345 results about "Track algorithm" patented technology

The invention relates to the technical field of target tracking, in particular to a photoelectric tracking algorithm and system for self-adaptive target tracking, and the system comprises an image collection module, a cross-spectrum self-adaptive sensing module, an image processing and target recognition module, a self-adaptive tracking algorithm module and a servo control module, and constructs a closed-loop feedback link to achieve the synchronous optimization of parameters in a whole link. The algorithm comprises the steps of image acquisition and adjustment, target detection, parameter optimization, servo driving and feedback. The system can accurately estimate a high-speed turning target, is adaptive to complex environments such as low illumination / haze and the like, completes sheltered target recapture within 0.5 s, has servo compensation precision of + / -0.1 degree, is suitable for airport bird monitoring, port unmanned aerial vehicle tracking and highway vehicle monitoring scenes, and significantly improves tracking stability and practical value.

The invention provides a three-dimensional tracking method applied to a single-photon laser radar to detect a far-field dynamic unmanned aerial vehicle point target, and belongs to the technical field of laser radar detection and target tracking. The problems that in the prior art, under the long-distance condition, the target imaging size is smaller than one pixel, and a detection and tracking algorithm based on shape, texture or edge features fails are solved. The method comprises the following steps: constructing a six-dimensional state vector containing a three-dimensional position and a three-dimensional speed of a target; predicting a state vector according to a state transfer function containing air resistance and turning acceleration; and an extended Kalman filtering framework is utilized to update the predicted state and obtain state estimation of the target, and the extended Kalman filtering framework comprises a self-adaptive process noise adjustment mechanism based on innovation feedback and is used for adjusting a process noise covariance matrix on line. The method is mainly used in the field of detection and multi-dimensional tracking of airspace moving targets.

The invention provides a financial statement automatic generation method and system based on machine learning, and the method comprises the steps: carrying out the matching and conversion of semantics of different fields through employing an intelligent recognition technology according to a preliminary data format difference classification result, generating a standardized data structure set, and obtaining a unified field mapping relation; performing dynamic recombination on the cross-system business process by adopting a process tracking algorithm through a complete business data link, obtaining a time sequence relation and a data flow path of each business link, and obtaining a complete process tracking graph; key financial indexes and associated data of all links are obtained through the repaired process data set, and a data integration module is adopted for multi-dimensional summarization to obtain a preliminary financial data report framework; and through the final financial data content, obtaining a preset report template and an output rule, and generating a financial report output result conforming to a specification.

The invention relates to the technical field of computer vision, in particular to a target tracking method based on adaptive template updating and lightweight Transform. The invention provides an efficient and robust target tracking algorithm for solving the problems that a traditional twin network is insufficient in robustness in a complex scene and a Transform architecture is highly dependent on computing resources. Firstly, a supervision feedback module is designed, supervision information related to a task is introduced in a feature extraction stage, and a network is guided to be more focused on a target area, so that the feature discrimination capability is improved, and background interference is effectively suppressed; secondly, a lightweight Transform structure is constructed, the calculation complexity and the parameter scale are remarkably reduced while the global modeling capability is maintained, and the balance between the model performance and the calculation efficiency is achieved; and finally, designing a self-adaptive template updating mechanism, and dynamically updating the template content in combination with the state information of the current frame and the environment change, thereby enhancing the adaptability of the model to the target appearance change and reducing the tracking drift risk.

The invention discloses an unmanned aerial vehicle autonomous information perception path optimization method fusing sparse Gaussian estimation and RLSAC, and particularly relates to the technical field of unmanned system multi-target tracking, and the method comprises the steps: employing Kalman filtering to carry out the optimization processing of target initial data obtained by a ground laser radar, achieving the filtering of the noise of a target, obtaining the dynamic trajectory of the target, and obtaining the optimal path of the target. Therefore, the output insufficiency of the RLSAC to the dynamic trajectory is made up. The data after Kalman filtering processing is used as original input data of a multi-unmanned aerial vehicle target tracking I PP algorithm based on information path planning, the own unmanned aerial vehicle adopts the I PP algorithm to track a target in real time and transmit target related data to provide input data for RLSAC, the RLSAC is used to process the input data, model fitting is carried out, and the RLSAC is used to carry out target tracking. And the RLSAC outputs rewards of different geometric morphology models adopted by the own unmanned aerial vehicle during mode recognition, and the model which is more consistent with the output of the RLSAC is determined by calculating the confidence coefficient.

