114results about How to "Increase weight" patented technology

The invention relates to the technical field of information security, and discloses a terminal clipboard sensitive information cross-application leakage real-time interception method, which comprises the following steps of: registering a clipboard access hook function in a kernel layer or a system framework layer of a terminal operating system, and capturing a clipboard reading operation in real time; comparing an application program identifier corresponding to the captured reading operation with an application program identifier of a clipboard owner, and judging whether the access is cross-application access or not; when the access is determined to be cross-application access, triggering a clipboard content scanning engine to perform sensitive information identification on texts, images and file paths in a clipboard; and obtaining an evaluation result based on a preset sensitive information rule base and the dynamic strategy engine. A clipboard access hook function is deployed in a kernel layer or a system framework layer, millisecond-level capture and blocking of cross-application clipboard reading operation are achieved, and traditional extensive management based on application permission declaration is upgraded to refined management and control based on content risk identification.

The invention relates to the technical field of computer vision, and discloses a camouflage target detection method based on cascade frequency domain perception and refinement feature guidance, comprising the following steps: a computer device obtains a camouflage image and extracts a multi-scale feature map; performing frequency domain separation by using a frequency domain sensing module, generating low-frequency and high-frequency characteristic sub-bands, and performing cross-hierarchy aggregation from bottom to top through a cascaded frequency domain sensor to obtain frequency fusion characteristics; receiving the frequency fusion features from top to bottom by using an aggregation guide decoding unit, performing guide refinement in combination with the deep decoding features, adapting to irregular edges by using a variable kernel convolution strategy in the refinement process, and generating a prediction feature map in combination with a partial convolution strategy; and outputting the predicted feature map as a detection result. According to the invention, through frequency domain information complementation and dynamic convolution refinement, the problem of difficult detection caused by highly similar texture of the camouflage target and the background is solved, and the positioning accuracy and edge segmentation integrity of the camouflage target are improved.

The application discloses a brain function state evaluation and psychological disease auxiliary analysis system based on a brain fluctuation map, comprising: a multi-modal electroencephalogram signal acquisition and preprocessing module; a brain fluctuation map feature extraction module; an individualized baseline modeling module; and an adaptive weighting evaluation module based on an attention mechanism, which constructs an individualized deviation vector and adaptively weights the current features and deviation features based on an attention network, and outputs a disease risk probability and a channel contribution weight map. Compared with traditional population norms or simple arithmetic means, the method can more truly reflect the stable ground state of the subject itself, effectively eliminate individual differences and measurement noise, and provide accurate dynamic reference system for subsequent abnormal deviation judgment. Meanwhile, the fluctuation threshold calculation unit sets an adaptive threshold based on a statistical confidence interval, and introduces an abnormal deviation index and a multi-dimensional joint determination, further reducing the false positive risk.

The application relates to a heat dissipation technology field of a fuel cell, in particular to a heat dissipation device integrated with anti-vibration and a fuel cell heat dissipation system, which comprises a shell, a fan arranged in the shell and a heat exchanger arranged on the side of the shell, and further comprises two opposite vertical arms arranged in the shell, a plurality of guide vanes corresponding to the gaps between two adjacent fins of the heat exchanger are rotatably arranged between the two vertical arms; the inclined guide vanes can change the flow direction of the airflow, so that the airflow can directly impact the wall surface of the fin, thereby effectively destroying the airflow layer wrapped on the fin due to vibration and not easy to flow, when the vibration intensity increases, the swing amplitude of the guide vanes increases, so that the stripping effect of the airflow on the low-speed airflow layer attached to the wall surface of the fin is more intense, the stripping effect of the airflow layer automatically matches the current vibration condition, the speed regulation mechanism increases the rotating speed of the fan, compensates for the additional wind resistance loss, and ensures that the convective heat exchange efficiency is maintained under different intensity vibration working conditions.

This invention discloses a method for constant temperature and humidity control of a cigar cabinet based on a fuzzy control algorithm, comprising the following steps: setting a target temperature and humidity range; collecting environmental data of each layer through a distributed sensor array and comparing it with the target range to obtain the temperature and humidity deviation of each layer; inputting the deviation into a temperature and humidity decoupling control module and outputting a weight adjustment signal; inputting the weight adjustment signal into a fuzzy controller to construct a multi-objective fitness function, optimizing it with the WFPSO algorithm, and outputting hierarchical fuzzy controller parameters; inputting the hierarchical fuzzy controller parameters into a hierarchical temperature and humidity model to monitor the status of each layer in real time, and outputting a drive command when the limit is exceeded; inputting the drive command and hierarchical parameters into an actuator, adjusting it, and outputting new environmental data; and using a feedforward control module to adjust in advance when a door opening / closing disturbance warning is issued, dynamically adjusting the status of the affected space layer. This invention improves the temperature and humidity control accuracy and disturbance adaptability of the cigar cabinet through hierarchical optimization and feedforward intervention control mechanisms.

