123 results about "Space time domain" patented technology

The invention relates to the technical field of geotechnical mechanics and engineering, and particularly discloses a soft soil foundation settlement monitoring method and system based on multi-field multi-source information, and the method comprises the steps: collecting multi-source data of a target soft soil area; performing constitutive parameter inversion, data standardization and discrete Fourier transform frequency domain conversion on the data to obtain a standardized frequency domain multi-field multi-source data set; based on the data set, a soft soil constitutive parameter library and a soil mechanics physical equation, constructing an FD-PINN frequency domain physical information neural network, embedding the physical equation as a prior constraint, and training the model by adopting an alternating optimization strategy; inputting the real-time frequency domain data flow into the model, and outputting the current settlement amount, the settlement rate and the multi-physical field frequency domain distribution; and in combination with a pre-established large scale model test result, through IDFT inverse transformation, a time-space domain settlement field is reconstructed, multi-stage early warning is triggered, a targeted reinforcement scheme is recommended, the soft soil foundation settlement monitoring precision and the engineering practicability are remarkably improved, and the method is suitable for construction, operation and maintenance of infrastructures such as high-speed rails and highways.

The invention discloses a space-time speckle projection three-dimensional imaging method based on multi-frame optical flow alignment. Firstly, a projector based on DLP is used for projecting a space-time speckle pattern to a measured scene, and a binocular camera synchronously collects a three-dimensional space-time speckle image. The calibration parameters of the binocular camera are used to carry out stereo correction on an acquired original speckle image, and a parallax image is generated frame by frame in combination with a coarse-to-fine single-frame speckle matching strategy. And estimating a two-dimensional inter-frame displacement field between continuous disparity maps by using an optical flow method by taking an intermediate frame disparity map as a reference, compensating motion artifacts in a space-time speckle image, and ensuring strict space-time registration of a dynamic target. And based on the speckle image after motion correction, a speckle matching strategy is expanded to a time-space domain, and high-precision multi-frame three-dimensional measurement of a complex dynamic scene is realized. The method is suitable for performing rapid and high-precision three-dimensional modeling on a moving target in an unstructured environment, and can perform accurate three-dimensional measurement on a high-speed dynamic target undergoing any translation or rotation motion.

The invention belongs to the technical field of robot path planning, and provides a path planning method for an electric power inspection mobile robot, which comprises the following steps of: firstly, constructing an environment grid map, establishing a corner directional expansion safety boundary, and performing safety expansion on an outer right angle of an obstacle; secondly, establishing an intelligent neighborhood selection mechanism based on fuzzy control, and performing adaptive adjustment neighborhood search; then, a multi-target heuristic function fusing the local obstacle density, the safety distance and the traffic difficulty is established, path search is carried out, and an initial path is obtained; then, constructing a path optimization strategy, and carrying out redundant point deletion and B spline curve smoothing processing on the initial path until a target executable path is obtained; and finally, optimizing an executable track in the target executable path in a time-space domain through an improved time-domain elastic band algorithm to obtain an optimal path. According to the invention, accurate obstacle avoidance of complex geometric obstacles can be realized.

The invention relates to an unmanned aerial vehicle track optimization method, system and device in an urban high-density airspace and a medium, and the method comprises the steps: dynamically dividing the urban airspace into space-time four-dimensional grid units of different scales according to the vertical height and the time axis based on the real-time airspace traffic density and environmental conditions; establishing a guide vector field pointing to a target position in each four-dimensional grid unit, and when an obstacle and / or a flight conflict is detected, generating a local disturbance vector and dynamically correcting the guide vector field in real time so as to autonomously generate a candidate flight trajectory; and calculating the energy consumption of the candidate flight path based on an unmanned aerial vehicle dynamics model and an environment model, and selecting a path with the lowest total energy consumption as a final flight path under the condition of meeting safety constraints and task time requirements. Compared with the prior art, in a complex dynamic environment of an urban high-density airspace, unification of real-time efficient planning of the flight path of the unmanned aerial vehicle, safe obstacle avoidance and low-energy-consumption optimization is realized.

