164 results about "Image geometry" patented technology

The invention relates to the technical field of medical image processing, in particular to a medical equipment optical system calibration method and system based on image processing, and the method comprises the following steps: capturing a calibration image in medical equipment, analyzing the boundary structure of an optical calibration grid, screening the pixel distribution characteristics of a segmentation boundary, and extracting the geometric description characteristics of the calibration image; and generating a geometric distortion distribution diagram. According to the invention, by capturing the calibration image and analyzing the boundary of the optical calibration grid, the geometric distortion area is accurately identified, the accuracy of the distortion compensation image is improved, the non-linear distortion is compensated in a partition manner, the image quality of each area is optimized, the continuity of the image boundary is improved, and the structural symmetry and curvature change area are optimized. The distortion of the detail area is restored, the edge details are smoothed, the image quality is improved, the amplitude distortion of the optical calibration grid is effectively controlled, the edge detail distortion is reduced, the image integrity and consistency are ensured, and the image precision and the reliability of an optical system are remarkably improved.

The invention relates to the technical field of power equipment state monitoring, and particularly discloses a transformer fault diagnosis method and system based on image recognition, and the method comprises the steps: collecting a transformer multi-mode image sequence in real time, and carrying out the definition and part integrity evaluation and screening to form an initial image set; performing multi-scale space registration on the initial image set and a transformer normal state standard template to generate a reference image, and reversely deriving a displacement vector field based on pixel-level difference; carrying out smooth optimization and geometric reconstruction on the displacement vector field under the geometric constraint of the transformer structure, and generating a correction image with a real structure; fault feature enhancement is carried out in a gradient domain of the corrected image, a fault area is identified through matching of multichannel feature extraction and a transformer typical fault feature library, and a diagnosis report integrating fault types, confidence coefficients and geometric parameters is generated; according to the method, the problem of image geometric deformation caused by shooting condition differences is effectively solved, and the accuracy and reliability of fault identification are improved.

The invention discloses a crack image processing method based on a neural network and plane geometry image transformation, and particularly relates to the field of general image data processing driven by plane geometry image transformation and neural network identification. Comprising the steps that an image collection platform projects a structured image coding pattern to the surface of a to-be-measured target through a preset image coding pattern template, and synchronously collects an image frame containing an image mark so as to construct a space anchor point used for image geometric reference in an image plane; view angle parameter change information in an image acquisition process is acquired through an image acquisition platform, and projection configuration parameters of an image coding pattern are combined. By constructing an image offset modeling mechanism jointly driven by an image coding pattern and a space anchor point, image in-plane correction of geometric distortion features in an image frame is realized, and structural consistency and texture continuity before subsequent image analysis are improved.

The invention discloses a reservoir area hydro-fluctuation belt rock crack image recognition method based on a deep learning network, and the method comprises the steps: S1, constructing a rock crack segmentation model, and training the rock crack segmentation model through employing a pre-processed rock crack image sample set; s2, recognizing a rock crack region in the target image by using a rock crack segmentation model to obtain a first type of recognition result; s3, on the basis of coordinate positions, pixel values and color space attributes of pixel points in the image, optimizing the first type of recognition result through a crack mask optimization algorithm of multi-algorithm weighted fusion to obtain a second type of recognition result; and S4, aiming at the second type of identification result, measuring the width of the rock crack through a maximum inscribed circle method and an orthogonal projection method so as to obtain geometric representation information of the rock crack. According to the invention, automatic identification of the geometric representation information of the rock crack image of the hydro-fluctuation belt can be realized, and the intelligent and informatization level of reservoir area hydro-fluctuation belt bank slope rock crack detection is improved.

The invention discloses an image acquisition and data processing system for a vertical take-off and landing unmanned aerial vehicle, and relates to the technical field of data processing, and the system comprises a region division unit, an image acquisition unit, an image preprocessing unit, an image feature matching unit and an image geometric transformation unit. The method comprises the following steps that: a plurality of sub-regions are divided into a plurality of sub-regions, each sub-region is numbered, the occurrence of region shooting omission in the acquisition process of an unmanned aerial vehicle is prevented, and meanwhile, an image preprocessing unit performs distortion correction on acquired image information, so that pixel points at offset positions in an image can be adjusted to actual positions, errors existing in the image information are reduced, and the accuracy of the image information is improved. The adjacent image information is smoothly spliced through the image feature matching unit and the image geometric transformation unit, and the existing feature matching points are adjusted by using the genetic algorithm and the RANSAC algorithm in the image feature matching unit, so that the mismatching points can be found and deleted in time.