This invention belongs to the field of distillation apparatus technology and discloses an automatic distillation range measuring instrument and its calibration method, including a housing, a distillation flask, and a tube cover. The left end of the housing has a working area A, and the right end has a working area B. A control box is fixedly installed on the top of the housing. A heating plate is fixedly connected inside the working area A. The invention simultaneously starts a first motor to drive a second pulley to rotate. This, in turn, drives the first pulley to rotate via a drive belt, causing the tube cover to rotate with the mounting end. The serpentine condenser then rotates, and small volumes of liquid adhering to the inner wall of the serpentine condenser, under centrifugal force, slowly slide along the inner wall and merge with other small volumes of liquid to form a larger volume of liquid. Simultaneously, propelled by gas, this liquid drips down the discharge pipe into the collection bottle, effectively cleaning the liquid inside the serpentine condenser and avoiding waste.

The application belongs to the field of vehicle-mounted refrigerators, and discloses a dual-temperature cooperative acquisition circuit, an intelligent temperature measurement method and a vehicle-mounted refrigerator, which comprise: a synchronous acquisition unit, which is used for synchronously acquiring temperature signals of at least two target regions of the vehicle-mounted refrigerator and outputting corresponding temperature analog signals; an anti-interference processing unit, which is connected with the output end of the synchronous acquisition unit and is used for filtering various interference signals in the vehicle-mounted environment; an adaptive signal conditioning unit, which is connected with the output end of the anti-interference processing unit and is used for adaptively amplifying, filtering and linearizing the temperature analog signals; a main control unit, which is connected with the output end of the adaptive signal conditioning unit and is used for digitizing the temperature signals, performing algorithm operation and logical control; and a wide-adaptive power supply unit, which is connected with the synchronous acquisition unit, the anti-interference processing unit, the adaptive signal conditioning unit and the main control unit and is used for providing a stable working voltage that is adapted to the vehicle-mounted power supply.

The invention discloses a rock core CT image crack segmentation method and device based on deep learning. The method comprises the following steps: making a training data set; expanding the number of training samples in the training data set to form a new data set; a deep learning network model is constructed, the deep learning network model comprises a U-Net network, a contraction path for feature extraction in the U-Net network adopts a feature extraction network based on a VGG-16 network, and an up-sampling expansion path in the U-Net network is added with a CBMA attention module. The CBMA attention module is used for changing the attention of the model to different spatial positions and different channels in the feature map and improving the weight of the area where the rock core crack is located; performing training and parameter adjustment on the deep learning network model by using the training set until the model converges; and testing the trained deep learning network model by using the test set, and evaluating the crack segmentation effect of the deep learning network model. According to the invention, the efficiency of rock core CT image automatic analysis work can be improved.

The invention discloses a power grid frequency prediction method based on LSTM and time feature attention. The power grid frequency prediction method comprises the following steps: step 1, collecting multi-dimensional time sequence feature data of a power system before and after disturbance; 2, preprocessing the multi-dimensional time sequence feature data, and extracting high-dimensional time sequence features in the feature data obtained in the step 1 by using a time convolutional neural network T-CNN; 3, inputting the high-dimensional time sequence features extracted by the time convolutional neural network T-CNN into the time-feature attention module to obtain a focusing key feature and a weighted feature sequence of a key moment; 4, taking the weighted feature sequence obtained in the step 3 as input, mining a time sequence dependency relationship through a frequency prediction model of a long short-term memory (LSTM) network, and outputting a prediction curve of the inertia center frequency of the power system; and step 5, according to the predicted frequency curve, calculating a frequency lowest point deviation and a frequency change rate, and according to a preset frequency response danger level standard, outputting a frequency safety evaluation result of the power system.