The invention discloses a dynamic visual guidance method and system for cooperative assembly of multiple mechanical arms, and belongs to the technical field of industrial robot visual guidance, and the method comprises the steps: synchronously collecting an image sequence through a binocular visual system, synchronously recording the pose sequence of each mechanical arm through a robot controller, and obtaining an observation data set; obtaining a two-dimensional trajectory flow of vehicle body assembly area feature points and instrument body feature points through an optical flow tracking algorithm, and obtaining a hand-eye matrix correction through a nonlinear optimization algorithm; obtaining an updated hand-eye matrix and a local pixel compensation field through a spatial interpolation algorithm based on the re-projection error of each feature point; according to the method, self-adaptive compensation of flexible deformation of a vehicle body and dynamic reflection of an instrument is realized by adopting unsupervised clustering and spatial interpolation according to a re-projection error, and the precision of cooperative assembly of multiple mechanical arms and the self-adaptive capability of a system in a complex industrial environment are improved.

The invention relates to the technical field of gas leakage detection and optical remote sensing, particularly discloses a laser natural gas leakage point positioning and imaging processing method based on an unmanned aerial vehicle, and aims to solve the problems of multiple coverage blind areas, low positioning precision and easiness in environmental interference in the prior art. The method comprises the following steps: acquiring a methane concentration integral value through a tunable diode laser absorption spectrum system carried by an unmanned aerial vehicle; constructing an initial concentration point cloud in combination with the high-precision pose and scanning angle data; performing space-time filtering and interpolation enhancement by adopting Kalman filtering and a Gaussian plume model to generate a high-confidence three-dimensional concentration field; and accurately positioning a leakage source by using a gradient rising tracking algorithm in combination with curvature and continuity constraints, and matching pipe network geographic information to perform verification alarm. According to the technical scheme, large-range, high-precision and full-automatic natural gas leakage detection and visual positioning are realized, and the signal-to-noise ratio, the anti-interference capability and the response efficiency are remarkably improved.

The invention discloses an intelligent alarm system based on a real-time video stream of a platform, and relates to the technical field of video processing early warning, and the intelligent alarm system comprises a data collection module which carries out the real-time monitoring based on a monitoring camera, reads the image data of a point position camera, carries out the standardization and normalization of the data, and carries out the unified input of the image data; and the target discovery module is used for inputting the preprocessed image frame and the monitored area, identifying and positioning pedestrians in the image frame by adopting an algorithm based on target detection, performing label processing on the identified pedestrians based on a detected tracking algorithm byterack, and outputting the residence time of the current frame target in the monitored area. According to the platform safety management system, various advantages are brought to platform safety management, the alarm threshold value can be flexibly adjusted, decision support is provided, management personnel are helped to analyze passenger flow, optimize deployment and troubleshoot hidden dangers by recording related data, a'early warning-processing-optimization 'closed loop is formed, the safety management and control level is improved, and safety accidents are effectively prevented.