The invention relates to the technical field of image processing, and discloses a fluorescent lifetime microscopic image large-view-field splicing method and system and electronic equipment, and the method comprises the steps: constructing a tissue region mask for a to-be-spliced image, and constructing a vignetting model in a tissue region to complete vignetting correction; counting brightness indexes in an organization area to determine a global brightness reference, and realizing image group brightness unification through global zooming and single image brightness adaptive correction; further realizing alignment of adjacent images through feature point matching, and finishing image block splicing in combination with a minimum color difference suture fusion method to obtain an image band; an overlapping area is extracted from an image belt to generate an effective content mask, relative displacement is estimated by using a phase correlation method after low-frequency suppression, image belt splicing is completed by multiplexing a minimum color difference suture fusion method, a large-view-field spliced image is output, the automation degree and robustness of splicing are improved, the spliced image is geometrically consistent and visually seamless, and the splicing efficiency is improved. And the requirements of medical research on high-resolution and large-field-of-view fluorescence lifetime microscopic images are met.

The invention relates to a GIS (Gas Insulated Switchgear) equipment mechanical defect diagnosis method based on a Gramb angle field and a dual-channel PCNN-Attention neural network, and belongs to the technical field of gas insulated switchgear mechanical vibration defect diagnosis. The method solves the problems that traditional diagnosis depends on artificial feature extraction, so that subjectivity is high, information mining is insufficient, and defect severity evaluation is missing. According to the technical scheme, the method comprises the steps that a one-dimensional vibration signal is converted into a GASF two-dimensional image and a GADF two-dimensional image through a GASF field so as to completely reserve time sequence topological features; carrying out data enhancement by adopting an image geometric transformation technology so as to improve the generalization ability of the model; and a dual-channel PCNN-Attention model is constructed, and synchronous intelligent identification of defect types and severity is realized through parallel feature extraction and dynamic weight optimization of an attention mechanism. According to the method, the diagnosis accuracy, reliability and adaptive capacity are improved, and support is provided for equipment state operation and maintenance.

The invention relates to the technical field of glass bottle detection, and discloses a glass bottle bottom defect detection method which comprises the following steps: building a detection platform, and adopting a multi-view imaging module of an annular LED and oblique light supplement combined light source, a vertical camera and 3-4 oblique cameras; after the to-be-detected glass bottle is positioned, synchronously collecting a bottle bottom full-view image set; graying, adaptive median filtering, CLAHE histogram equalization and Otsu binarization preprocessing are carried out on the image; extracting geometric and textural features of the image, inputting the geometric and textural features into an improved YOLOv5 model for identification and classification, and judging defects by combining with multi-view result fusion; and sorting the glass bottles according to a detection result and generating a traceable report. The identification rate of tiny defects is larger than or equal to 99%, the detection time of a single bottle is smaller than or equal to 0.5 second, and the method is compatible with glass bottles with the diameter of 30-100 mm, is suitable for large-scale quality control in the fields of food and beverage, medicine packaging and the like, and has extremely high practical value.

The invention relates to the field of image recognition, discloses an image recognition method and an image recognition system, and systematically solves the core pain point of traditional document recognition through optical-topology fusion processing and a dynamic resource allocation mechanism. The image recognition method is composed of an acquisition module, a grid module, a phase module, a setting module and a distribution module, pixel brightness is calculated and coded based on document RGB data, and a two-dimensional coding matrix is generated; constructing a geometric correction grid, forming a nonlinear constraint field, and enhancing the anti-deformation capability of the image; generating spiral phase light waves in the constraint field by using a spatial light modulator, and generating a time-varying phase map; determining a local topology index by detecting the number of phase jump times; and extracting the closed boundary region as a character block, and generating an analysis result file. The system breaks through traditional limitation, improves image geometric correction precision, feature extraction sensitivity and boundary judgment accuracy, efficiently completes document analysis, and is suitable for scenes such as document digitization and information retrieval.

