning method and system based on fusion of vision and inertial navigation. The method comprises the following steps: collecting vision information of a current landing area; the visual information of the front landing area is mapped to an aligned grid of a map coordinate system through homography transformation, correlation matching is conducted on the visual information and a map through a frequency domain FFT algorithm, a matching result of rough matching is obtained, and rough matching is completed; the method comprises the following steps of: mapping an area subjected to rough matching and positioning to a grid aligned with a map coordinate system by utilizing homography transformation, performing spatial domain multi-window division on a larger range of images around the area subjected to rough matching and positioning, selecting a landmark area in each sub-window by applying an interest operator, performing spatial correlation matching on the landmark area and a map by utilizing a spatial correlation method, and performing spatial correlation matching on the landmark area and the map by utilizing a spatial correlation method. Finishing fine matching; inputting all matched observation results into an extended Kalman filter, and carrying out joint estimation on the matched observation results and the observation results of the inertial measurement unit; the method has the advantages that the robustness, the precision and the engineering applicability of an existing navigation positioning method are improved.

The invention discloses a BIM-based intelligent construction progress management system, electronic equipment and a storage medium, and the system comprises a component decomposition module which constructs a BIM model based on a current construction plan state, carries out the component decomposition of the BIM model through geometric analysis, and carries out the correlation matching of the split components and construction sections; the data synchronization module is used for collecting actual construction data of a construction site, extracting multi-source features of the construction data and fusing the multi-source features to obtain fused multi-source features; the dynamic evaluation module is used for dynamically evaluating the deviation between the actual progress and the planned progress based on the fused multi-source features and the associated and matched component information; and the deviation correction adjustment module is used for generating a deviation correction strategy based on the deviation by using a heuristic optimization algorithm, and completing intelligent construction progress management based on the deviation correction strategy.

The invention provides an AI intelligent recommendation method and system applied to an online patent operation digital platform, and the method comprises the steps: obtaining user full-amount behavior data and patent full-period attribute data, carrying out the demand intention analysis and value dimension modeling, generating a user demand intention feature set and a patent value dimension feature set, and carrying out the calculation of the user demand intention feature set and the patent value dimension feature set; and on this basis, an AI intelligent recommendation model is used to carry out multi-dimensional association matching on the two, an initial recommendation candidate set is generated, dynamic adjustment is carried out based on real-time interaction information of the user, and finally an optimized patent recommendation sequence is generated and output to a user terminal interface, so that precision and dynamic patent recommendation are realized, and the user experience is improved. And the patent operation efficiency and the user experience are effectively improved.

The invention discloses a mining area subsidence dynamic monitoring method, device and equipment, a medium and a product, and relates to the field of mining area mining subsidence monitoring. The method comprises the following steps: firstly, carrying out data preprocessing on an unmanned aerial vehicle image and laser radar point cloud time sequence data observed in the same period to obtain DOM and DEM data in each period; based on each period of DOM data, extracting east-west and south-north horizontal displacements between any two periods of DOM data by using a normalized cross-correlation matching algorithm; for the newly-added DOM data, dynamically updating historical and newly-added east-west and south-north direction horizontal displacements according to a sequential least square adjustment principle to obtain adjusted east-west and south-north direction horizontal displacements; carrying out plane position correction on DEM data of each period based on the adjusted horizontal displacement in the east-west direction and the south-north direction; and performing difference calculation on the corrected DEM data of each period and the reference DEM data before deformation to obtain time sequence subsidence data of the mining area, thereby realizing dynamic monitoring of mining area subsidence, and improving mining area subsidence monitoring precision and efficiency.

The invention discloses a fault self-diagnosis method and system for a temperature transmitter under real-time monitoring, and relates to the technical field of fault self-diagnosis, and the method comprises the steps: collecting the spatial position information and working state parameters of high-frequency electrical equipment, carrying out the electromagnetic interference analysis through combining the spatial position information of a target temperature transmitter, and generating an electromagnetic interference distribution map; establishing an electromagnetic-temperature influence correlation, arranging an electromagnetic interference sensing module to obtain an electromagnetic interference change curve, and monitoring a temperature signal fluctuation curve in real time; performing association comparison analysis to generate an association matching result; and when abnormal matching occurs in the correlation matching result, temperature transmitting signal reconstruction is carried out to generate transmitter compensation parameters for transmitting signal correction, and fault risk early warning is carried out. According to the invention, the technical problem in the prior art that the fault is difficult to find in time and position accurately when the temperature transmitter is influenced by electromagnetic interference is solved, and the technical effect of improving the accuracy and real-time performance of fault diagnosis of the temperature transmitter is achieved.

