3228 results about "Video image" patented technology

The present disclosure relates to the field of computer vision and discloses a three-dimensional dynamic scene reconstruction method and apparatus, and a storage medium. The three-dimensional dynamic scene reconstruction method comprises: acquiring synchronized videos of a plurality of viewpoints of a dynamic scene; computing matching points between video images of different viewpoints, and estimating intrinsic and extrinsic parameters of each camera; obtaining a Gaussian splatting point set {p0} on the basis of a sparse point cloud constructed according to the depth of each matching point; for the first image frame of each video, using {p0} to perform static training thereon, to obtain a Gaussian splatting point set {p}; for the remaining image frames, dividing {p} into a static point set {S} and a dynamic point set {D}, performing dynamic training on {D}, and constructing a dynamic Gaussian splatting point set {P} from {p}, {S}, and the final {D}; and, in view of the intrinsic and extrinsic parameters of each camera, rendering {P} using a Gaussian splatting rendering pipeline, to obtain rendered images at different moments from new viewpoints.

The invention discloses an insurance claim settlement-oriented multi-modal image video evidence analysis method and system. The method comprises the following steps of: acquiring video / image and multi-source data such as metadata, audio, IMU (Inertial Measurement Unit), GPS (Global Positioning System), OBD (On-Board Diagnostic) and the like; calculating content Hash of the video and the audio according to frames, connecting the content Hash with time information in series to form chained Hash, and adding a verification digital signature and a credible timestamp; realizing cross-modal time sequence alignment based on self-adaptive time anchor-attitude coupling; tampering detection is carried out in combination with PRNU fingerprints, noise field consistency, dual compression, copy-movement and the like; multi-view geometry and monocular depth are fused, IMU scale constraint and micro rendering are introduced, three-dimensional reconstruction and re-projection optimization are completed, and collision dynamics verification is carried out; and constructing an event cause and effect graph, judging responsibility in combination with traffic rules, outputting a confidence coefficient vector and a structured report, and generating a verifiable evidence packet. The scheme has the advantages of high efficiency and traceability in the aspects of space-time restoration and interpretable responsibility judgment.

The invention discloses a long video multi-modal understanding and question-answering method and system based on large model and retrieval enhancement generation. The method comprises the following steps: 1) a multi-modal feature extraction module; 2) a multi-modal synchronization and alignment mechanism; 3) constructing a structured memory pool; 4) querying a drive generation mechanism; 5) incremental updating and memory compression strategy; and 6) unifying the multi-modal representation space. The invention provides a long video multi-mode understanding method fusing a large language model and retrieval enhancement generation, and aims to break through the limitation of a traditional method in the aspects of single-mode processing and semantic fragmentation. According to the method, video image features are extracted through a visual model (such as YOLO and ViT), voice transcription and environment voice description are obtained in combination with an audio model (such as Whisper and Qwen-Audio), and unified coding of vision, voice and audio in a long video is achieved. Then, a structured memory pool is constructed through semantic consistency segmentation and timestamp alignment technologies to store time slice data of different modalities.

The invention relates to the technical field of engineering construction supervision, and discloses a building engineering construction supervision system based on big data analysis, and the system comprises a multi-modal data collection module which is used for collecting multi-source data of a construction site, and the multi-source data comprises structure sensor data, environment monitoring data, video image data, construction log data and building information model (BIM) state data; performing standardization processing and time synchronization on the data to generate a construction state data sequence; and the construction event modeling module identifies key events in the construction process based on the construction state data sequence and constructs a construction event graph, and the construction event graph is composed of event nodes representing construction events and event edges representing event collaboration or time correlation. By introducing an event atlas construction mechanism based on multi-source construction data driving, structured expression and semantic association mapping of key behavior units of a construction site are realized, and the problem of insufficient non-structured information processing capability in construction monitoring is overcome.

The invention discloses a high-precision instrument assembly fault backtracking method and system, belongs to the field of precision manufacturing, and aims to solve the problems that in a traditional backtracking method, assembly data are scattered and unreliable, fault root positioning is fuzzy, and new scene adaptation depends on a large amount of data. The method comprises the following steps: collecting assembly structured data, video images and environment data in a multi-source manner, filtering out low-quality images, and distributing unique identifiers for products; fusing the multi-modal data to generate a depth feature matrix; constructing an anomaly detection model to output a risk score and a label; hashing the data and then storing the data into a product exclusive private block chain; when a fault occurs, extracting data on the chain through a unique identifier, reconstructing an assembly process by using a graph neural network, and comparing a standard positioning root; and based on the fault report incremental training model, parameters are optimized in combination with meta-reinforcement learning. According to the method, the data authenticity is guaranteed, the fault backtracking precision and efficiency are improved, a new scene is quickly adapted, the production rework rate is reduced, and the stable assembly quality is maintained.