This invention discloses a method for constructing a fast-growing, sterile zebrafish model, belonging to the field of fish animal model preparation technology. This invention is the first to discover the role of the CK gene in regulating fish growth rate and fertility: fish with simultaneous knockout of ck1 and ck2 genes exhibit rapid growth and gonadal abortion. Using the method provided by this invention, a fast-growing, sterile zebrafish model was successfully constructed. These zebrafish exhibited gonadal dysplasia and sterility, but also rapid growth. Compared to wild-type zebrafish, the 4-month-old double mutant zebrafish showed significantly increased body length and weight. The disclosed gene function and zebrafish model construction method can be used for breeding fast-growing, sterile fish varieties and studying related mechanisms, which is of great significance for improving the economic benefits of aquaculture and protecting ecological security.

The invention relates to an optimization method and device of a fine arrangement model, a storage medium and electronic equipment. The method comprises the following steps: acquiring user behavior data, marking an object clicked by a target user in the user behavior data as a positive sample, and marking an object which is not clicked as a negative sample; obtaining a model prediction probability, an exposure frequency and a sample feature of each negative sample, and calculating a dynamic weight of each negative sample; inputting all browsing objects and object information thereof in the user behavior data into a fine arrangement model to perform recommendation probability prediction on all the browsing objects to obtain a prediction result, and calculating to obtain a basic loss function according to the prediction result and click information of all the browsing objects; and performing weighted calculation on the basic loss function according to the dynamic weight of each negative sample to obtain a weighted loss function, and iteratively updating parameters of the fine arrangement model according to the weighted loss function. The technical problems that an existing fine ranking model does not make full use of difficult-to-load samples, and the judgment and ranking precision is limited are solved.

The application relates to a negative pressure filter control system and method based on multi-scene application, belonging to the technical field of pool filtration, which comprises a detection module, the detection module is used for acquiring the number of swimmers and turbidity and uploading to a control module, the control module calculates a load coefficient f according to the number of swimmers n and the turbidity z, the control module increases the starting frequency of a water pump when the load coefficient f exceeds f0, wherein f=(n / n0xk1+z / z0xk2) / (k1+k2), k1 and k2 are weight numbers, and the control module increases the starting frequency of the water pump when the load coefficient exceeds a threshold value; the application has the characteristics of accurate data collection and high detection precision.

The present application belongs to the field of microseismic monitoring, and relates to a microseismic P wave and S wave separation method based on CNN-LSTM and a multi-head attention mechanism. The technology comprises: constructing a data set, using microseismic data collected in a hot dry rock hydraulic fracturing field, and manually labeling the data; the data set includes microseismic data of different time periods and different background noise intensities; inputting the data set into a neural network composed of a convolution layer, an excitation layer, a pooling layer, a random hidden neuron (Dropout) layer, a long short-term memory (LSTM) layer, a multi-head attention mechanism layer and a full connection layer for training; after the training is completed, the model is saved; inputting the microseismic data to be separated into the trained model for separation, and finally obtaining the separation result of P wave and S wave.

This application provides a method for identity verification and authentication of generating events of interest from video streams. The identity verification method includes: extracting and preprocessing frames from the input video stream to remove noise interference and enhance the image quality of keyframes; identifying objects of interest in the video stream using a target detection algorithm and extracting their biometric information; matching the extracted features with a preset identity feature database to determine the object's preliminary identity; verifying the preliminary identity by combining it with scene information of the event of interest, and outputting the identity verification result. The authentication method includes receiving the identity verification result and verifying the accuracy of the identity verification through multimodal feature cross-validation, time series consistency checks, etc., to ensure the reliability of identity recognition in events of interest. The method of this application can effectively improve the reliability of identity recognition for events of interest in video streams and is applicable to various scenarios such as sports videos and intelligent transportation.

The embodiment of the present application discloses a feeding method for improving the survival rate and weight of Polyphylla laticornis larvae in the process of processing and converting fungus dungs. The feeding method comprises the following steps: S1: obtaining the feces generated in the feeding process of 2nd instar and / or 3rd instar Polyphylla laticornis larvae; S2: obtaining the feces substrate formed by the feces; S3: putting the 2nd instar Polyphylla laticornis larvae into the feces substrate, and placing them in an environment with a temperature of 28±2 DEG C and a light cycle of L:D = 16h:8h for feeding for 10 days, and the substrate humidity is controlled at 50%-55%; S4: picking or screening the larvae fed for 10 days from the feces substrate, and putting them into fungus dung substrate for further feeding. The present application can improve the survival rate and weight of Polyphylla laticornis larvae in the process of processing and converting fungus dungs, improve the utilization rate of insect feces, simplify the pretreatment process of fungus dungs, and reduce the feeding cost. The feeding method is simple and feasible, convenient to obtain materials, and has a great promoting effect on the processing and converting fungus dungs industry of Polyphylla laticornis.