The invention discloses a monitoring image multi-target tracking method based on deep learning, and relates to the technical field of digital image processing, and the method comprises the following steps: S1, generating a multi-scale feature pyramid; s2, extracting causal feature vectors; s3, generating a discrete codebook index; s4, constructing a dynamic graph; s5, inputting the dynamic graph into an improved ASTGCN network, iteratively fusing space-time attention weights and Hamilton dynamics evolution characteristics between nodes through cascaded Hamilton space-time blocks, and generating a final prediction state of each historical track node; s6, solving an optimal correlation matching matrix by using a Hungary algorithm; and S7, extracting all active track information. According to the method, the limitations of identity drift caused by variable appearance characteristics and inaccurate prediction caused by lack of physical priori in a traditional multi-target tracking method are overcome, and an efficient, accurate and robust solution is provided for intelligent video monitoring.

The invention discloses an image retrieval method and device, and relates to the technical field of knowledge distillation, and the method comprises the steps: taking a query text input by a user as a retrieval, and screening and outputting a target image consistent with text content description from a preset image library through text semantic comprehension and an image feature matching algorithm, the technical problems that in the related technology, it is difficult to deeply mine core knowledge such as fine-grained probability distribution and coarse-grained structural features of an attention mechanism in the image retrieval process, a distillation process cannot cover full-dimensional knowledge of an attention level, and the integrity of knowledge migration is insufficient are solved. The technical effects that through deep correlation matching of the text and the image, the user can be helped to rapidly position the target image through simple text description, the image search operation process is effectively simplified, and efficient and convenient image retrieval and browsing experience are provided for the user are achieved.

The invention provides a health big data-oriented production and education integrated intelligent prediction and implementation method thereof, and the method comprises the steps: collecting the internship work data of internship students through the connection with a hospital information system, carrying out the correlation matching of the collected internship work data and a student health record data set, and constructing a production and education integrated health big data platform; according to the birth and education fusion health big data platform, individual dietary habit factors, work and rest rule factors and internship work intensity key features of the students are extracted, and the probability of gastrointestinal diseases of the students in the internship period is predicted by adopting a decision tree model; and obtaining an internship performance score and an internship unit evaluation in the internship period of the student, performing correlation analysis with the health condition data in the internship period of the student, and determining a correlation rule between the internship performance and the health level of the student.

The invention relates to the technical field of model training, in particular to an AI algorithm model training management system, which comprises the steps of analyzing a data value association matching degree in an AI algorithm model training process based on training data information and model deployment environment information of an AI algorithm model in the model training process; in combination with training configuration information and model architecture information of the AI algorithm model in the model training process, the effective utilization degree of computing resources in the AI algorithm model training process is analyzed; evaluating the training effect of the AI algorithm model based on the data value association matching degree and the effective utilization degree of the computing resources; according to the training effect evaluation result, the training scheme of the AI algorithm model is adjusted, the reliability and efficiency of the training process of the AI algorithm model can be remarkably improved, and the high performance and high robustness of the AI algorithm model in a real scene are ensured.

The invention belongs to the technical field of flow measurement, and particularly relates to an ultrasonic flow measurement method which remarkably improves measurement precision and system energy efficiency by intermittently updating zero flow velocity reference flight time and pre-storing zero flow velocity waveform data in combination with a correlation algorithm and a parabola interpolation technology. Compared with a traditional threshold value method and a TI scheme, the method abandons threshold value dependence, dynamic correlation matching is carried out by utilizing a standard waveform of factory calibration and a real-time signal, absolute flight time is calculated, and a wave hopping error caused by long-term signal degradation is effectively avoided; meanwhile, a pre-stored waveform template and a regular self-calibration mechanism suppress noise, temperature drift and transducer aging influences, breakthrough is achieved in the aspects of anti-interference performance, long-term reliability and low power consumption, and a high-cost-performance solution is provided for gas flow metering.