The invention provides a lottery store violation detection method and system based on multi-modal data fusion, and relates to the technical field of intelligent monitoring, and the method comprises the steps: detecting a current abnormal event, carrying out time sequence perception target detection on monitoring video data, and recognizing violation electric equipment in a video image frame based on adaptive feature enhancement and multi-target tracking. The method comprises the following steps: acquiring time sequence data of current abnormity and illegal electric equipment detection, constructing a time sequence incidence relation of current abnormity, equipment detection and an electric state based on a dynamic causal network, calculating a time-varying weight and instantaneous causal intensity by utilizing conditional entropy increment, acquiring optimal time lag in combination with eigenvector conversion and a dynamic programming algorithm, and determining the current abnormity and illegal electric equipment detection according to the optimal time lag. And weighting the integral of the instantaneous causal intensity and the exponential function of the optimal time delay to obtain an event matching score, and distinguishing a temporary power utilization event and a continuous illegal power utilization event based on the event matching score.

The invention relates to the technical field of transformer substation intelligent patrol, and provides a transformer substation three-dimensional fusion patrol method and system based on digital twinborn and autonomous identification. According to the method, a fused three-dimensional model is constructed through multi-source data acquisition and a three-dimensional Gaussian splash algorithm, and in combination with deep learning-based point cloud semantic segmentation and clustering, an equipment-patrol means coverage relationship is generated. Creating a virtual inspection proxy object based on a three-dimensional virtual environment, and controlling terminals such as an unmanned aerial vehicle to collect real-time video image data; the system carries out automatic identification on pictures, automatically completes equipment level alignment and standard point location identification, generates fine control of camera zooming, horizontal rotation, pitching and the like, and realizes standardized view finding and acquisition. By combining an enhanced recognition algorithm, traditional image processing and a deep learning model are fused, model self-evolution is realized through incremental learning, flexible expansion and collaboration of various patrol terminals are supported through a unified interface, and refined, real-time and intelligent patrol operation and maintenance requirements of an intelligent substation are met.

The invention discloses a smart mine safety production violation behavior identification method, system and device and a medium, belongs to the technical field of smart mine safety identification, and aims to solve the technical problem of how to improve the accuracy and efficiency of mine safety production violation behavior identification, realize real-time and accurate safety supervision of the whole process of mine operation and improve the safety of mine safety production violation behaviors. According to the technical scheme, the method comprises the following steps: data acquisition and preprocessing: installing a camera in a key operation area of a mine to acquire video image data, and carrying out denoising, graying and normalization preprocessing operation on the video image data to obtain preprocessed video image data; and constructing a deep learning model based on a convolutional neural network: introducing an attention module into the network structure of the deep learning model, and training the deep learning model by using the marked video image data including the safety production violation behavior and the normal operation behavior, and adopting a transfer learning method in the training process.

The invention relates to the technical field of image target tracking, and discloses a video image target tracking method and processing device based on multi-feature fusion, and the method comprises the steps: obtaining an initial position parameter and a current frame detection parameter of a target; calculating according to the initial position parameter to obtain a pixel-level motion vector field; performing prediction according to the pixel-level motion vector field to obtain a prediction position parameter of the target; comparing the predicted position parameter with the current frame detection parameter, and when the difference between the predicted position parameter and the current frame detection parameter is smaller than a preset threshold value, judging that the target is not shielded; and when the difference between the predicted position parameter and the current frame detection parameter is greater than a preset threshold value, tracking error accumulation may be caused by too early view angle switching in the prior art, and in a resource-limited scene, the calculation overhead of a deep learning model is generally large, resulting in poor tracking accuracy and real-time performance.