The invention, which belongs to the technical field of remote sensing image processing and geological disaster monitoring, discloses a high-risk slope multi-temporal remote sensing image deformation monitoring and early warning system comprising a multi-platform remote sensing data acquisition module, a deep neural network registration module, a sub-pixel-level deformation detection module, a meteorological data fusion module and a dynamic early warning module. The centimeter-level registration precision is realized by introducing an adaptive feature registration mechanism based on a graph attention network, the deformation estimation precision of a texture sparse region is improved by adopting a time-space domain combined sub-pixel-level deformation field estimation algorithm, and the early warning timeliness is improved by establishing a dynamic early warning threshold self-adjustment mechanism fused with meteorological data. The system supports multi-scale multi-platform cooperative monitoring, the early warning advance period is prolonged by 7-15 days, and the system is successfully applied to high-risk slope monitoring projects along the expressway.

The invention provides a video coding and compression transmission method for intelligent equipment, which belongs to the technical field of video coding and compression transmission, realizes advanced channel quality estimation by establishing a channel prediction model enhanced by a spatial transformation network, configures layered coding parameters according to a prediction result and optimizes motion search in combination with inertial measurement unit data. An optimal coding mode is selected by an application rate distortion optimization algorithm, enhancement layer sending is adjusted according to real-time feedback by adopting a dynamic transmission scheduling strategy, and differential forward error correction protection and a time-space domain error concealment mechanism are matched; the technical problem that video transmission is unstable due to violent fluctuation of channel quality when intelligent equipment moves and works in complex electromagnetic environments such as a transformer substation is solved.

The invention relates to a figure operation planning method and device based on satellite remote sensing, equipment and a medium. The method comprises the steps that structured processing is conducted on multi-source satellite remote sensing data, an integrated data packet is obtained, and the integrated data packet comprises a cloud catalysis potential matrix, a terrain slope matrix, an obstacle distribution matrix and a real-time airspace occupation state; analyzing the cloud catalysis potential matrix in the integrated data packet, generating an executable task set, performing task division on the task set, obtaining decomposed tasks, performing resource verification on the decomposed tasks, and generating a verified task set; and obtaining a historical risk coefficient matrix, and constructing an executable operation instruction based on the verified task set, the terrain gradient matrix, the obstacle distribution matrix and the historical risk coefficient matrix. According to the method, by means of multi-source data integration, multi-constraint task generation, instruction construction based on historical risks and real-time environments and the like, the efficiency and accuracy of figure operation planning and the safety and reliability of operation execution can be improved.

The invention discloses a local motion and frequency domain knowledge-guided motion infrared small target detection method, which comprises the following steps of: firstly, constructing a motion infrared small target detection network model, inputting a continuous infrared image, respectively carrying out feature extraction on a spatial domain and a local time-space domain through a fast multi-scale backbone network and an efficient local motion Mama module, and carrying out feature extraction on the continuous infrared image; and the multi-directional adaptive frequency domain enhancement module effectively reduces the sampling information loss of the target information in the fast multi-scale backbone network, and the efficient local motion Mama module fully utilizes the local multi-frame information to carry out time sequence modeling, and finally a key frame target prediction image is obtained, and the motion infrared small target detection is completed. The method can effectively improve the efficiency of feature extraction, reduces false alarms generated by background shaking, and guarantees the real-time performance of multi-frame target detection by adopting an efficient module and a multi-position lightweight design.

The invention discloses a nuclear safety simulation verification method and device based on digital twinning and a medium, and relates to the technical field of nuclear engineering and digital twinning, and the method comprises the steps: training a physical information neural network based on a digital twinning initial model in combination with a nuclide transport equation and a thermal hydraulic equation, and outputting a nuclear safety multi-field simulation result; based on a nuclear safety multi-field simulation result, establishing a space-time diagram structure, extracting dynamic evolution characteristics by using a space-time diagram convolutional network, and outputting a nuclear accident propagation path and a risk distribution diagram; according to the method, the space-time diagram structure is constructed, the space-time diagram convolutional network is utilized to extract the dynamic evolution characteristics, the nuclear accident propagation process is modeled as cascade response in the space-time domain, and the reliability of the nuclear accident propagation process is improved. And the accuracy, the foresight and the intelligent level of nuclear security simulation verification are improved.

The invention relates to the technical field of satellite navigation and positioning, in particular to a Beidou signal anti-interference method based on space-time state prediction.When the Beidou signal anti-interference method is used, time-space domain parameters, including key information such as directions and phase differences of direct and multi-path signals, of Beidou signals are extracted, and multi-path special suppression measures are combined, so that the anti-interference performance of the Beidou signals is improved; a multi-path signal spectrum peak is subjected to graded attenuation, a positioning result is calibrated in cooperation with a scene correction coefficient, finally, the urban canyon scene positioning precision is achieved, high-precision scene application such as automatic driving and precise surveying and mapping can be stably supported, meanwhile, a prediction model integrating LSTM and Kalman filtering is constructed, a multi-path signal prediction result is output in advance, and the accuracy of urban canyon scene positioning is improved. In cooperation with the rapid matching function of the interference feature library, the method can adapt to the dynamic change of multi-path signals, ensures that interference suppression measures take effect in time, and avoids signal receiving interruption or positioning deviation sudden increase.