The invention relates to the technical field of automatic guided vehicle control, in particular to a multi-differential wheel set AGV trajectory tracking method and system, and the method comprises the steps: building a special kinematics forward / inverse solution model, and determining the speed and angle mapping relation between a driving wheel and the center of a vehicle body; constructing a double-layer framework of upper-layer pure tracking trajectory tracking and lower-layer wheel set cooperative control, and introducing a steering cooperative optimization strategy; an intelligent self-correction mechanism is set, and a strong correction mode or a smooth correction mode is selected according to the deviation distance; and deep fusion of a pure tracking algorithm and a special chassis is realized, and a complete closed-loop control system is formed. The system comprises a chassis module, a kinematics modeling module, a double-layer control module, a self-correction module and a closed-loop feedback module. According to the method, the problems of poor adaptability, motion coupling, unbalanced heavy AGV tracking precision and stability and the like of a traditional algorithm are solved, high-precision trajectory tracking is realized, tire wear and energy loss are reduced, the robustness of the AGV in a complex environment is enhanced, and the method is suitable for heavy and heavy-load AGVs.

The invention discloses an electric field inspection unmanned aerial vehicle path planning method, and belongs to the technical field of electric field inspection, and the method comprises the following steps: collecting multi-dimensional environment data and task demand data for an electric field inspection scene, forming a path planning basic data set, carrying out the preprocessing and modeling of the basic data set, and obtaining a path planning result; the method comprises the following steps: constructing an electric field inspection environment grid model and quantifying a path planning constraint rule, designing a dynamic path planning algorithm adaptive to an electric field scene based on the grid model and the constraint rule, generating an initial optimal path, and converting the dynamic path planning algorithm into an operable path planning system. And a data processing, path calculation, real-time adjustment and result output function module is developed. According to the electric field inspection unmanned aerial vehicle path planning method, a power side-load side core inspection path is locked through a power flow tracking algorithm, searching is only carried out in the core path and two adjacent grids, non-key branch lines are excluded, and compared with traditional full-scene searching, the calculation space is reduced by 60% or above.

The invention discloses a digital twin mechanical arm dynamic obstacle avoidance method based on multi-view fusion and intelligent decision, and belongs to the technical field of industrial robot control, machine vision and artificial intelligence. Comprising the steps that environment data are collected in real time, a mechanical arm digital twinborn model is constructed, a target detection and tracking algorithm is used for recognizing and tracking a dynamic obstacle, a virtual model and a motion track of the dynamic obstacle are reconstructed in a three-dimensional mode through multi-view information fusion, and construction of a digital twinborn scene is completed. Performing global path planning by adopting an improved A * algorithm which introduces a direction change penalty factor and a path smoothing optimization strategy; and performing local real-time planning by using a local planner integrated with a pre-trained DQN decision optimization module, issuing and executing the optimal local trajectory, and performing real-time feedback to form a closed loop. According to the method, the problems that dynamic obstacle avoidance is not flexible and is lack of predictive ability are solved, and the autonomy, safety and efficiency of the mechanical arm in a complex dynamic environment are remarkably improved.

The invention relates to the technical field of diaphragm pump aging test, and discloses a two-parameter intelligent closed-loop diaphragm pump aging test system and method. According to the method, a temperature spectrum and a vibration spectrum are obtained through fast Fourier transform, a frequency domain coherence coefficient is calculated to identify temperature-vibration coupling characteristics, a resonance frequency band and a drift trend thereof are identified in real time by adopting a spectrum peak tracking algorithm, and a multi-objective optimization function considering aging acceleration efficiency and resonance risk is constructed. And solving the optimal temperature stress amplitude and vibration stress frequency by using a particle swarm algorithm, generating a stress loading instruction avoiding a resonance area through a self-adaptive PID controller, and dynamically updating model parameters based on a closed-loop feedback mechanism. According to the invention, the problem of resonance effect amplification caused by traditional fixed frequency stress loading is solved, and refined adaptive stress control is realized.