The invention provides an under-mine video stream dynamic denoising method based on multi-modal fusion, which comprises the following steps: constructing a time sequence synchronous fusion mechanism of visible light, infrared and laser radar data, and realizing time-space alignment of multi-source heterogeneous data; a dynamic noise model is established by introducing a fractional calculus optical flow field concept and combining a Gaussian mixture model, so that a dynamic noise region is accurately identified; an improved self-adaptive wavelet threshold function is constructed, a function threshold parameter can be linked with a dust concentration sensor in real time, and the de-noising intensity is dynamically adjusted according to the actual dust concentration; designing a dual-path feature enhancement neural network to effectively separate and enhance structural features and texture features in the video image; a cascaded detection decision system is created, a lightweight network is used as a primary detector, a high-confidence detection result is directly output, and a low-confidence detection result is input into a Transform correction module for secondary reasoning. According to the invention, dynamic denoising, feature enhancement and target intelligent monitoring of the video stream under the mine can be realized.

The invention discloses an electric locomotive rail obstacle identification method, device and system and electronic equipment, and relates to the technical field of rail transit. The method comprises the following steps: acquiring geographic information along a track and current positioning information, acquiring current radar information to perform obstacle detection when a driving state is determined according to the current positioning information, and performing a first alarm and acquiring current video image information when a suspected obstacle is detected; performing time-space synchronization and data fusion on the current positioning information, the current radar information and the current video image information, and inputting the fused data, the locomotive running state data and the geographic information into a pre-established obstacle recognition model to recognize a suspected obstacle to obtain a recognition result; and determining a risk level according to the identification result, so that the electric locomotive gives a second alarm according to the risk level and takes countermeasures. According to the invention, multi-source data is acquired through various sensors, accurate identification of small targets is realized, and the safety of railway transportation is guaranteed.

The invention provides a high-altitude operation risk early warning method and system based on camera image recognition, and relates to the technical field of computer vision, and the method comprises the steps: firstly collecting a video image sequence of a high-altitude operation scene, and generating a fusion feature map containing environment and operation main body features through multi-level feature extraction; performing spatial dimension segmentation and regional feature comparative analysis on the fusion feature map to obtain a spatial risk distribution map containing risk region identification information, processing the spatial risk distribution map of continuous frames based on a time sequence feature fusion rule to generate a dynamic risk evolution map, and calling a risk decision model to perform mode recognition to obtain a dynamic risk evolution map; and generating a risk level classification result and a risk position coordinate set according to the risk level classification result and the risk position coordinate set, and finally generating a risk early warning signal and sending the risk early warning signal to the monitoring terminal, thereby comprehensively, accurately and dynamically monitoring the high-altitude operation risk, and improving the accuracy and timeliness of risk early warning.

The invention provides a safety production behavior monitoring method and system based on AI video analysis. The method comprises the steps of collecting a real-time video data stream of a production area; inputting each frame of video image in the real-time video data stream into a target detection model for target detection to obtain a personnel target output by the target detection model and a target position coordinate of the personnel target in each frame of video image; cutting out a local image area of the personnel target in each frame of video image based on the target position coordinate, and performing feature recognition based on the local image area to obtain personnel features; performing comparison on the basis of the personnel characteristics and the personnel standard behavior characteristics to obtain personnel behavior states, and performing track association on the basis of the personnel behavior states corresponding to the continuous multi-frame video images to obtain personnel behavior tracks; and performing safety production behavior monitoring based on the personnel behavior state and the personnel behavior track, and generating an abnormal behavior early warning signal. According to the method and the device, the real-time performance and the accuracy of safety monitoring in a production scene are improved.

The invention discloses a gesture recognition method and device based on deep learning, and the method comprises the steps: collecting a video stream, and obtaining a hand key point data set in the video stream; performing data enhancement and preprocessing on the hand key point data set; establishing a gesture recognition model based on deep learning, and performing training optimization; and carrying out lightweight processing on the trained gesture recognition model, and outputting a gesture recognition result in real time based on the lightweight gesture recognition model. According to the method, through pre-segmentation of the video image, a progressive polynomial attenuation pruning strategy, collaborative optimization of pruning and quantification and a comprehensive callback mechanism, real-time, accurate and efficient operation of a gesture recognition model on mobile equipment and an embedded system is achieved, and the high-standard requirement of a modern intelligent interaction system is met.

The invention provides a photoelectric pod target identification and tracking system based on a multi-scale attention mechanism, and belongs to the technical field of intelligent vision. Through combination of a multi-scale convolution module and a multi-head self-attention mechanism, accurate detection and tracking of a target in a photoelectric pod video image are realized. The multi-scale convolution module adopts convolution kernels of different sizes, and can extract local features of different scales to adapt to the change of the size of a target; the multi-head self-attention mechanism is used for capturing global features, especially long-distance dependency relationships between targets and backgrounds and between targets. Through fusion of local features and global features, the system improves the precision and robustness of target recognition in a complex scene. Meanwhile, by optimizing the structural design and the feature aggregation method, the calculation complexity of the system is remarkably reduced, and the requirements of the photoelectric pod for real-time performance and high efficiency are met.