The invention discloses a swimming skeleton point coordinate data denoising method and device based on spatio-temporal topological structure learning, and belongs to the field of data processing. The method comprises the following steps: acquiring and splicing a swimming video of a target object; detecting a target object in the video, and obtaining skeleton point coordinates and corresponding confidence scores of the target object by adopting a posture estimation model; preprocessing the coordinates to obtain feature vectors; adjacency information of each skeleton point is obtained based on the human body skeleton topology, the feature vectors of all the skeleton points are input into a spatial domain noise removal network, the adjacency information of the skeleton points is aggregated, multi-scale pooling is carried out, and denoised spatial domain features are obtained; sampling the high-dimensional spatial-temporal characteristics by adopting a deformable time convolutional network; and carrying out deformable convolution operation and decoding operation on the sampling features to obtain skeleton point coordinates after time-space domain denoising. Linear and nonlinear noise in data is processed in a time-space domain in a cooperative manner, so that the limitation of a traditional low-pass filter method and an existing ST-GCN method is effectively overcome.

The invention provides a multi-view three-dimensional virtual-real fusion rendering method and related equipment, and relates to the field of computer vision, a target scene is divided into a space-time flow line or local space-time blocks by constructing a geometric field, an appearance field and a space-time flow field on a four-dimensional time-space domain, fragmentation caused by pure frame-by-frame rendering and scene specific rules is avoided, and the rendering efficiency is improved. The problem of multi-scene splitting is solved, and space-time importance distribution at each moment and under each virtual view angle is calculated through the updated feature vector and the attention weight corresponding to the updated feature vector so as to render a geometric field and an appearance field on a four-dimensional time-space domain. And the rendering picture quality is improved on the premise that the total calculation overhead is not obviously increased.

The invention provides a frequency-wavenumber domain-based matrix array full-matrix capture three-dimensional ultrasonic imaging method and system, and belongs to the technical field of ultrasonic imaging. Five-dimensional time-space domain radio frequency data are collected through a two-dimensional full-sampling matrix array and are converted into a frequency-wave number domain through five-dimensional fast Fourier transform, angular frequency is mapped into axial wave number through Stolt interpolation, data resampling is achieved, then three-dimensional image frequency spectrum data are synthesized by defining spatial wave number coordinates of an image and conducting coherence stacking, and the image frequency spectrum data are obtained. And finally, reconstructing a high-quality three-dimensional ultrasonic image through three-dimensional inverse fast Fourier transform. Compared with a traditional delay summation beam forming method, the method has the advantages that the calculation efficiency is remarkably improved, the speed is increased by several times to dozens of times, the spatial resolution, the contrast noise ratio and the spot signal-to-noise ratio are remarkably improved, and the method is suitable for high-resolution real-time three-dimensional and four-dimensional ultrasonic imaging application.

The invention relates to the technical field of information security, in particular to an AI video detection method based on multi-feature branch fusion, and the method comprises the following steps: extracting a frame extraction color image from a video, and obtaining a standardized frame sequence after preprocessing; calculating an inter-frame differential volume; obtaining a time sequence spectrum volume according to the standardized frame sequence; performing bilateral filtering decomposition on the standardized frame sequence to obtain an illumination consistency volume; respectively inputting the three types of volume features into a deep convolutional neural network, and after feature extraction and fusion, outputting an AI forgery probability through a classifier; and comparing the AI forgery probability with a decision threshold to generate a video category label, and calculating an index. According to the method, the multi-dimensional features of the time-space domain, the frequency domain and the physical illumination domain are fused, the potential traces of the deeply-forged video are effectively captured, the stable detection performance is kept under various video quality conditions, the accuracy is improved, and reliable technical guarantee is provided for media information security.

