8381 results about "Image sequence" patented technology

The present disclosure provides an underwater detection method for contact leakage of a tunnel joint of a dam culvert, including: obtaining an underwater image sequence of the culvert by an underwater robot; preprocessing and registering the underwater image sequence; extracting particle information appearing in the underwater image sequence based on the registered underwater image sequence; constructing a three-dimensional fluid velocity distribution map based on the particle information, and determining leakage situation according to the three-dimensional fluid velocity distribution map; and superimposing the three-dimensional fluid velocity distribution map on a preconfigured three-dimensional culvert model, and rendering and displaying it. The present disclosure can adapt to the complex and changeable underwater environment inside the culvert, it has a high sensitivity to fine leakage in the deep layer, it can realize the accurate measurement and trend prediction of leakage, and it can automate the detection process to reduce manual dependence.

The invention discloses an unmanned aerial vehicle routing inspection line adaptive obstacle detection method and system based on a monocular camera. The method comprises the steps of obtaining initial flight parameters, evaluating and optimizing the flight parameters based on scene complexity, collecting an original image sequence and carrying out quality evaluation and exposure compensation. Constructing a multi-resolution feature pyramid, extracting feature descriptors, and performing feature matching and verification; performing multi-constraint pose optimization and depth estimation based on a feature matching result, and reconstructing a three-dimensional point cloud; calculating direction enhancement response to extract a power line candidate area, performing multi-view line segment matching and three-dimensional reconstruction, and analyzing dynamic characteristics of the power line; and based on the power line three-dimensional model and the depth map, carrying out obstacle segmentation, calculating an obstacle distance, evaluating a risk level, and generating early warning information and a monitoring scheme. According to the invention, different scenes can be adaptively processed, the obstacle risk can be accurately identified and evaluated, and the inspection efficiency and reliability are improved.

The invention belongs to the technical field of visual special effects, and discloses an AI visual special effect dynamic generation system fusing multi-modal perception. Through cooperative work of six core modules of multi-modal input processing, importance analysis, parameter configuration, parameterized rendering, parameter optimization and rendering and output, deep fusion of special effects and contents is realized. An image sequence, audio data and scene parameters can be processed at the same time, a multi-modal feature data set is constructed, a scene key visual area is scientifically recognized, a visual importance distribution map is generated, and the special effect parameter configuration and rendering process is guided. An improved neural radiation field technology and a physical constraint model are adopted to ensure the consistency and reality of the special effect at different visual angles; and a multi-dimensional quality evaluation and parameter automatic optimization mechanism is introduced, so that the visual expressive force and the artistic value of the special effect are guaranteed. The creation threshold is remarkably reduced, the cooperative expression ability of the special effect and the content is improved, and the special effect becomes a powerful tool for enhancing the narration and enhancing the emotion.

The invention discloses an unmanned aerial vehicle autonomous routing inspection conductor detection method and system based on Bezier curve modeling. The method comprises the steps that a binocular image sequence, GPS positioning and IMU attitude data are collected through an unmanned aerial vehicle, synchronous alignment processing is carried out on multi-source data, a scene dynamic evaluation value is calculated, adaptive image enhancement is executed, and space-time features are extracted and fused. Constructing a multi-scale feature pyramid, extracting wire topological features, and generating enhanced features in combination with environment context information; and generating a control point candidate set based on the enhanced features, optimizing Bezier curve parameters, and performing time sequence consistency optimization. According to the invention, high-precision lead detection and robust tracking in a complex scene are realized, and the reliability of autonomous inspection of the unmanned aerial vehicle is improved; the key technical problems of scale change, feature matching, dynamic modeling, data fusion, parameter optimization and the like in the wire detection process can be solved.

The invention provides a reconstruction method of a three-dimensional reconstruction model based on two-dimensional Gaussian splashing, which comprises the following steps: S1, carrying out sparse reconstruction on an input image sequence through a multi-view stereoscopic vision algorithm to generate an initial sparse three-dimensional point cloud and a corresponding camera pose parameter; s2, inputting an improved two-dimensional Gaussian radiation field by using the sparse three-dimensional point cloud and the camera pose as information; s3, dynamically screening a visible anchor point subset based on the current view angle parameter, and generating a rendered image through a differentiable rendering pipeline; s4, calculating a loss function of the rendering image of the training track and the input image to optimize a reconstruction scene; and S5, starting a special visualization tool, and inputting a rendering result. According to the method, by introducing a trimmable anchor point parameterization framework and a multi-scale feature fusion mechanism, light-weight and high-precision three-dimensional scene modeling is achieved, and the problems that traditional 2D Gaussian sputtering is insufficient in multi-view geometric consistency, storage overhead and weak texture region reconstruction and an existing 2D Gaussian splashing method is insufficient in self-adaptive mechanism are solved.