The invention provides a real-time video image compression method based on deep learning, and relates to the technical field of video image compression, and the method comprises the steps: carrying out the key feature recognition through employing an attention mechanism; performing convolution training optimization on the video image sample data set by using a deep learning network structure; a self-encoder structure is designed to carry out feature map encoding compression; a video image compression adaptive network is generated through series fusion; a real-time video image frame is collected for preprocessing, and feature compression processing is performed on a standard video image frame based on a video image compression adaptive network. According to the method and the device, the technical problem that the video compression quality is reduced due to the fact that the generalization ability is insufficient in the face of various scenes and the video compression strategy is difficult to adaptively adjust according to different scenes in the prior art can be solved, the adaptive network is constructed through the combination of deep learning and the auto-encoder, and the video compression quality is improved. And the video compression strategy is dynamically adjusted according to the contents of different video images, so that the video compression quality is improved.

The invention relates to the field of traffic management, and discloses a traffic situation prediction method based on multi-source heterogeneous data fusion, which comprises the following steps of: firstly, acquiring traffic situation related data of a target area from a plurality of data sources, including traffic flow data, vehicle speed data, video image data, meteorological data and historical traffic statistical data; secondly, preprocessing the acquired traffic situation related data, including data cleaning, normalization or standardization processing, and performing time-space synchronization and matching; wherein the data cleaning comprises noise removal, abnormal value processing and missing value filling; and finally, inputting the preprocessed data into a pre-trained traffic situation prediction model, and outputting the predicted traffic jam degree of the target area. According to the invention, comprehensive analysis is carried out through the traffic-related situation data and the emergency data, and finally, the purpose of improving the prediction comprehensiveness through a multi-source cooperation mechanism is achieved.

The invention relates to the technical field of computer vision and virtual reality, and discloses a three-dimensional video fusion method based on camera self-calibration and projection texture mapping, and the method comprises the following steps: S1, obtaining video data, and collecting at least one frame of two-dimensional image in a to-be-fused video stream; optionally, the two-dimensional image is preprocessed; and S2, calibrating internal reference of the camera, detecting linear features in the two-dimensional image by using an image processing algorithm, estimating the position of a vanishing point by using a least square method or other optimization algorithms based on the detected linear segment, and calculating an internal reference matrix of the virtual camera according to the optimized vanishing point position. The perspective relation between the video image and the surface of the three-dimensional model is determined through the vanishing point detection technology, accurate fusion of the video image and the three-dimensional model is achieved, the sense of reality of a virtual scene is improved, and the tedious camera calibration process and the complex three-dimensional reconstruction process based on a calibration plate are avoided.

The invention relates to the technical field of water flow monitoring, and particularly discloses a flow monitoring method and system based on video image analysis and processing. The method comprises the following steps: carrying out communication acquisition of image monitoring and water level monitoring at a plurality of monitoring acquisition sites; performing video frame extraction and image preprocessing; feature point detection and displacement analysis; converting the water surface flow velocity into a plurality of station water surface flow velocities, and performing fitting correction; and calculating station flow data of the plurality of monitoring acquisition stations. Water surface image data are obtained through image monitoring and water level monitoring communication collection of a plurality of monitoring collection sites in combination with video frame extraction and image preprocessing, the site pixel flow velocity is calculated through feature point detection and displacement analysis, the site pixel flow velocity is converted into corrected water surface flow velocity through a space projection model, and site flow data are calculated. Non-contact full-flow monitoring is achieved, the potential safety hazard that equipment needs to be arranged in water in a traditional method is eliminated, the maintenance difficulty is remarkably reduced, and the method is particularly suitable for flow monitoring in flood periods and dangerous water areas.