The invention discloses a self-adaptive optimization method of a time-space domain intelligent sensing collaborative compression algorithm. According to the method, the adaptive coupling matrix can dynamically adjust the time and space feature weights according to the fed-back game conflict, and when the demand conflict occurs between the time compression strategy and the space compression strategy, the matrix can enhance the proportion of key features in a targeted manner, so that information imbalance caused by unilateral optimization is avoided; therefore, the retention capability of the core spatio-temporal information in the compression process is enhanced, the redundancy loss of non-key details is reduced, the compression strategy is enabled to be more fit with the internal law of data, and the adaptive adjustment capability of the algorithm on dynamic scene and resource constraints is enhanced by the linkage mechanism. Game optimization takes the fused feature vector as a decision basis, and the coupling matrix is continuously optimized through conflict feedback, so that the algorithm can respond to scene demand changes in real time, the compression effect stability under different application scenes is effectively improved, the utilization efficiency of system resources is enhanced, and the problem of resource waste or overload is avoided.

The invention provides a light-weight pedestrian motion prediction method based on a mask, and is suitable for the technical field of human-computer interaction. According to the method, human body 3D skeleton point data is processed through space and time masks, key features are extracted by using a local perceptron constructed by a lightweight multilayer perceptron module and a cross-frame fusion device, the features are fused by using 1 * 1 convolution and splicing technologies, then a frequency-space domain pedestrian prediction action sequence is generated through a prediction module, and the frequency-space domain pedestrian prediction action sequence is obtained. And converting into a time-space domain sequence through an inverse discrete cosine converter. The method has the advantages of light model, quick response and suitability for real-time interaction. In addition, the strategy of gradually increasing the number of training frames improves the accuracy and stability of prediction.

The invention discloses an image processing method based on deep learning and Fourier time-space domain transformation, and the method comprises the specific steps: image processing: collecting an original image frame from a video, inputting the original image frame into a convolutional neural network, carrying out the convolution of horizontal and vertical latitudes on an image, outputting an operation result through a convolution kernel, and carrying out the image processing; a modified image is obtained; and carrying out feature extraction through BN operator operation: carrying out normalization processing on a result output by the deep convolution operation, unifying data distribution of different network layers, carrying out data transformation and reconstruction after normalization, and recovering feature information of original data. The BN operator operation is carried out on the result output by the deep convolution operation, the BN operator operation enables the distribution of input data of each layer in the network to be relatively stable, the learning speed of the model is increased, the regularization effect on the model is achieved to a certain extent, and the analyzed result is more accurate.

This invention relates to a vehicle trajectory prediction method based on heterogeneous node spatiotemporal domain perception. It establishes an online perception map generated from the perspective perception results of the driving vehicle and uses a spatiotemporal domain attention mechanism to complete the vehicle trajectory prediction task by understanding and fusing different types of information. It can effectively obtain map information within the local range of the target vehicle without using high-precision maps, and complete the trajectory prediction while ensuring high robustness and estimation accuracy, which is more in line with the needs of practical applications.

The application discloses a complex action-based spatio-temporal frequency domain feature fusion action recognition method and system. The application preprocesses the electroencephalogram signal of the complex action motor imagery paradigm, extracts the global spatio-temporal feature by using a time-space partition branch network TG-Net, extracts the frequency domain feature by using a frequency branch network F-Net for the power spectrum density of the preprocessed signal, and performs action recognition after flattening, splicing and the like of the frequency domain feature and the global spatio-temporal domain feature. The application constructs a time-space frequency domain double branch network, deeply fuses and splices the global spatio-temporal feature and the power feature, breaks through the limitation of the traditional method which is limited in the time-space domain, realizes the complementary feature extraction of the motor imagery electroencephalogram signal in the time, space and frequency dimensions, and adopts the partition weighting splicing mode of the functional brain area in the space dimension to highlight the important channel action recognition method, so that the comprehensive feature is acquired, and the classification precision is improved.

The invention provides a submarine cable anchor damage detection method and device, electronic equipment and a computer medium. The submarine cable anchor damage detection method comprises the following steps: acquiring target time-space domain data, wherein the target time-space domain data represents phase difference data generated by vibration of a submarine cable in a time-space domain; calculating a mask matrix of the target time-space domain data, and taking the mask matrix as target mask data; obtaining a target mask image based on the target mask data; determining a plurality of target contour images from the target mask image, wherein the plurality of target contour images represent the edge distribution condition of the submarine cable in the time-space domain when the submarine cable is subjected to external force; a plurality of morphological characteristic curves are obtained based on the plurality of target contour images, one morphological characteristic curve is matched with one target contour image, and the morphological characteristic curves represent edge characteristics of the submarine cable in a time-space domain; and according to the plurality of morphological characteristic curves, generating an anchor damage detection result of the submarine cable. Submarine cable anchor damage can be timely alarmed, and submarine cable damage is avoided.