The invention discloses a multi-target sensing and tracking method for detecting an unmanned aerial vehicle swarm through a single-photon laser radar, and belongs to the field of photoelectric detection and signal processing. The problem that a traditional image reconstruction algorithm and a tracking algorithm are difficult to adapt to ultra-low signal-to-noise ratio, high dynamic and multi-target complex sensing requirements is solved. The method comprises the following steps: carrying out systematic prepeak noise self-adaptive suppression on global photon arrival time data acquired by a single-photon laser radar; performing photon event physical statistical modeling and background noise suppression based on the suppressed data to obtain an enhanced residual signal; performing space-time consistency enhancement processing on the enhanced residual signal to realize extraction and reconstruction of a multi-target area; based on the extracted photon events in the multi-target area, performing trajectory fitting and prediction by adopting a Bayesian weighted RANSAC (Random Sample Consensus) fitting method; and designing a target tracking algorithm based on scale adaptive updating and template matching of an optical flow method and pyramid search assistance, and carrying out robust tracking on multiple targets. The method is used in target tracking field.

The invention relates to the technical field of image target detection, and discloses an inland ship intelligent inspection method based on deep learning, and the method comprises the steps: collecting video image frame data containing a ship target in an inland waterway region, and constructing a ship detection data set and a ship compliance detection data set; on the basis of the YOLOv5s, a shallow high-resolution characteristic path is introduced, and an improved YOLOv5s model is constructed; respectively training a ship detection model and a compliance detection model based on the improved YOLOv5s by utilizing the ship detection data set and the ship compliance detection data set, and storing optimal model parameters; the ship detection model processes a real-time video, outputs ship position information and transmits the ship position information to a ByteTrack tracking algorithm, when the PTZ PTZ camera is regulated and controlled to meet a preset condition, the camera is controlled to shoot a static image and put the static image into the compliance detection model for detection, and a detection result is output; according to the method, the detection capability and the detection efficiency of small targets in ship compliance detection can be improved.

The invention provides a low-altitude visual analysis and illegal behavior recognition method for a port, and the method comprises the steps: collecting and carrying out the time-space alignment of a video stream through a multi-view high-definition camera, and extracting a standardized behavior event sequence through a deep neural network and a tracking algorithm; the events are sequenced according to time, behavior frequency, transition probability and co-occurrence modes are counted in a sliding time window, a dynamic feature vector is generated to drive an incremental graph neural network, and continuous embedding updating, confidence calculation and anomaly detection are performed on a behavior semantic graph; if structural abnormity is detected, a new behavior mode is identified through clustering analysis and causal reasoning assistance, and map self-adaptive adjustment is achieved after expert feedback verification; and finally, a composite semantic map is constructed by fusing conventional and abnormal behavior reasoning results through a double-channel learning architecture, port violation behaviors are accurately identified and pre-warned, and the scene understanding ability and the safety monitoring intelligent level are improved.

The invention relates to the technical field of agricultural equipment control, in particular to an intelligent fusion control navigation system of a vegetable collection and transportation all-in-one machine, which comprises an image acquisition and preprocessing module, a cabbage ridge line identification and positioning module, a multi-source information fusion decision module and a path tracking and cooperative control module. The method comprises the following steps: collecting ridge row images through a visual sensor, executing multi-scale fusion denoising, histogram equalization and perspective correction, and extracting a clear visual navigation datum line; gNSS and IMU data are fused with visual information, and extended Kalman filtering is adopted to realize high-precision pose estimation; switching different operation modes based on the operation state machine, and dynamically generating a corresponding expected path; path tracking control is completed in combination with a pure tracking algorithm and a speed mapping table, and intelligent cooperative control over the whole receiving and operating process is achieved. According to the method, the ridge row identification robustness, the path tracking precision and the system collaborative operation capability are improved, and the method is suitable for automatic harvesting operation of cabbages in a complex farmland environment.