The invention provides an early warning analysis method based on intelligent vision and a server, and the method comprises the steps: obtaining a multi-mode video stream containing a visible light and infrared thermal imaging image sequence through a vision collection device in a monitoring region, and carrying out the spatial-temporal feature extraction of the multi-mode video stream; the method comprises the steps of generating dynamic behavior sequence characteristics including a target object moving track, posture change and thermodynamic distribution characteristics, analyzing the dynamic behavior sequence characteristics through a preset abnormal behavior judgment model in a multi-dimensional matching mode, judging whether the target object triggers abnormal behavior characteristics or not, and if yes, generating a graded early warning signal in combination with an association weight. In addition, according to the spatial distribution and the time length of the abnormal behavior characteristics, the parameters of the acquisition equipment and the coverage range of the monitoring area are dynamically adjusted, and the subsequent video stream anomaly detection precision is optimized, so that intelligent and accurate early warning analysis is realized, and the security efficiency is improved.

The invention relates to the field of defect detection, in particular to a PCB defect detection method and system based on image recognition. The method comprises the following steps: collecting a multidirectional PCB detection image, carrying out pixel-level registration correction and adaptive pixel stability compensation, and constructing a space-time stability compensation image sequence; performing reverse pyramid structure division on the space-time stability compensation image sequence, performing normalized similarity probability calculation, and constructing an initial region classification result; based on an initial region classification result, depth image visual analysis is carried out, pseudo defect comprehensive elimination optimization is carried out, and a pseudo defect purification high-confidence image is constructed; pCB connection defect identification is carried out on the pseudo defect purification high-confidence image, global defect point distribution marking is carried out, and a defect point space distribution diagram is constructed. According to the invention, high-credibility, high-precision and high-closed-loop PCB defect detection is realized.

The invention relates to the technical field of unmanned vehicle control, and discloses a smart city environmental sanitation unmanned vehicle path planning and real-time monitoring method. The method comprises the following steps: firstly, acquiring a multi-source environment sensing data set such as laser radar point cloud data, a camera image sequence and real-time traffic flow information; local road network features are extracted based on laser radar point cloud data, a dynamic target motion prediction map is generated according to a camera image sequence, and real-time traffic flow information is processed to generate a regional traffic efficiency evolution curve. And inputting the data into a path optimization model to generate an initial path sequence, dividing cleaning task priorities, fusing related data and generating a final path planning scheme through a reinforcement learning algorithm. In addition, operation state data of the unmanned vehicle are collected in real time, and a path correction instruction set is generated through an anomaly detection model to update the strategy network. According to the method, the rationality and the operation efficiency of the path planning of the environmental sanitation unmanned vehicle can be improved, real-time monitoring is realized, and the operation safety and the management intelligence level are enhanced.

The invention relates to the technical field of welding defect identification, and discloses a welding defect identification method and system based on a molten pool image. According to the method, a multispectral high-speed camera is used for collecting a molten pool dynamic image sequence in the welding process, after multi-scale morphological filtering preprocessing is conducted, a fusion feature vector is extracted through a depth separable convolutional network, and then the fusion feature vector is input into a defect classification model adopting a heterogeneous graph neural network architecture to obtain a defect probability distribution matrix. Then, constructing a dynamic sparse optimization model to position defects, generating a defect space coordinate set, and finally, outputting welding defect types and position information through hierarchical verification framework processing. According to the method, the problems of welding image noise interference, complex defect characteristics and the like are effectively solved, the accuracy and reliability of welding defect identification are improved, and powerful technical support is provided for welding quality control.