The application provides an infrared image processing method based on a space-time field and related equipment, and relates to the technical field of infrared image processing. The method comprises the following steps: acquiring continuous infrared images, constructing a unified background image, identifying key points of each target moving object and mapping the key points, determining an image sequence group according to a time sequence, and obtaining a complete motion trajectory curve according to the distribution of the key points. When a split-screen target infrared image is detected, the key points of each part are extracted after splitting, the two parts of the background are matched with the unified background image to determine an initial background area, the distance between the key points and the expected position predicted based on the trajectory, adjacent frames and motion rules is calculated, the current frame and the delayed frame are determined, and finally the complete sequence is obtained by processing the delayed frame. The method in the application can accurately repair the split-screen image in a dynamic scene, and ensure the time sequence coherence and integrity of the infrared image sequence.

The invention provides an efficient cooperative construction method, device and equipment for simultaneous operation of multiple geophysical prospecting teams, relates to the technical field of seismic prospecting, and aims to solve the problem that the construction efficiency is low under the condition that the distance between shot points constructed by the multiple geophysical prospecting teams in the same time window in adjacent work areas is smaller than the minimum distance of synchronous excitation in the prior art. The method comprises the following steps: integrating adjacent work areas for analysis, calculating the lower limit value of the root-mean-square energy of the upper cannon interference of a conventional production cannon, and drawing a first time distance curve meeting the condition that the upper cannon interference is not greater than that of the production cannon by utilizing a single-cannon data single-cannon energy attenuation time-space domain distribution plane graph of a long-record long-arrangement-point test; according to the seismic data of the work area, obtaining the lowest refracted wave speed, and calculating a second time distance relation curve when the last shot refracted wave does not enter the single shot data; and an area clamped by the first gun loading interference time distance curve and the second gun loading interference time distance curve is used as a distance range of staggered excitation of two guns in adjacent work areas, a blasting time interval is dynamically calculated, and staggered excitation is carried out.

The invention discloses a small target detection method based on a time-space domain recursive network. The method comprises the following steps: step 1, preprocessing image data as input data; 2, constructing a spatial-temporal feature extraction backbone network, inputting the input data into the backbone network, and generating feature maps of four stages; 3, constructing a time domain recursion module, and generating a fused spatial-temporal feature based on the feature map of the third stage; and step 4, constructing a lightweight detection head, constructing multi-scale features based on the spatial-temporal features and the feature maps of the other three stages, and outputting a classification result and bounding box coordinates through the lightweight detection head. The method has high detection precision, and the spatial-temporal feature fusion effectively enhances a target signal and inhibits background interference. The method is suitable for a static platform and a motion platform, does not need complex inter-frame registration, and has no strict constraint on a target motion mode; and meanwhile, the algorithm complexity is low, a fixed weight convolution kernel or a lightweight network can be adopted, and hardware implementation is facilitated.

The application provides a three-dimensional recursive filtering method and device for image space-time domain joint noise reduction. The image space-time domain joint noise reduction method of the application is different from the traditional three-dimensional recursive filtering method which is directly connected in series by three one-dimensional filters. Instead, the application integrates the information of time domain recursion in the process of point-by-point space domain recursion to improve the effect of space domain filtering. Then, the information of space domain recursion is integrated in the process of point-by-point time domain filtering to improve the effect of time domain filtering. Such repeated recursion traverses the whole image. Under the same level of operation logic and storage cost, compared with the traditional space-time domain series and space-time domain parallel methods, the application can better improve the image definition and signal-to-noise ratio while reducing the phenomena of motion trail lack and raindrop noise.

The application discloses an ultrasonic signal classification and identification method, medium and equipment, and belongs to the technical field of signal processing and deep learning; the method comprises the following steps: collecting ultrasonic signals of excitation channels and receiving channels under different working conditions in structural ultrasonic detection, reconstructing channel signals into time sequence samples; inputting the time sequence samples into a double-branch deep learning model, and outputting working condition category probability distribution results; the double-branch deep learning model obtains the working condition category probability distribution results through a time domain feature extraction module, a space domain feature extraction module, a feature fusion module and a classification and prediction module; the application fully considers and utilizes the excitation and receiving signal features in ultrasonic detection, adopts a time and space domain feature fusion mode to fully capture multi-scale feature information and perform deep-level fusion, solves the problem that a single network is difficult to represent ultrasonic signal features from a multi-scale level, has strong engineering practicability, and can be widely applied in the fields of structural nondestructive detection, structural health monitoring and the like.