The utility model relates to finished product basket automatic weighing technical field, and disclose a kind of finished product basket automatic weighing warehouse entry structure, including support frame A, the one side of support frame A is fixedly connected with weighing platform, the one side of weighing platform is fixedly connected with support frame B, the surface of support frame A is provided with centering mechanism.The finished product basket automatic weighing warehouse entry structure, by the setting of centering mechanism, when using, goods is placed on the transmission belt A of support frame A surface, then starts transmission belt A and the hydraulic rod of both sides, and hydraulic rod drives centering plate and arranges neatly on transmission belt A, subsequent manual weighing is facilitated, and then transmission belt A is conveyed to weighing platform, and weighing operation is carried out, and after weighing, the article is placed on the transmission belt B of one side, and the article is conveyed to specified location using transmission belt B and is uniformly stored, and the whole process does not need manpower to carry, improves the automation level and overall efficiency of warehousing operation.

The application provides a method and system for automatically generating a topology of a power communication network, and relates to the technical field of intelligent operation and maintenance of a power system. The method comprises: collecting initial images containing at least one entity of a site, a device and a port at different time points; adjusting the weight of different entities in the initial images to obtain target images corresponding to the initial images, and arranging the target images corresponding to the initial images in the order of time points to obtain a time sequence image set; identifying the connection relationship between entities in any target image in the time sequence image set; and generating a dynamic topology according to the connection relationship between entities in each target image in the time sequence image set. The application adjusts the weight of different entities in the initial images to enhance the information expression ability of each entity and reduce the probability of missing detection. In addition, the dynamic topology is constructed according to the connection relationship between entities, which facilitates subsequent accurate maintenance of the power communication network.

The disclosure provides a recruitment post big data mining method and device, the method comprises: semantic normalization step: the original post name is processed by word segmentation, the word segmentation result is matched with the pre-constructed technical noun library and the function verb library, the core word item is extracted, and the normalized post name is generated according to the preset structured template reconstruction; skill extraction step: extracting skill keywords from the original post description corresponding to the normalized post name, generating a skill keyword set of the post category; feature vectorization step: calculating the term frequency-inverse document frequency of the core word item and the skill keyword set of the post category, generating a sparse feature vector of the post category; adaptive clustering step: based on the sparse feature vector of the post category, the clustering result is obtained.

This invention proposes a psychological diagnosis and treatment dialogue method based on dynamic knowledge evolution and adaptive retrieval, belonging to the fields of natural language processing and intelligent medical technology. The method includes constructing a psychological counseling generation model, an urgency prediction module, and a background process auxiliary model; initializing a patient knowledge base; sending the patient's current dialogue input to the background process auxiliary model to extract key information and determine whether retrieval is needed; if no retrieval is needed, the psychological counseling generation model generates a response, and the current dialogue ends; otherwise, the retrieval yields a set of candidate structured memory units, which are adaptively reordered to obtain context information; the context information is concatenated with the current dialogue input and sent to the psychological counseling generation model to generate a therapeutic response conforming to psychological counseling standards; and the knowledge base undergoes dynamic evolution and incremental updates. This invention overcomes the limitations of the traditional "static profile" of dialogue systems, achieving long-term continuity and dynamic adaptability of diagnostic and treatment memory.

The utility model relates to the technical field of river side slopes, in particular to a flood control device for a river side slope. Comprising a vertical part and an inclined part at the upper end of the vertical part, the upper end of the inclined part inclines towards the outer side of a river channel, the vertical part and the inclined part are integrally formed to form a slope protection part, the slope protection part is hollow, the upper end of the inclined part is open, and a plurality of through holes are formed in the side, facing the river channel, of the inclined part in a rectangular array. A supporting plate is fixed to the portion, below each row of through holes, of the inclined part, and a groove and a protruding edge are arranged at the two ends of the protection slope correspondingly and correspond to each other. And the vertical part and the inclined part are integrally formed, so that the problem of weak joints of a traditional segmented structure is solved, and the overall anti-scouring capacity is enhanced. The grooves and the convex edges of the adjacent protection slopes are quickly embedded to form a continuous and compact protection surface, so that the risks of water flow permeation and local damage are reduced, and the protection structure is particularly suitable for river channels with high silt content.