The invention discloses a video super-resolution reconstruction method and system, and relates to the technical field of video restoration processing, and the method comprises the steps: collecting a to-be-reconstructed video frame sequence and a reference frame sequence; extracting a dense optical flow field between the reference frame and the target frame; performing geometric alignment on the target frame to the reference frame according to the dense optical flow field to obtain a motion consistency feature; performing feature extraction on the video frame sequence through a parameterized residual scaling module; inputting the extracted features into a bidirectional propagation module to process forward and backward image sequences respectively, extracting time sequence forward features and time sequence reverse features, and fusing the features to obtain detail features; carrying out adaptive fusion on the motion consistency features and the detail features, and sending the fused features into a reconstruction network to obtain deep visual features; pixel-level adaptive reconstruction is realized according to the deep visual features, and a super-resolution video sequence is obtained; according to the method, the feature space consistency is kept, and meanwhile, the feature matching precision in a complex motion scene is remarkably improved.

The invention provides a textile defect classification detection method, which comprises the following steps: adjusting the radiation intensity of an X-ray source according to a textile yarn movement speed parameter and the receiving sensitivity of a detector, and obtaining a high-resolution and high-signal-to-noise-ratio yarn perspective image sequence through a high-speed image acquisition system; inputting the multi-dimensional defect feature vector into a pre-trained support vector machine classifier, performing preliminary judgment on the yarn defect type, and if the classification confidence is higher than a preset threshold, determining the defect type; if the defect category is a fiber fracture defect, calculating the position of a fracture point and the fracture severity through a fracture degree quantification method, if the defect category is a fiber entanglement defect, analyzing the size and density distribution of an entanglement area, and if the defect category is an impurity defect, estimating the size of impurity particles.

The invention provides an unmanned aerial vehicle multi-dimensional information fusion method and system based on acousto-optic-electric composite detection. According to the method, acoustic frequency band signals, visible light image sequences and electromagnetic field intensity change data are collected, acoustic characteristic flow, optical characteristic flow and electrical characteristic flow are generated, and based on the time deviation between a low-frequency vibration mode and transient electromagnetic pulses, multi-channel dynamic calibration is carried out on an image distortion area in the optical characteristic flow. The method comprises the following steps: acquiring an acoustic feature flow, an optical feature flow and an electrical feature flow, correcting space-time coordinate parameters of the acoustic feature flow, the optical feature flow and the electrical feature flow, generating a calibration feature flow, carrying out anti-interference fusion processing on the calibration feature flow, extracting multi-dimensional coupling features, inputting the multi-dimensional coupling features into an unmanned aerial vehicle flight path planning model, and generating a three-dimensional space situation model; according to the invention, the physical space-time consistency alignment and noise suppression of the multi-modal data are realized, and the obstacle positioning precision and the dynamic trajectory prediction real-time performance are significantly improved.

The invention relates to the technical field of rail traffic safety detection, in particular to a railway freight train key part online monitoring system based on an unmanned aerial vehicle, which comprises an unmanned aerial vehicle control module, a multi-modal data acquisition module, an edge calculation module, a central processing module and a feedback execution module. A dynamic three-dimensional grid flight path is generated by fusing a train Beidou positioning signal and a millimeter wave radar sensing result, a system is provided with an infrared thermal imager, a laser radar and a high-speed polarization camera, bearing temperature, train body point cloud and a train coupler image sequence are obtained, and temperature rise area identification, structural deformation extraction and coupling state modeling are completed on the edge side. The central processing module outputs a multi-dimensional safety assessment result, and the feedback module generates a compensation control instruction and a graded early warning signal based on the risk fusion index. The method has the advantages of being high in autonomy degree, high in recognition precision and high in response speed, and is suitable for full-time structural intelligent inspection and early warning of the freight train in a high-speed operation environment.

The invention discloses a railway track damage detection method, and belongs to the technical field of railway track detection. According to the method, an image acquisition module and an ultrasonic detection module are installed at the bottom of a track detection vehicle, the detection vehicle is controlled to run, and track top face and side face image sequences and ultrasonic reflection signals are acquired; preprocessing the image, and respectively inputting the image into a deep convolutional neural network model and a support vector machine classifier to obtain a crack identification result and a wear level; processing an ultrasonic reflection signal, and judging a layering defect; and finally fusing the data, marking a damage position and generating a structured detection report. According to the method, the problems of incomplete detection, low precision and the like in the existing railway track damage detection are solved, efficient detection of track surface cracks, side abrasion and internal layering defects is realized through collaborative acquisition of the multi-modal sensor, intelligent algorithm processing and data fusion, and the comprehensiveness and reliability of detection are improved.