The invention relates to a YOLO algorithm-based automatic detection method and system for pedestrians who do not give way to vehicles. The method comprises the following steps: collecting a road monitoring video stream and converting the road monitoring video stream into an image sequence for preprocessing; inputting the preprocessed image sequence into an improved YOLO target detection model to identify vehicles and pedestrians; utilizing a ByteTrack tracking algorithm to generate continuous motion trails of the vehicle and the pedestrian; analyzing and detecting whether a vehicle which does not give way to pedestrians exists based on the motion trail, and obtaining a detection result; and when the detection result is that the vehicle does not give way to the pedestrian, a detection image is intercepted and output. A vehicle and a pedestrian are accurately recognized through an improved YOLO target detection model, a ByteTrack algorithm is used for matching and tracking a target in a cross-frame trajectory mode, the trajectory is analyzed in real time to judge violation, an image is automatically intercepted when the violation is detected, and information is marked and output, so that full-process automatic processing from traffic data collection to violation evidence output can be realized, and the traffic violation evidence output efficiency is improved. Therefore, accurate detection and continuous tracking of vehicles and pedestrians are realized.

The invention discloses an unmanned aerial vehicle autonomous approaching and netting control system based on visual servo, and belongs to the technical field of unmanned aerial vehicle control. The high-precision two-axis pan-tilt mechanism, the binocular vision module, the netting launching device and the electromechanical system main control unit of a hierarchical control algorithm are integrated, automatic recognition, accurate distance measurement, stable tracking and intelligent striking of a dynamic target are achieved, and the two-axis pan-tilt mechanism enables the netting launching device to have the motion freedom degree independent of an unmanned aerial vehicle platform. The effective striking range of the system is effectively expanded by combining a high-precision target positioning and servo tracking algorithm based on binocular vision, the capturing precision and the adaptive capacity of a dynamic target are remarkably improved, and through fusion of deep learning target detection, binocular vision pose calculation and a hierarchical control strategy, the dynamic target capturing precision and the adaptive capacity of the system are improved. The full-process independence from target identification, tracking approximation to attack decision making is realized, and a multi-subsystem efficient and collaborative working architecture is formed through unified scheduling of the main control unit.

The invention relates to the technical field of photovoltaic cell MPPT, and discloses an intelligent tracking method, system and device for the maximum power point of a solar cell panel and a medium, and the method comprises the steps: obtaining a target data sample, and carrying out the analog-to-digital conversion of the target data sample; establishing a first maximum power point tracking algorithm, wherein an analog-to-digital conversion result is input data of the first maximum power point tracking algorithm; and performing intelligent tracking of the maximum power point of the solar cell panel according to the output of the first maximum power point tracking algorithm. The output voltage and output current analog signals of the solar cell panel can be accurately obtained in real time, the analog signals are converted into digital signals through analog-to-digital conversion, and an accurate data basis is provided for subsequent algorithm processing. Through the established first maximum power point tracking algorithm, the maximum power point of the solar cell panel can be quickly and intelligently determined and tracked, so that the power generation efficiency and stability of the solar cell panel are improved.

The invention discloses a similar wild protected animal interference data processing method and system, and belongs to the field of computer vision, and the method comprises the steps: obtaining one or more wild protected animal monitoring reference data sets, and constructing a data space D; m single-target tracking algorithms are selected to construct a determiner set A; performing similar object interference feature extraction and judgment on each video sequence in the data space D to obtain a frame-level judgment result sequence and sequence-level interference statistical data; and performing grading processing according to the sequence-level interference statistical data to obtain a structured data set containing similar object interference characteristics. Aiming at the problem of tracking failure caused by similar object interference in a wild animal tracking technology in the prior art, the invention provides an interference identification mechanism based on multi-determiner cooperation, and automatic detection, quantitative analysis and grading processing of similar object interference in a video sequence are realized.

The invention relates to a target positioning method based on radar photoelectric linkage guidance, and belongs to the technical field of edge-sea-air defense three-dimensional management and control of multi-source sensor fusion. According to the method, a dynamic calibration mechanism is provided, a real-time timestamp alignment and navigational speed calibration algorithm is introduced on the basis of traditional calibration, and radar measurement errors and coordinate conversion errors are effectively weakened; through an SAMURAI tracking algorithm of improved Kalman filtering, stable target tracking under shielding, appearance change and similar target interference is realized, and identity and position consistency is kept; the three-dimensional-to-two-dimensional projection conversion is adopted, the calculation precision of the pitch angle and the azimuth angle is improved, and the influence of radar latitude and longitude errors is reduced. Compared with an existing method, the method has the advantages that the precision and stability of joint positioning of the perimeter target in a complex environment are remarkably improved, the problem of target tracking jump is effectively solved, and the method has wide application prospects in scenes such as boundary security and protection and marine monitoring.