The invention discloses a lawn coverage evaluation platform based on image recognition, which belongs to the technical field of lawn measurement and comprises an image acquisition module, an image preprocessing module, a lawn area recognition module, a coverage calculation module, a scoring and analysis module and a result output module which are connected in sequence. The method comprises the following steps: performing standardized image acquisition on a lawn through an image acquisition module, and preprocessing an image to realize geometric standardization of the image; an artificial intelligence image segmentation model is adopted to improve the accuracy and adaptability of lawn area identification; the sample areas are scored and analyzed by calculating the coverage, and the method is used for comparative analysis of multiple varieties. By adopting the method, the acquisition efficiency, the identification accuracy and the evaluation standardization level of the coverage data in the lawn grass seed regional test can be improved, and the problems of non-uniform acquisition conditions, fragmentation of an identification method, incomparable results and the like in the prior art are solved.

The invention relates to the technical field of industrial visual inspection, and discloses an asynchronous exposure double-view-field imaging system and method, and the system can achieve the merging of independent view field images collected by two parallel lenses into a single linear array camera through a reflector and a non-polarization beam splitting cube based on the asynchronous exposure double-view-field imaging system. Meanwhile, in cooperation with precise time sequence control, the system drives two independent stroboscopic light sources to be synchronous with line trigger signals of the camera, different view fields are alternately illuminated in odd-even line periods, the camera is made to perform asynchronous exposure, and two paths of image information are collected in a time-sharing mode; and finally, reconstructing two complete, independent and seamlessly spliced detection images from the interleaved data streams through an odd-even line separation algorithm. According to the invention, with the hardware cost of a single camera, the detection visual field is obviously expanded, the detection efficiency and the image geometric precision under the high-line scanning frequency are ensured, and the excellent balance of the performance, the cost and the efficiency is realized.

The invention discloses a target positioning method and system based on target height correction monocular distance measurement, and the method comprises the steps: carrying out the calibration of a ship-borne camera through a data set, employing a YOLO target detection model, carrying out the detection of video data collected by the ship-borne camera, and obtaining the bounding box information of a target; and secondly, according to bounding box information, calculating a yaw angle of the target relative to a main point of the shipborne camera, and obtaining an azimuth angle by combining an orientation angle of the main point of the shipborne camera. And then height correction is carried out to measure the distance between the target ship and the shipborne camera. And finally, according to the azimuth angle and the distance, combining the latitude and longitude coordinates of the shipborne camera to obtain the actual geographic latitude and longitude coordinates of the target ship. According to the method, the goodness of fit between distance measurement and actual image geometry is improved, the actual object distance is reflected more accurately, and accurate and efficient target positioning is carried out.

The invention discloses an end-to-end SLAM (Simultaneous Localization and Mapping) method based on a deep cyclic convolutional neural network. According to the method, an image sequence frame is collected in real time through a depth camera, converted into a set image size and then output; creating a convolutional neural network model to receive and process an image sequence frame output by the depth camera, and outputting extracted feature point cloud coordinates and descriptor information and camera pose information for establishing topological information of an image geometric structure; creating a recurrent neural network model to receive and process an image sequence frame output by the depth camera, and outputting a credible semantic region and object-level road sign information of which the credibility is greater than a set first threshold value; according to the method, the three-dimensional map with semantic recognition and perception is constructed based on the deep learning model, and the parameter indexes such as the positioning precision, the map construction precision, the loopback detection precision, the system efficiency and the robustness are remarkably improved.

The invention discloses a multi-scale feature fusion-based atmospheric turbulence video correction method and system, and the method comprises the steps: extracting shallow texture, middle structure and high semantic information through a multi-scale feature pyramid, and carrying out the channel weighting processing of different frequency components through the combination of frequency adaptive enhancement and a deformable attention alignment mechanism; adaptive optimization is realized for turbulence sensitivity difference, feature level alignment is completed under the guidance of optical flow, tilt distortion is eliminated, and reference features are dynamically maintained and turbulence noise accumulation is inhibited by adopting a gating reference updating module. In addition, cross-frame space-time fusion is completed through a multi-head time channel self-attention mechanism, time consistency is improved, and image geometry and high-frequency details are gradually recovered by using a two-stage decoder. According to the method, the definition of the turbulence degraded image can be remarkably improved, and technical support is provided for high-precision imaging.