The invention provides a rehabilitation training action evaluation method and device based on multi-view vision. According to the method, a multi-view camera is adopted to synchronously collect a rehabilitation training image sequence, two-dimensional coordinates of key points of a human body are extracted by improving an HRNet deep learning model, three-dimensional reconstruction is performed in combination with a Gaussian process, motion feature data are extracted, and rehabilitation training quality is evaluated. Accurate motion capture without wearing mark points by the patient is realized, the training constraint feeling of the patient is effectively reduced, the rehabilitation evaluation accuracy is improved, and a personalized rehabilitation scheme can be generated.

Image series transformation for optimal compressibility is performed with neural upsampling and error resilience. A novel correlation network composed of convolutional layers for feature extraction that extract multi-dimensional features from the image and a channel-wise transformer with attention to capture complex inter-channel dependencies. An angle optimizer enhances compressibility of an image and an error resilience subsystem improves robustness against transmission errors and data loss. The error resilience subsystem applies forward error correction coding, data partitioning based on importance, and embeds error concealment hints. This hybrid approach addresses both local and global features, mitigates compression artifacts, improves image quality, and enhances data integrity during transmission. The correlation network incorporates error correction and concealment techniques during decoding. The model's outputs enable effective image reconstruction, achieving advanced compression while preserving information for accurate analysis.

The embodiment of the invention discloses an outer wall thermal insulation defect diagnosis method and system based on artificial intelligence, and the method comprises the steps: firstly obtaining an infrared thermal imaging and visible light image sequence of a target building outer wall, the former comprising continuous temperature distribution data, and the latter comprising textural feature data in time-space alignment with the latter; performing dynamic temperature gradient analysis on the infrared thermal imaging image sequence to generate a three-dimensional heat conduction abnormal map, extracting surface deformation characteristics from the visible light image sequence to generate a structure deformation distribution map, and performing multi-modal characteristic fusion on the two to obtain a joint defect characteristic matrix; performing defect type classification and region positioning on the matrix based on a pre-trained deep residual neural network model, outputting a defect type identifier and a corresponding region boundary coordinate, and finally generating a diagnosis report containing a repair priority score and a material matching suggestion according to the defect type identifier and the corresponding region boundary coordinate, and sending the diagnosis report to a user terminal for visual display. And efficient and accurate external wall thermal insulation defect diagnosis is realized.

The embodiment of the invention discloses a bridge bolt monitoring image recognition method and system based on deep learning, and belongs to the technical field of bridge image processing. The method comprises the following steps: acquiring a multi-angle monitoring image set of a bridge bolt part through an image acquisition device, and preprocessing the monitoring image set to generate a standardized target image sequence; positioning a bolt monitoring area for each image frame in the target image sequence based on a pre-trained positioning model, and generating an area coordinate set containing bolt position information; inputting the region coordinate set into a state recognition model, extracting structural state features of each bolt monitoring region, and generating a looseness degree prediction result of the bolt part according to the structural state features; according to a matching result of the looseness degree prediction result and a preset safety threshold range, a bolt state abnormity alarm signal is output; and associating the bolt state abnormity alarm signal to the corresponding monitoring image frame, and generating a monitoring report containing a visual mark.

The invention provides a textile blending proportion defect detection method, which comprises the following steps: carrying out continuous image acquisition on a blended textile in a vibration state, and adjusting the exposure time and frame rate of camera equipment according to vibration frequency and amplitude characteristics to obtain a textile surface image sequence; the obtained textile surface image sequence is preprocessed, and warp and weft areas are segmented according to the surface texture features of the blended textile; based on the abnormal degree evaluation model, constructing a blending proportion flaw classifier, and training the classifier to identify different types of blending proportion flaws by using the warp and weft abnormal features, the fiber distribution features and the texture features as input; according to a blending proportion defect classifier, in combination with defect positioning requirements in a vibration environment, performing real-time tracking on detected different types of defect proportion areas, compensating image displacement caused by vibration, and ensuring that a defect detection result corresponds to an actual textile position.

The invention discloses an adaptive teaching real-time feedback method based on multi-modal fusion, and the method comprises the following steps: synchronously collecting text modal information, voice modal information and image modal information generated by students in a teaching process, forming multi-modal original data information, and extracting historical student interaction behavior data; preprocessing the multi-modal original data information, and respectively generating corresponding text, voice and image sequence features; a visual feature encoder and a sequence feature encoder are adopted to encode each modal sequence feature to obtain a high-dimensional feature; inputting the modal high-dimensional features into a cross-modal fusion network for deep fusion; parameters of the feedback model are optimized through a model-independent element learning feedback regulation and control algorithm, and a personalized feedback strategy is generated; generating comprehensive feature representation according to the fusion features, and outputting personalized teaching feedback; and the interaction information is updated based on the feedback behavior data to realize closed-loop optimization.