The invention discloses a alumen ustum characteristic quantification method based on a target detection and tracking algorithm, and belongs to the field of computer vision, and the method specifically comprises the steps: 1, collecting a alumen ustum data set; step 2, an alumen ustum detection and tracking algorithm based on improved YOLOv11-DeepSORT is carried out; comprising an improved YOLOv11 target detection algorithm and an improved YOLOv11-DeepSORT target tracking algorithm. The intelligent addition of the water treatment coagulant provides a quantitative technical basis, is beneficial to realizing automatic and fine control of the coagulation process, and improves the water quality safety and the treatment efficiency. In future work, the robustness of the algorithm is further optimized in a more complex actual underwater environment, and how to establish a more direct quantitative relationship between the alumen ustum form and motion parameters and the coagulant dosage is explored, so that more intelligent and more accurate coagulation process regulation and control are realized.

The invention provides a traffic collision early warning method, system and device based on multi-source perception and a medium, and belongs to the technical field of intelligent traffic. The method comprises the following steps: deploying multiple types of sensor arrays and edge computing terminals in a traffic accident high-incidence area, configuring V2X communication and traffic signal interfaces, and constructing an operable hardware infrastructure; a sensor array is used for collecting target data, and standardized multi-source data streams are generated through time synchronization, format unification and noise filtering; carrying out fusion perception and target identification by adopting a multi-source attention mechanism fusion algorithm, an OfficientDet model and a tracking algorithm, and outputting a dynamic target tracking data set; predicting a target trajectory based on a time sequence prediction model, calculating a collision probability and dividing collision risk levels in combination with space-time crossing judgment, and generating a risk assessment result and a distribution diagram; triggering a multi-channel early warning mechanism according to the collision risk level; and recording full-process data and labeling, and carrying out model optimization regularly.

The invention discloses a method and system for preventing screen information leakage based on UWB positioning, relates to the field of smart screens, and aims to capture facial features, eyeball movement and other information of an observer carrying shooting equipment in front of a screen in real time by combining a UWB positioning technology and a high-speed camera, and identify whether the observer is a legal user or not through a deep learning algorithm. For an unauthorized observer, the system can generate false information which is similar to the screen content but has no practical significance in real time to replace the display content of the sensitive area. According to the dynamic replacement technology, illegal observers cannot obtain effective information without affecting legal user experience, further, an attention tracking algorithm is introduced, the sight focus of the observers is predicted, only areas which are possibly concerned are protected, and therefore the calculation load is reduced.

The invention relates to the technical field of satellite communication, and discloses a low-orbit satellite communication Doppler frequency offset processing method, system and device and a medium, and the method comprises the steps: separating a pilot signal from a received signal; taking the Doppler frequency as a first system state, taking the received signal as first observation data, and performing Doppler frequency offset estimation on the pilot signal by adopting a frequency offset estimation algorithm to obtain a frequency offset estimation value; taking the Doppler frequency as a second system state, taking the frequency offset estimation value as second observation data, and adopting a frequency offset tracking algorithm to perform Doppler frequency offset tracking on the pilot signal to obtain a frequency offset tracking value; and performing frequency offset compensation on the received signal according to the frequency offset tracking value to obtain the received signal after frequency offset compensation. According to the method, the Doppler frequency offset is subjected to streaming processing through Bayesian filtering in the estimation and tracking stages, the complexity of data storage and calculation is effectively reduced, the data updating rate is increased, and therefore the accuracy of low-orbit satellite data is effectively improved.