The application provides a multi-view three-dimensional virtual-real fusion rendering method and related equipment, and relates to the field of computer vision. By constructing a geometric field, an appearance field and a space-time flow field on a four-dimensional space-time domain, a target scene is divided into a space-time flow line or a local space-time block, fragmentation caused by pure frame-by-frame rendering and scene-specific rules is avoided, a multi-scene fragmentation problem is solved, a space-time importance distribution under each time and each virtual view is calculated by using an updated feature vector and an attention weight corresponding to the updated feature vector to render the geometric field and the appearance field on the four-dimensional space-time domain, and the rendering picture quality is improved without significantly increasing the overall computing overhead.

The application provides a gas leakage three-dimensional reconstruction method and system based on thermo-optical fusion, relates to the technical field of three-dimensional reconstruction, and comprises the following steps: multi-scale features are extracted from a thermal radiation image sequence and a visible light image sequence, and are fused based on mutual information; a space-time causal graph is constructed to perform directed message passing to obtain enhanced node features; the enhanced node features are mapped to a continuous space-time domain and are transformed by a Fourier neural operator to obtain a feature field; the feature field is used as a condition to guide iterative denoising; and finally, a three-dimensional space distribution model of gas leakage is generated. The application can accurately reconstruct the three-dimensional space distribution of gas leakage, and improve the accuracy and reliability of leakage monitoring.

The invention provides a non-reference cloud game video quality evaluation method and equipment based on local and global information fusion, and the method comprises the following steps: 1, obtaining a cloud game video needing to be evaluated, extracting K frames of images from the whole video frame at equal intervals as key frames, step 2, respectively sending the extracted key frame into a depth feature extraction branch and a traditional feature extraction branch, for the depth feature extraction branch, cutting an area with a specified size from the key frame, processing the area to obtain global texture features, and for the traditional feature extraction branch, obtaining global texture features; and step 2, obtaining local depth features from a specific time-space domain local region of the key frame through a residual convolutional network, and step 3, splicing the features obtained by the two branches and sending the spliced features to a multi-layer perceptron quality regression device to obtain a final video quality score. According to the scheme, global and local information can be jointly utilized, the quality problems of blocking effect, fuzziness, texture distortion and the like are effectively reflected, and high-precision and light-weight cloud game video quality evaluation is achieved.

This invention discloses a method and system for representing lane-changing scenarios involving intelligent vehicles and human vehicles in mixed traffic based on a spatiotemporal network. The method includes: acquiring motion state data and environmental interaction data of all vehicles in the mixed traffic flow within a target road segment; constructing a basic architecture of a discretized spatiotemporal network covering the target spatiotemporal domain based on the motion state data and environmental interaction data; modeling the lane-changing behavior of intelligent vehicles as deterministic trajectories in the discretized spatiotemporal network, and modeling the uncertain lane-changing behavior of human vehicles as probability clouds in the discretized spatiotemporal network, while simultaneously constructing a risk coefficient field; establishing interaction constraints between the deterministic trajectories and the probability clouds, and transforming the interaction constraints into mathematical constraints; constructing an objective function for the discretized spatiotemporal network, and obtaining the final discretized spatiotemporal network based on the risk coefficient field; and obtaining the optimal spatiotemporal trajectories of intelligent vehicles and human vehicles based on the final discretized spatiotemporal network.

The present application belongs to the technical field of signal-to-noise ratio enhancement, and particularly relates to a dark and weak extended target signal-to-noise ratio enhancement method based on time-space domain joint detection. The method comprises the following steps: S1: performing preliminary noise reduction on a dark and weak extended target image sequence; S2: processing each dark and weak extended target image contained in the dark and weak extended target image sequence after preliminary noise reduction by using an adaptive weighted fusion noise reduction function; S3: inputting the dark and weak extended target image sequence after processing and deducting time domain noise into a trained deep convolutional neural network model for processing, to obtain a dark and weak extended target image sequence after deducting time-varying background noise; and S4: performing image enhancement on the dark and weak extended target image sequence after deducting time-varying background noise, to obtain a dark and weak extended target image sequence with high signal-to-noise ratio. The present application realizes effective suppression of time domain random noise, real-time estimation and elimination of sky light time-varying background, and greatly improves the detection capability of ground-based photoelectric detection equipment under strong sky light background in the daytime.