The invention relates to the field of robot attitude control, in particular to a robot attitude control method and system based on intelligent recognition. The method comprises the following steps: continuously acquiring a real-time sandbox scene video stream based on a high-definition camera; and carrying out time sequence frame segmentation on the real-time sandbox scene video stream, and carrying out global contrast enhancement processing to obtain a global contrast optimization image sequence. Carrying out dynamic object identification on the global contrast optimization image sequence, and carrying out continuous frame optical flow motion tracking to generate a dynamic object optical flow change track; optimizing the image sequence according to the global contrast to extract scene ground and scene obstacle data; performing dynamic trajectory rendering on the scene ground and scene obstacle data based on the dynamic object optical flow change trajectory so as to construct a dynamic scene twinborn model; robot position posture parameters and user operation instructions are recognized. According to the invention, timely attitude adjustment response and accurate movement position control are realized.

The invention provides a human body posture detection method and equipment based on machine vision, which are applied to posture recognition in physical fitness detection, and are used for generating a dynamic human body region set of a target user by acquiring a multi-frame continuous image sequence of a physical fitness training scene and performing human body region positioning processing on the multi-frame continuous image sequence. Performing posture feature extraction processing on the dynamic human body region set to obtain a space-time posture feature set of the target user, performing posture state recognition processing on the space-time posture feature set based on a preset physical fitness evaluation model to generate a posture state recognition result of the target user, and generating physical fitness evaluation parameters according to the posture state recognition result, and the physical fitness evaluation parameters are fed back to the physical fitness training guidance system. According to the method, the identification capability of fine posture change in the fitness action of the complex body can be improved, the synchronous evaluation of the action specification degree and the physiological load state is realized, and the limitation of the traditional single-dimensional action evaluation is broken through.

The invention discloses an end-to-end monocular visual odometer method fusing space-time semantic information. According to the method, continuous image sequence frames are collected through a color monocular camera, and a multi-information fusion end-to-end deep learning framework is constructed; a heterogeneous training domain is adopted to set various data set course sharing parameter fusion training, continuous image sequences are input, and the end-to-end deep learning framework is dynamically coupled with hidden state feature vectors output historically, so that a feature mapping relation of time sequence perception is formed; and interpretable feature decoupling of the static background elements and the dynamic entity objects in the scene is realized. After iterative feature fusion, the system outputs sparse depth and camera motion poses which conform to scene geometric constraints, so that a camera trajectory estimation model with high robustness and strong generalization ability in a complex environment is constructed. According to the method, the positioning precision and stability of the monocular vision odometer are remarkably improved.

A method for early diagnosis of Parkinson's disease based on multimodal deep learning is provided. Audio-visual data of a to-be-diagnosed subject while performing a speech task is acquired. The audio-visual data are preprocessed to extract a plurality of audio segments and a plurality of video segments. A face image sequence is extracted from each of the plurality of video segments. A Mel-spectrogram of each of the plurality of audio segments is calculated. The face image sequence and the Mel-spectrogram are input into a multimodal deep learning model to output a classification result for Parkinson's disease early diagnosis of the to-be-diagnosed subject. A system for early diagnosis of Parkinson's disease based on multimodal deep learning is also provided.

The invention relates to an automatic data labeling method of an ISP image signal processing visual sensor. The method comprises the following steps: carrying out noise reduction and signal enhancement on an original image data stream acquired by an ISP image signal processing visual sensor to obtain a standardized image sequence; performing time sequence feature extraction to obtain a time sequence feature tensor; performing spatial feature extraction to obtain spatial-temporal feature representation; inputting a double-fold teacher-student network for semi-supervised training to obtain an initial image annotation model; performing quality evaluation based on the prediction probability distribution and the time sequence consistency to obtain a labeling result with a quality score; and storing a high-quality sample in the labeling result with the quality score into a sample buffer pool, carrying out sample screening based on entropy and distance measurement, and carrying out parameter updating on the initial image labeling model through an incremental learning strategy to obtain a target image labeling model. According to the invention, the reliability evaluation of the labeling result is realized, and the accuracy of automatic labeling is ensured.