The invention relates to the field of video processing, in particular to a video data processing method and device, equipment and a medium. In a background replacement link, based on precise operation of video processing requirements, an adaptive algorithm can be selected according to scene characteristics, and errors are preliminarily reduced. And subsequently, error compensation is carried out on the generated intermediate video data, the error region is corrected in a targeted manner, the edge is filled and optimized by utilizing the edge pixel characteristics of the foreground region, and iteration processing is carried out until the error is lower than a preset value, so that the accuracy of background replacement is greatly improved. And finally, the target video data is injected into the virtual camera of the cloud mobile phone, the method is applied to a mobile terminal scene, the background and the foreground are naturally fused under high-precision scenes such as live broadcast and virtual conferences, the image flaws are remarkably reduced, the strict requirements of a user on the video image quality are met with a high-precision background replacement effect, and the overall user experience is improved.

A real-time badminton action detection system, consisting of: a video recording module configured to continuously record video images at a frame rate of at least thirty frames per second; a pose estimation processing unit configured to detect and output two-dimensional skeletal landmark coordinates for a variety of body joints, including at least wrists, elbows, shoulders, hips, knees and ankles, from each video frame; a Motion Bidirectional Encoder Representation Transformer (Motion-BERT) configured to receive sequential skeleton landmark coordinates over a defined time window and encode motion trajectories using multi-head self-attention mechanisms across past and future frames; and a classification controller module operationally coupled to the Motion-BERT, wherein the classification controller module comprises a dense neural network with a softmax output layer configured to generate real-time probability distributions over a variety of badminton-specific action classes, the end-to-end system being configured to produce recognition results with a processing latency of less than 100 milliseconds; and wherein the video recording module comprises a high-speed digital camera with a wide-angle lens positioned at a point on the perimeter of the court, the camera being calibrated with intrinsic and extrinsic parameters for perspective correction, and wherein the system includes a calibration routine that aligns detected skeletal landmarks with a reference badminton court coordinate system.

The invention discloses a tunnel face construction area safety monitoring system and method, and belongs to the technical field of construction safety monitoring. The method comprises the following steps: collecting multiple types of monitoring data of a tunnel face construction area in real time, and preprocessing the monitoring data; extracting abnormal event features in the tunnel face video / image based on a target detection model; the integrity of the tunnel face is judged based on a tunnel face integrity judgment module; fusing the collected multi-source data, and generating a comprehensive safety coefficient S in real time; and triggering different levels of alarm response measures according to the interval in which the safety factor S is located. According to the method, the risk early warning precision of the tunnel face construction area can be effectively improved by integrating multi-modal data fusion, the self-adaptive warning rule and the risk quantitative evaluation index.

The invention provides a CLAHE image enhancement optimization method and system based on an FPGA and a medium. The method comprises the steps that firstly, an input image is divided into a plurality of histogram sub-regions; on the basis of histogram statistics of each sub-region, the histograms of the sub-regions are cut, pixels exceeding a threshold value are redistributed according to a given algorithm, and the histograms are corrected; mapping is carried out through CDF operation to obtain an equalized gray level, and then gray stretching is carried out to generate a new gray mapping result; and finally, reading the mapping gray level stored in the previous frame, processing the gray value of the pixel point through combined operation of bilinear interpolation, completing block effect elimination between the sub-regions, and outputting an image after local contrast enhancement. According to the method provided by the invention, the controllability of the overall brightness level of the image can be ensured on the basis of ensuring the enhancement effect of the video image. And meanwhile, the method is realized on an FPGA platform, so that the processing speed of image enhancement is ensured, and the dual requirements on the processing speed and quality can be met.

The invention relates to the technical field of urban rail transit, in particular to an urban rail transit passenger monitoring data processing and analyzing system and method, and the system comprises a data collection layer which is used for collecting video streams, gate passing records, passenger positioning data and environment parameters of temperature, humidity and illumination in real time; the spatio-temporal feature fusion layer is used for converting the video image data into analyzable feature vectors and integrating card swiping records and position information to form spatio-temporal trajectory data of passengers; the three-level anomaly detection layer comprises an individual layer detection unit, a group layer detection unit and a system layer prediction unit; the dynamic decision-making layer is used for calculating an abnormal score based on a dynamic threshold value, and dynamically adjusting a behavior coefficient according to a historical disposal effect through a PPO reinforcement learning algorithm; and the execution layer comprises an edge computing node and a cloud analysis platform. Therefore, the problems of single data acquisition, lack of comprehensive anomaly detection means, fixed and lagged decision response, ineffective utilization of resources and the like in the prior art are solved.