The invention discloses a distorted image correction method based on a fast multi-scale residual network. Relates to the field of image geometric distortion correction, in particular to the technical field of distortion image correction based on a fast multi-scale residual network. Comprising the following steps: S1, constructing a data set, and generating a paired data set of a distorted image and a standard image and a deformation displacement field file of a corresponding number of the distorted image by using a geometric distortion model; s2, a distorted image correction model based on a fast multi-scale residual network is constructed, the model is divided into a backbone network and a tributary network, the backbone network is used for recovering image features, and the tributary network is used for outputting image classification features; s3, training the distorted image correction model based on the fast multi-scale residual network by using the pairwise data set of the distorted image and the standard image and the deformation displacement field file; and S4, processing the to-be-corrected image by using the trained distortion image correction model based on the fast multi-scale residual network to obtain the corrected image and the corresponding classification thereof.

The invention discloses a short-focus projection image correction optimization system based on artificial intelligence, and the system comprises an image collection and structure modeling module which is used for generating a structure perception feature map; the structure guide matching module is used for constructing a stable matching subset and executing matching point screening; the dense matching and geometric correction module is used for outputting a coarse correction image; the structure residual error fine tuning module is used for constructing a structure residual error image and inputting the structure residual error image into the local fine tuning sub-network, guiding the model to carry out fine-grained correction on a structure error region, and generating a final correction image; the structure consistency evaluation module is used for evaluating key point deviation, edge direction similarity and geometric alignment degree; the correction convergence judgment module is used for tracking the correction score change; and the correction result output module is used for outputting a final correction image. The method has the advantages of high precision, strong robustness and iterative optimization, and is suitable for image geometric alignment and enhancement requirements in complex optical projection scenes such as smart classrooms, interactive display, immersive projection and the like.

The invention relates to the technical field of automobile detection, in particular to a wheel base difference vision measurement method based on a rim circle common tangent, which comprises the following steps: S1, image acquisition: acquiring projection images of two rims on one side of a vehicle through a monocular camera fixed on a bracket; s2, determining internal parameters of the camera and the circle center of the rim: carrying out edge detection and ellipse fitting on the rim image, and calculating the internal parameters of the camera by combining a circular point and absolute conic constraint; s3, determining external parameters of the camera: resolving a rotation matrix and a translation vector of the camera based on the geometrical relationship between the inner and outer common tangent of the ellipse and the image; s4, solving a wheelbase difference: obtaining the positions of the circle centers of the left and right rims in the world coordinate system, and calculating the spatial distance difference as the wheelbase difference; through image processing and geometric modeling, the method is low in cost and suitable for accurate measurement of the wheelbase difference of multiple vehicle types, and has the advantages of being easy and convenient to operate, small in error and low in equipment requirement.

The invention belongs to the technical field of wireless communication, and particularly relates to a machine vision integrated communication enhancement integration method and system, a communication part and a machine vision part are equipped, and target equipment can be identified by the communication part and the machine vision part; when the machine vision-fused communication enhancement integrated system deals with massive equipment access, accurate identification and management of equipment are realized through a multi-mode multi-level identification system, a dynamic association algorithm and a collaborative management mechanism; the equipment identified by the communication part is matched with the equipment identified by the machine vision part to ensure that the target equipment identified by the communication part and the machine vision part is the same in the processing flow; perceptual data obtained by the communication part and including parameters such as distance, position / angle, speed and the like of the target equipment are fused with perceptual data obtained by the machine vision part and including parameters such as images, geometric features, position / angle, speed, distance and the like of the target equipment.

The invention discloses an imaging geometry online calibration method, which comprises the following steps that: a marker motif is scanned by a high-degree-of-freedom CT (Computed Tomography) to obtain a first group of projection drawings, and the marker motif comprises an upper half part and a bottom part; processing the projection data based on the first group of projection drawings to obtain initial geometric information; removing the upper half part of the marker motif, and scanning the bottom part of the marker motif and the measured object by the high-degree-of-freedom CT to obtain a second group of projection drawings; correcting the initial geometric information based on a global optimization algorithm and the second group of projection drawings to obtain final optimized geometric information; three-dimensional reconstruction is carried out based on final optimized geometric information, the marker motif device used in the method occupies little imaging view and is located at the bottom of the view, imaging of a measured object is not interfered, and the engineering applicability is high.