The invention provides a square power battery shell visual defect rapid detection method and system, and relates to the technical field of image processing, and the method comprises the steps: obtaining a high-resolution multi-view image sequence, outputting standardized image data, carrying out front visual scratch detection, outputting scratch defect positions and features, and outputting defect candidate regions containing confidence scores; and recognizing bubble defects, summarizing recognition results of the defect candidate areas, and generating a labeling image. The method solves the problems that in the prior art, due to the fact that multi-view image information is not fully fused, traditional scratch detection depends on a single edge feature, and a main defect recognition model is insufficient in recognition capacity for tiny defects such as bubbles, defect detection accuracy is low, missing detection and false detection phenomena are serious, and detection efficiency is high. According to the method, the problems that the defect is difficult to comprehensively and accurately locate and classify in the prior art are solved, the multi-view information fusion capability and the identification accuracy and detection efficiency of multi-class defects are improved, and high-precision, comprehensive and real-time quality monitoring of defect detection results is realized.

The invention provides an artificial intelligence-based ultrasonic image data classification method and system, and the method comprises the steps: firstly obtaining a real-time ultrasonic scanning signal sequence containing the time sequence change characteristics of a tissue elastic parameter and a hemodynamic parameter, carrying out the noise suppression and motion artifact compensation processing, generating a standardized ultrasonic image sequence, and marking the coordinates of an anatomical boundary; then performing multi-scale anatomical structure decomposition on the ultrasonic image to obtain a local feature map set of different organization levels, inputting the local feature map set into a cascade deep classification network, and realizing cross-frame feature fusion and dynamic weight adjustment through spatial-temporal feature alignment and a multi-granularity attention distribution module to obtain a spatial-temporal feature fusion model; and the abnormal region classification probability distribution and the spatial topological relation graph are output, finally, a multi-modal diagnosis report is generated according to the abnormal region classification probability distribution and the spatial topological relation graph, an interactive three-dimensional visual interface containing risk level labels and treatment suggestions is generated after the multi-modal diagnosis report is compared with historical cases, and ultrasonic image classification accuracy and diagnosis efficiency are improved.

The invention provides a magnetic core intelligent cutting parameter self-adaptive optimization system based on multi-mode perception, and relates to the technical field of data processing.The method comprises the steps that a multi-mode sensor module is integrated on magnetic core cutting equipment, and the module comprises a force sensor, a visual sensor and a temperature sensor; the acquisition units are respectively used for acquiring cutting force dynamic signals, cutting track image sequences and cutter temperature time sequence data in real time; magnetic core surface texture features and three-dimensional contour data are captured through a visual sensor, and an initial cutting parameter set is generated in combination with a magnetic core material type recognition result, associated parameters in a historical process database and preset process constraint conditions; and first workpiece trial cutting is executed based on the initial cutting parameter set, multi-modal data fusion collection is synchronously started, cutting force frequency domain feature vectors, a tool temperature change rate curve and cutting surface defect image features are obtained, and multi-modal data are obtained. According to the invention, multi-objective collaborative optimization of processing efficiency and energy consumption is realized.

A computer implemented method, system and non-transitory computer-readable device for a remote deposit environment. Upon receiving a user request, based on interactions with the UI, the method implements an electronic deposit of a financial instrument by activating a camera on the client device to generate a live stream of image data of a field of view of at least one camera, wherein the live stream includes imagery of at least a portion of the financial instrument. The method continues by extracting in real-time, based on the formation of byte array objects from the live stream of image data, data fields from a ranked sequence of imagery to be processed by an optical character recognition (OCR) program resident on the client device. The OCR process extracts one or more data fields from of the financial instrument that are communicated to a remote deposit server to complete the electronic deposit.

The invention relates to the technical field of camera motion compensation, and discloses an IMU-based camera motion compensation method and device and a storage medium, and the method comprises the steps: processing original data collected by a plurality of IMU sensors, and obtaining six-degree-of-freedom IMU fusion data; processing through a linear convolution aliasing model to obtain a high-dimensional motion feature vector; performing fusion to obtain a camera motion parameter prediction value; constructing an advanced network compensation model, carrying out jitter intensity evaluation, and dynamically adjusting a compensation model parameter matrix; performing geometric transformation to obtain an initial stable image sequence and a foreground mask; according to the method, IMU data and image information can be effectively fused, camera motion parameters can be accurately estimated, and a compensation strategy can be dynamically adjusted according to scene features to realize camera motion compensation.