The invention discloses a community comment management method and system, and relates to the technical field of data processing, the method can capture user interest changes in real time based on a dynamic interest map generated by registration information and short-term interaction behaviors, quickly screen high-correlation comments through association matching of interest levels, and improve the user experience. It is ensured that the recommended content is deeply matched with user preferences, and meanwhile the redundant data processing burden is greatly reduced. And through comprehensive analysis of a multi-modal quality evaluation system fused text, image and user behavior, and dynamic adjustment of community type differentiation weight distribution and real-time interactive feedback, false, contradictory or low-quality content is effectively identified, and compliance review and high-quality content exposure are enhanced. According to the embodiment, high-quality comments are accurately recommended through the dynamic interest map and multi-modal evaluation, the content ecology is optimized in real time, the timeliness is guaranteed, and the system stability is enhanced.

The invention discloses a small sample point cloud semantic segmentation method, computer equipment and a storage medium, and relates to the technical field of three-dimensional computer vision, the small sample point cloud semantic segmentation method mainly comprises the following steps: according to a support set point cloud and a query set point cloud, using an embedded network to obtain support set features and query set features, using a feature correlation matching module to obtain enhanced support set features and enhanced query set features; according to the enhanced support set features, obtaining a support prototype by using a similarity perception prototype number constraint method and a multi-prototype generation method; according to the enhanced query set features and the support prototype, utilizing a prototype correction module to obtain a corrected support prototype; and according to the enhanced query set features and the corrected support prototype, obtaining a prediction mask by using a KNN graph construction and label propagation method. By implementing the small sample point cloud semantic segmentation method, the computer equipment and the storage medium provided by the invention, the small sample point cloud semantic segmentation precision can be improved.

The invention relates to a customer service method based on seat text processing analysis. A pre-trained pre-analysis model analyzes the business field of customer demands and provides the reply difficulty of the customer demands; when the difficulty does not exceed a difficulty threshold value, the pre-analysis model adds the business field to a cue word of a pre-trained first language model; matching reply or configuration operations related to customer demands in cue words from contents of corresponding business fields in a customer service vector library through correlation matching, wherein the first language model is limited by the system cue words to organize reply or configuration operations with correlation ranked in front as reply results; submitting a reply result obtained by the customer demand and the first language model to the second language model to verify whether the reply result is valid, if so, replying to the customer, otherwise, providing an invalid reason by the second language model; and when the difficulty exceeds a difficulty threshold value, the customer demand is switched to a manual service module according to the business field, and idle customer service in the corresponding field is configured for service.

The invention discloses a power bank with a data storage function and an electromagnetic interference suppression method thereof, and the method comprises the steps: obtaining the operation parameters of a wireless charging coil and the operation indexes of a high-speed data link, and carrying out the segmentation statistics and time-frequency correlation matching processing, thereby obtaining an interference characteristic parameter set; performing spread spectrum migration and duty ratio-phase joint modulation on the coil driving signal; executing cache rearrangement and time sequence revision on the to-be-sent data frame; carrying out adaptive adjustment on the error control coding parameter according to the interference intensity value; and performing comprehensive optimization processing on each parameter set to generate an updated control parameter set, and forming a closed-loop control system. According to the technical scheme, under the condition that hardware shielding is not increased, fine cooperative control over the charging magnetic field and data transmission can be achieved through a software algorithm, electromagnetic interference is effectively restrained, wireless charging and high-speed data transmission can be stably and parallelly operated under the condition that space is limited, and the reliability of data transmission and the overall user experience are improved.

The invention discloses a multi-target water floating garbage tracking method based on D-StrongSORT (Dempster-StrongSORT). Relates to the technical field of image recognition and tracking. A D-StrongSORT multi-target tracking algorithm combining an appearance model and a motion model is used for tracking floating garbage, a previous detection result video is used as input, a previous position state is associated with a next position state by calculating motion features and appearance features, prediction and tracking of a target position can be well achieved, DIOU is adopted by the D-StrongSORT to replace IOU, and the prediction and tracking efficiency is improved. According to the method, a new feature updating strategy is used to match appearance features of a tracking target to reduce ID switching times of the tracking target, adaptive Kalman filtering is used to reduce error accumulation of a motion model, a new feature updating strategy is used to match the appearance features of the tracking target when an appearance model is established, and better correlation matching of the motion model and the appearance model is realized by designing adaptive weighted cost matrix calculation.