The invention discloses a correlated region feature assisted binocular vision depth estimation method, which comprises the following steps of: 1, matching geometric shapes of object images in left and right views through barycentric coordinates and rotational inertia of the geometric shapes of the object images; 2, through intersection of a key point connecting line on the geometric shape contour of the image and an epipolar line, obtaining a search range of a minimum scale matched with imaging points on the epipolar line of the left and right views; 3, constructing geometric figures of imaging points and key points on the geometric shape contour of the image, calculating the rotational inertia of constructing the geometric figures, matching the imaging points in the left and right views, and calculating the parallax and depth of the matched binocular imaging points; according to the invention, through subdividing the geometric shape area of the image, the imaging point matching range of parallax calculation is reduced, and the binocular matching efficiency is improved; according to the method, the associated region features are used for assisting parallax calculation, and missing of matching feature information of poor regions such as shielding, weak texture and exposure is made up, so that the parallax calculation accuracy is enhanced, and the binocular stereo matching precision is improved.

The application discloses a method for intelligent component identification and segmentation of an aircraft CAD model based on three-mode fusion, comprising: obtaining an aircraft CAD model with a boundary representation structure; constructing a multi-modal training data set that fuses surface images, geometric parameters and component semantic labels; training a multi-modal identification model; using the model for preliminary identification; for low confidence or small area surfaces, starting a geometric rule auxiliary module based on geometric topology and parameterization rules for semantic completion and correction to form a final segmentation result; automatically extracting key geometric feature lines; and converting and outputting into a structured engineering semantic format that can be used by grid generation software. The application realizes full automation of component identification, solves the problems of low efficiency and consistency caused by manual dependence, improves the identification integrity and engineering practicability of complex geometry, and the output result can directly drive subsequent grid generation, thereby significantly improving the automation level of the digital simulation pre-processing process.

The application provides a task-driven defogging perception method and system, and belongs to the technical field of cooperative perception of Internet of Vehicles. The method comprises the following steps: collecting a first foggy image stream and a second foggy image stream; obtaining a first processed image by intercepting and processing the first foggy image stream; obtaining a second processed image by intercepting and processing the second foggy image stream; respectively performing geometric alignment on the first processed image and the second processed image to obtain a first aligned image and a second aligned image; performing defogging and image enhancement processing on the first aligned image and extracting a first feature; performing defogging and image enhancement processing on the second aligned image and extracting a second feature; respectively repairing the first feature and the second feature by using a diffusion model; fusing the repaired first feature and the second feature, and determining a final perception result based on the fused feature. The application improves the robustness of automatic driving and balances the low power consumption of AI intelligent glasses and the high performance computing demand of a vehicle-mounted system.

The invention relates to the field of remote sensing image processing and terrain correction, and particularly discloses a radar geometric field-driven SAR distortion information adaptive neighborhood propagation identification method, which comprises the following steps of: S1, acquiring a Golden digital elevation model of a research area and a pixel-by-pixel ellipsoid incident angle graph generated by SAR satellite precision orbit parameters and imaging geometry, and carrying out geometric unified processing on the Golden digital elevation model and the pixel-by-pixel ellipsoid incident angle graph; s2, calculating a terrain parameter, an effective slope angle, an R geometric response field and an R field directivity gradient based on the unified raster data; s3, main distortion region classification is carried out according to the effective slope angle and the pixel-by-pixel ellipsoid incident angle; s4, defining an adjacent gradient and constructing a self-adaptive threshold value and a shadow self-adaptive threshold value; and S5, performing passive region propagation from the boundary of the main distortion region, and generating a final classification grid map after space overlapping correction. According to the method, the defect of blind propagation of the existing method is overcome, and high-precision automatic identification of SAR geometric distortion under complex terrains is realized.

The application discloses a kind of underwater concrete crack width automation measurement system and method based on double laser, it is related to underwater structure health monitoring and optical measurement technical field.The application includes: double laser projection module, for projecting a pair of circular laser points with fixed physical spacing to underwater concrete surface to be measured as physical scale reference;Image acquisition module, including underwater camera, for collecting image containing laser point and crack area;Data processing module, for performing underwater camera calibration, laser point detection, crack image segmentation and crack physical width calculation.The method corrects the image geometric distortion caused by "water-glass-air" refraction by establishing an accurate underwater multi-interface refraction imaging model;Nonlinear mapping relationship from image pixels to physical space is established using double laser points;Based on unsupervised crack segmentation model, crack area is automatically segmented, the physical width of each position of crack is calculated and width distribution cloud chart is generated.