The invention discloses a visual anti-occlusion tracking method based on AIS and video fusion. The method comprises the following steps: synchronously obtaining AIS data and corresponding video data; preprocessing the AIS data, predicting latitude and longitude coordinates of the AIS data, projecting the latitude and longitude coordinates to a video pixel plane, and generating a ship AIS track; meanwhile, identifying a ship bounding box in a video frame by using a video detection algorithm, and tracking by using a three-stage association anti-occlusion tracking algorithm; the similarity between the AIS trajectory and the video trajectory is calculated through a dynamic time warping algorithm, and the correlation matching of the AIS trajectory and the video trajectory is realized by using a Hungary matching algorithm, so that a stable and anti-shielding ship tracking trajectory fusing AIS and video information is obtained. According to the invention, the problem that pure visual tracking is easy to fail under the shielding condition is effectively solved.

This invention discloses a target tracking focusing method and apparatus, comprising: acquiring infrared image data of the target being tracked using an infrared optical system; determining a sharpness parameter of a sub-image region related to the target in the infrared image data; determining focusing parameters of the infrared optical system based on the sharpness parameter of the sub-image region; focusing the infrared optical system using the focusing parameters; and ending the focusing process when the desired sharpness is achieved. This embodiment utilizes the sharpness parameter of a sub-image region related to the target in the infrared image data of the tracked target for focusing. This allows for improvements in target detection and tracking distance, distance limits, and target tracking stability without altering the target tracking algorithm.

The invention discloses a federated learning method and system with a communication efficient mechanism in a heterogeneous environment, and the method comprises the steps: carrying out each round of training by a client through employing a gradient tracking algorithm with a momentum mechanism, obtaining the variable quantity of a trained local model matrix parameter, compressing the variable quantity through employing an error feedback mechanism, and uploading the compressed variable quantity to a central server; the central server calculates the average value of the compressed variable quantity uploaded by the participating clients and sends the average value to all the clients, and the clients and the central server calculate a new round of global model based on the average value; and repeating until the global model converges, and completing the training of federal learning. According to the invention, the calculation amount of the client in federal learning can be reduced, and the communication traffic between the client and the central server is reduced, so that the algorithm convergence time is reduced, and the federal learning efficiency is improved.

The present application belongs to the technical field of vehicle-mounted communication antenna servo control, and particularly relates to a tilt angle sensor drift elimination method based on a bias tracking algorithm, comprising the following steps: synchronously collecting original tilt angle readings and chip temperature readings of the tilt angle sensor at a fixed collection period; calculating a short-time standard deviation of the original tilt angle readings in a sliding window, determining as stationary when the short-time standard deviation is lower than a stationary threshold, determining as disturbance when the short-time standard deviation is not lower than the stationary threshold, obtaining a stationary state flag, and taking the square of the short-time standard deviation as an observation noise variance; calculating a temperature rise rate according to the chip temperature readings, and outputting a high value as a bias process variance when the stationary state flag is stationary and the temperature rise rate is higher than a thermal equilibrium threshold. The present application solves the problem of transient bias drift of the tilt angle sensor caused by chip temperature rise, so that the output error of the corrected tilt angle value continuously meets the leveling accuracy requirement in the task period.

The invention is suitable for the field of computer vision and artificial intelligence, and provides an automatic target detection and tracking method and system based on deep learning, and the method comprises the steps: obtaining a video stream and a scene structure diagram, extracting a target position bounding box, a category label and an appearance feature vector frame by frame, and calculating a behavior parameter and a priority score; generating a detection frame interval and a tracking algorithm complexity parameter according to the priority score, and initializing a special tracker to realize automatic tracking; when the target is lost, generating a space-time probability graph based on historical data and the scene structure graph; determining retrieval parameters according to the priority scores, and performing directional detection on the screened space-time regions; and after the target is retrieved, the identity is associated and parameters are redistributed to recover tracking. According to the method, through differentiated resource allocation, accurate loss prediction and directional retrieval, tracking precision, efficiency and robustness are considered, and the problems of target tracking interruption, resource waste and the like in a complex scene are effectively solved.