A vehicular driver monitoring system includes a vehicular interior rearview mirror assembly having a mirror head that accommodates a mirror reflective element. A video display is disposed behind the mirror reflective element and operable to display video images captured by a rearward viewing camera of the vehicle. A driver monitoring camera and a near infrared light emitter are accommodated by and move in tandem with the mirror head. The near infrared light emitter is accommodated within the mirror head so that, with the mirror head adjusted relative to the mounting base to set the rearward view of the driver of the vehicle, a beam of near infrared light emitted by the near infrared light emitter is directed toward a driver's region of the vehicle. The driver monitoring camera is disposed adjacent to the video display screen so as to not view through the video display screen.

The invention discloses a face posture recognition method, an anti-dazzle lamp regulation and control method, a lamp, equipment and a medium, and relates to the technical field of lamp control. According to the method, firstly, face region recognition is performed based on the integral image corresponding to the video image, then the face image is intercepted from the video image according to the recognition result, and then face pose recognition is performed based on the trained CNN model and the trained backbone network model, so that recognition of each face pose in the target scene is realized, and the recognition efficiency is improved. And then the irradiation angle and brightness of the lamp are adjusted based on the recognized face posture, and adaptive adjustment of the glare problem is achieved.

The embodiment of the invention provides a multi-mode unmanned material checking method based on combination of RFID and machine vision, the method is applied to a mobile module and a camera module, an RFID read-write module is arranged on a mobile trolley, and the method comprises the following steps: in response to movement of the mobile module, the RFID read-write module scans a target material to obtain RFID identification data, and filters and de-duplicates the RFID identification data to obtain a target material; an RFID identification result is determined; dividing a shooting range of the camera module based on a scanning area of the RFID read-write module, and detecting a material box and a material surface sheet label in video image data when the video image data is shot; and identifying the material surface sheet label to obtain material data information, and matching the material data information with the RFID identification result to obtain a multi-modal material matching result.

The invention discloses a video anti-shake method and system based on multi-scale fusion and adaptive smoothing, relates to the technical field of video image processing, and aims to effectively solve the image quality problem caused by shake in a video shooting process. Gradient histograms and wavelet energy distribution characteristics of video frames are extracted through graying and normalization processing, and the gradient histograms and the wavelet energy distribution characteristics are input into a jitter type recognition network to recognize translation, rotation and Z-axis jitter probabilities. And further extracting motion, frequency domain and edge features, and generating multi-modal coupling features through combination of a dynamic feature interaction network and a dot product attention mechanism. And constructing a motion trajectory by using the features, optimizing the trajectory by using a texture perception double-layer smoothing strategy, introducing an adaptive penalty term into a dynamic planning cost function, and outputting a smooth motion compensation parameter. And finally, processing the boundary region through motion compensation and image extrapolation to generate an anti-shake video frame. Through multi-scale feature fusion and a self-adaptive smoothing strategy, the video anti-shake effect is effectively improved, and the method is suitable for complex scenes.

The invention discloses a character loss compensation processing device and method for video subtitle extraction, which effectively solve the problem of character loss caused by video image quality, complex background interference, OCR (Optical Character Recognition) limitation, dynamic subtitle change and the like. According to the method, key technologies such as self-adaptive subtitle denoising, semantic compensation and multi-frame fusion are adopted, and the completeness and accuracy of subtitle extraction are remarkably improved. The scheme is suitable for various types of movies, television dramas, short videos, conference videos and the like. The core innovation of the method is that lost or wrong characters can be effectively detected, compensated and corrected by intelligently analyzing the context and integrating multi-frame information, so that the subtitle output quality is greatly improved, and the effect is remarkable particularly in a scene that a single character is easy to lose.

The invention discloses a video geographic positioning method and device based on iron tower monitoring, and the method comprises the steps: constructing a monitoring video homonymy point sample library, collecting the homonymy point data of a monitoring video and a remote sensing orthoimage, and storing a reference image; establishing an initial calibration mapping relation, and realizing coordinate transformation of the remote sensing image and the video image through projection transformation; matching a reference image closest to the real-time video frame from a reference image library through image retrieval; dense matching of inclined images is carried out by using a deep learning model, and automatic alignment of feature points is realized; and resolving a video image coordinate to a geographic coordinate through homography transformation and an initial calibration model to complete high-precision positioning. According to the method, the bidirectional conversion function of finding the land by the video and finding the video by the land is supported, the application efficiency of iron tower monitoring in the fields of cultivated land protection and the like is effectively improved, a good positioning effect is still achieved for complex scenes such as plains, mountainous regions and the like, and corresponding technical support is provided for wide application of iron tower video monitoring.