The invention discloses a remote sensing image classification method based on an improved Markov random field, and belongs to the technical field of remote sensing image object classification. The invention comprises the following steps: step one: determining the classification number of remote sensing images, and randomly classifying the remote sensing images; step two: based on the classical Markov random field model, considering the relationship between pixels in the observation field and the distance relationship between each pixel and the central pixel in the spatial neighborhood potential group, improving the prior energy function formula in the classical MRF algorithm; step three: based on the classical Markov random field model, considering the relationship between adjacent elements in the observation field, improving the likelihood energy function formula in the classical MRF algorithm; step four: using the improved prior energy and likelihood energy function formulas in step two and step three, solving the label field corresponding to the minimum global energy, and updating the classification label of the central pixel according to the solution result; step five: judging whether the pixel value mean of the classification label updated in step four converges.

This invention relates to a method for stitching and adjusting the grayscale of multispectral images using a filter array, belonging to the field of aerial image processing technology. The purpose of this invention is to utilize the SURF algorithm to register projected images, using the mean of the grayscale average ratio of overlapping regions in single-band images as the grayscale adjustment coefficient, and adjusting the grayscale of the registered images sequentially based on the grayscale of a reference image. This invention comprises three steps: determining image geometric relationships, adjusting image grayscale, and stitching images. This invention solves the problem of grayscale differences in single-band multispectral stitched images and minimizes the loss of spectral information of ground features. It not only effectively solves the problem of inconsistent grayscale of ground features in images but also maximizes the preservation of spectral information characteristics of ground features in different bands.

The invention relates to an urban green land health situation monitoring method and device based on a mobile phone camera visible light image. The method comprises the steps that the visible light image of an urban green land canopy is collected through a mobile phone camera; distortion correction and vertical calibration are carried out on the acquired image; compressing and re-sampling the image; noise reduction processing is performed on the image, and image details are reserved; converting the image from an RGB color space to an HSI color space, and eliminating the influence of an illumination condition on the image; calculating an EXG index, a leaf area index and a vegetation coverage; performing pseudo-color enhancement display on the vegetation index result; and outputting a monitoring result. According to the invention, through the square preview frame and the gyroscope vertical calibration system, the geometric error of the image is significantly reduced, and the area measurement and calculation precision is improved; matrix parallel operation and self-adaptive compression factors are adopted, so that the computing power bottleneck of a mobile terminal is broken through, and the image processing efficiency and the equipment compatibility are remarkably improved; through the whole-process anti-interference design, stable monitoring in a field complex scene is realized.

To reduce uncomfortable feeling given to a user when switching between display of a part of an image and display of a part of an image obtained by performing distortion correction of the image.SOLUTION: An electronic device includes: acquisition means which acquires a first image and a second image which is an image obtained by subjecting the first image to geometric conversion processing for distortion reduction; switch means which switches a display mode between a first mode which displays a range of a part of the first image on a display region and a second mode which displays a range of a part of the second image on the display region; and adjustment means which, when the display mode is switched, adjusts a display position of an image to be displayed in the display region after the display mode is switched so that an object which appears at a first position of the display region before the display mode is switched appears at the first position after the display mode is switched.SELECTED DRAWING: Figure 7

The application is suitable for the technical field of power equipment state monitoring, and provides a flow blocking valve state identification method, system, device and program product. The method comprises: acquiring a reference image and a to-be-inspected image of a flow blocking valve; identifying first structure key points from the reference image, and identifying second structure key points from the to-be-inspected image; performing geometric correction on the to-be-inspected image based on the first structure key points and the second structure key points, to obtain a corrected image consistent with the reference image in geometric space; calculating a reference angle in the reference image and a current angle in the corrected image respectively, and determining the state of the flow blocking valve according to the angle difference between the current angle and the reference angle. The accuracy and universality of flow blocking valve state identification are improved.

The application discloses an image processing method and an electronic device, and relates to the technical field of image processing, wherein the method comprises the following steps: extracting a first geometric edge line set of an initial depth image according to the initial depth image corresponding to a target monocular image; screening geometric edge lines with gradient scores greater than a target score threshold from a depth gradient image generated according to the initial depth image and an initial edge line set of the target monocular image to determine a second geometric edge line set; and fusing the first geometric edge line set and the second geometric edge line set to determine image geometric edge lines of the target monocular image and construct a final geometric edge line set. The method solves the problems of serious texture interference, insufficient utilization of depth information, difficulty in balancing the number and purity of edge lines, and insufficient real-time performance in the prior art, and achieves the technical effects of filtering texture noise while ensuring the number and robustness of edge lines and meeting the needs of multi-sensor fusion tasks.