The application discloses an engineered exosome targeting human urokinase-type plasminogen activator receptor uPAR and a preparation method and application thereof. Firstly, a high-affinity anti-human uPAR single-domain antibody VHH is prepared; then, a human umbilical cord mesenchymal stem cell engineering cell line expressing a VHH-PTGFRN fusion protein is constructed, and the engineered exosome is prepared and purified by combining three-dimensional hypoxic culture, chemical induction and size-exclusion chromatography technology. The engineered exosome comprises four comparative samples, wherein the core sample Exo^(VHH / miR) can simultaneously display the anti-uPAR VHH on the membrane surface to realize targeted delivery and enrich miR-193b-3p in the chamber to realize gene silencing, and the synergistic effect of the two can efficiently remove the uPAR high-expression senescent cells and inhibit the senescence-associated secretory phenotype SASP. The engineered exosome has the advantages of standardized preparation process, high yield and high purity, strong targeting, excellent curative effect and good safety, and can be used for preparing a treatment preparation for anti-aging and aging-related diseases, and has important clinical application value.

This invention provides a deep multi-view clustering method based on multi-granularity contrastive learning, belonging to the field of computer vision technology. It solves the technical problem that single-granularity feature representations are insufficient to comprehensively capture the multi-level semantic information of data and the quality differences between different views. The method includes the following steps: S10, designing an independent encoder for each view to extract latent embedding features; S20, mapping the attention-enhanced features of each view to the same semantic space; S30, dynamically assigning weights based on the similarity between the features of each view and the global representation; S40, granular-level contrastive learning can enhance the compactness of the cluster structure within the granular sphere and align the granular sphere structure between cross-view granular spheres; S50, performing K-means clustering on the final global features to obtain the final prediction result. This invention enhances the weight of key features by adding a feature attention enhancement module to the view features.

This invention discloses an intelligent identification method for street-front shops based on an improved TF-IDF model, belonging to the field of intelligent shop identification technology. The method includes: extracting and normalizing internal customer information from the operator to obtain a local customer pool; performing a three-dimensional matching between a clean shop pool and the local customer pool to obtain a first pool of matched shops and a second pool of unmatched shops; forming text pairs using corresponding fields from the second shop pool and the local customer pool, calculating semantic similarity using an improved TF-IDF model, and identifying shops with similarity greater than a threshold as model-matched local shops; merging the first shop pool with the model-matched local shops to form the final local shop pool, and removing them from the clean shop pool to obtain a pool of shops from other networks; adding industry and product tags to the final local and other network shop pools to generate a shop profile library and outputting it to the marketing end. This invention can improve the accuracy of shop identification, reduce the cost of manual visits, and achieve one-time separation of local and other network shops for precise marketing.

The utility model discloses a long-distance high-speed transmission system, which relates to the technical field of hydraulic transmission, and comprises a main supply pump source, a discrete impact pump source, a transmission pipeline and a steady flow valve, the main supply pump source enables transmitted oil to flow from the starting section of the transmission pipeline to the outlet section of the transmission pipeline through pressure difference; the discrete impact pump source comprises a boosting device and a shock wave emitter; the boosted oil is stored in a high-pressure energy accumulator or conveyed into a shock wave emitter, pulse impact is conducted on the oil in the conveying pipeline through the shock wave emitter, a shock wave string formed by the pulse impact disperses and segments the oil into pulse flow sections, and a shock wave front face forces the oil at the front end to roll to the outlet section of the conveying pipeline at a high speed; pulsating impact oil at the outlet section of the transmission pipeline enters a buffer flow channel and an oil outlet flow channel in the steady flow valve through an oil inlet flow channel in the steady flow valve, and shock waves entering the buffer flow channel are buffered and stabilized by the shock wave buffer and then flow out of the oil outlet flow channel of the steady flow valve.

The invention relates to the technical field of tunnel construction, and provides a machine-rock coupling state detection method and system based on multivariate information perception in order to solve the problem that a dynamic contact state cannot be distinguished in real time at present. The machine-rock coupling state detection method based on multivariate information perception comprises the following steps: acquiring three-dimensional point cloud data of a tunnel face, and constructing a tunnel environment dynamic map and marking key features by combining internal structure information of a rock mass; based on a mechanical arm kinematic model, in combination with a tunnel environment dynamic map and mechanical arm tail end motion state information, a mechanical arm obstacle avoidance path and an optimal tunneling track are planned; the mechanical arm pose offset, the contact stress gradient and the rock mass fracture density in the operation process of the mechanical arm along the mechanical arm obstacle avoidance path and the optimal tunneling track are obtained, and after processing is conducted through the dynamic machine-rock contact recognition model, the classified contact state is obtained, and the rock mass contact trend is predicted; and when the predicted rock mass contact trend is abnormal, the early warning level is determined according to the current classification contact state, and a control instruction is generated, so that motion parameters of the mechanical arm are adjusted in real time, and an emergency protection mechanism is triggered.

This invention discloses a method, system, and storage medium for weighted and rate optimization of wireless networks assisted by intelligent reflectors, relating to the field of wireless communication, and addressing the problem of the lack of such optimization methods in existing technologies. The method includes: Step 1: Under transmit power constraints, establishing a weighted and rate optimization model for wireless networks assisted by intelligent reflectors based on the joint optimization of the transmit beamforming matrix at the base station and the diagonal matrix at the intelligent reflector; Step 2: Initializing a population of cultured Remora fish using the optimization model as the fitness function; Step 3: Allowing the cultured Remora fish to roam freely; Step 4: Allowing the cultured Remora fish to feed; Step 5: If the number of iterations is less than the preset maximum number of iterations, setting ε = ε + 1 and returning to Step 3; otherwise, terminating the iteration, outputting the global optimal position of the cultured Remora fish, returning the optimal fitness value, and finally obtaining the weighted and rate optimization method for wireless networks assisted by intelligent reflectors.

This invention discloses a method, apparatus, and device for evaluating the answer quality of a large language model, comprising: acquiring data to be evaluated; extracting the input question, standard answer, and model answer generated by the large language model to be evaluated for the input question; vectorizing the standard answer and model answer using a vector model to obtain standard sentence vectors and model sentence vectors; identifying semantic units in the standard answer and model answer to obtain a set of standard units and a set of model units; calculating the importance weight of each semantic unit in the set of standard units and the set of model units based on information entropy theory; performing semantic similarity matching between the semantic units in the set of standard units and the semantic units in the set of model units; determining the matched units covered in the set of standard units based on the matching results; and calculating multi-dimensional evaluation indicators based on the standard sentence vector, model sentence vector, importance weight, and matched units, and performing weighted fusion to obtain a comprehensive score.

This invention provides an adaptive fusion retrieval enhancement method for EDA technical document question answering, belonging to the field of electronic design automation. The method includes: acquiring a document set; constructing a semantic unit set by performing structured semantic parsing on the documents and query text, and obtaining confidence information for keywords and key sentences; calculating document-level relevance scores using term matching relationships and performing preliminary screening to obtain a candidate document set; calculating a comprehensive relevance score based on the semantic unit set by introducing a multi-granularity semantic alignment mechanism at the keyword, sentence, and document levels, and ranking the candidate document set to obtain enhanced retrieval results; and inputting the enhanced retrieval results into a preset generation model to complete the adaptive fusion retrieval enhancement. This invention solves the problems commonly found in existing retrieval enhancement generation systems, such as single semantic matching granularity, coarse relevance modeling, and insufficient utilization of key information within documents.

The application belongs to the field of multi-target tracking. An anti-occlusion target tracking method fusing multi-granularity dynamic appearance is provided. The purpose is to solve the technical problems of target ID jump and tracking effect affected in the occlusion scene in the prior art. The main scheme includes obtaining an original image. A target detector is called on the original image to obtain a target detection result, and a target appearance feature is extracted from the target image in the target detection result; a target appearance contribution factor alpha of the video frame is calculated; based on the target detection result of the previous frame, a target motion feature of the current frame is obtained. The target appearance feature, the target appearance contribution factor alpha and the target motion feature are data fused to obtain a target similarity. The targets in two frames are cascade matched and IOU matched, and each target is assigned an ID. The positions of the targets associated with the same ID in each frame are obtained, the running track of the same target ID in the image sequence is obtained, and the target tracking result is output.

The application provides a method and system for automatically generating a topology of a power communication network, and relates to the technical field of intelligent operation and maintenance of a power system. The method comprises: collecting initial images containing at least one entity of a site, a device and a port at different time points; adjusting the weight of different entities in the initial images to obtain target images corresponding to the initial images, and arranging the target images corresponding to the initial images in the order of time points to obtain a time sequence image set; identifying the connection relationship between entities in any target image in the time sequence image set; and generating a dynamic topology according to the connection relationship between entities in each target image in the time sequence image set. The application adjusts the weight of different entities in the initial images to enhance the information expression ability of each entity and reduce the probability of missing detection. In addition, the dynamic topology is constructed according to the connection relationship between entities, which facilitates subsequent accurate maintenance of the power communication network.