577 results about "Perspective transformation" patented technology

The invention discloses a self-adaptive three-dimensional scene reconstruction method and system based on a single panorama, and belongs to the field of computer vision processing. The method comprises the following steps: firstly, generating a depth map through an indoor scene panorama and constructing an initial three-dimensional grid; then generating a multi-view image and a shielding mask thereof based on view conversion, forming a training sample pair for fine tuning of the diffusion completion model, and injecting scene prior information for the model; then extracting a grid boundary contour and calculating a central axis, and adaptively constructing a camera pose set; after the integrity of the grid is optimized through iteration completion, the grid is converted into a 3D Gaussian sputtering field, and Gaussian point parameters are optimized through multi-view data; in the optimization process, an up-sampling strategy of rendering error feedback and edge Gaussian points is adopted, and finally a scene model with complete geometric textures is output. According to the method, a high-quality three-dimensional scene is reconstructed from a single panorama through a structure self-adaptive completion and Gaussian refining technology, and the method is particularly suitable for immersive roaming reconstruction of indoor scenes.

The invention discloses a control method for an underwater cleaning robot, and the scheme can comprise the steps: S1, collecting an environment image of a to-be-cleaned region through a global camera, and carrying out the distortion correction and perspective transformation preprocessing, and obtaining a preprocessed image; s2, generating a static environment map according to the preprocessed image, wherein the static environment map comprises obstacles and dirt sub-regions; s3, adopting a multi-scale path planning strategy: firstly, dividing a map by using a first-scale grid, determining a cleaning end point, and planning an access sequence based on an A * algorithm; setting the first dirt sub-region as a target, and planning a path to a boundary turning point of the first dirt sub-region; dividing a map by using a second scale grid, and planning a path from a turning point to a cleaning end point; and updating the operation starting point and removing the processed area, and circulating until the operation path sequence is generated. And S4, the underwater cleaning robot completes cleaning operation on all the dirt sub-areas according to the planned path.

The invention discloses a method for a mowing robot to move along a lawn boundary based on a local grid map, and relates to the field of mowing robot edge navigation movement, and the method comprises the steps: recognizing a target region and an obstacle region through a semantic segmentation algorithm, constructing a dynamically updated local grid map through the combination of improved inverse perspective transformation and depth point cloud information, and carrying out the navigation of the target region and the obstacle region; no physical boundary dependence is realized, the deployment cost is reduced, and the real-time processing efficiency of acquired images can be improved; according to the local grid map, the edge navigation deflection angle is calculated through lawn boundary feature extraction in combination with the position of the robot, precise navigation of the robot along the lawn boundary is achieved, rapid response of dynamic obstacles is supported, and the edge movement precision and operation efficiency of the mowing robot are remarkably improved. The problems that an existing mowing robot is low in edge movement precision and poor in operation efficiency due to the fact that an edge track cannot be decided in real time and calculation efficiency cannot be considered at the same time are solved.

The invention provides an evidence obtaining method and system based on image processing, and the method comprises the following steps: S1, generating a pixel-level depth-of-field distribution diagram of an input image through a multi-scale encoder-decoder network, employing an edge perception optimization layer in a decoding stage, and improving the depth-of-field boundary precision through minimizing a local gradient consistency loss function; s2, performing depth-of-field rationality verification based on an optical imaging physical model, and triggering a first-level tampering alarm by calculating a defocusing fuzzy radius and a gradient direction of a selected region when a difference between the defocusing gradient directions of a target region and a background region exceeds a preset threshold value; and S3, dynamically positioning a key pixel region, identifying a depth-of-field mutation boundary by using an edge detector, calculating by combining local texture complexity, screening a pixel set of which the entropy value is higher than a threshold value and which is located at the mutation boundary, correlating metadata to verify the rationality of the physical size and the spatial position of an object, and eliminating false detection caused by perspective transformation.

The invention discloses a heading equipment cutting head position calibration method based on image acquisition, which comprises the following steps: S1, acquiring image data and preprocessing the image data; s2, extracting cutting head feature points, and constructing a two-dimensional feature point set; s3, calculating a three-dimensional coordinate by using perspective transformation, and generating a point cloud set; s4, performing rigid registration on the point cloud set and a standard structure model of the tunneling equipment, and iteratively solving a rotation matrix and a translation vector to obtain pose parameters; s5, simulating change tracks of the feature points in continuous frames, constructing a time consistency constraint, and optimizing pose parameters; s6, the optimized three-dimensional coordinates are converted to a reference coordinate system, space alignment is completed, and a position calibration result is generated; and S7, outputting a calibration result for control instruction generation and path planning. According to the invention, high-precision and non-contact three-dimensional position calibration of the cutting head of the tunneling equipment is realized, and operation automation and control precision are improved.

The invention discloses an inclined pointer instrument accurate reading identification method based on a cascade model. The problem that a traditional method is low in reading accuracy under the scenes of complex backgrounds, inclined shooting and the like is mainly solved. The method comprises the following steps: constructing a cascade model: positioning an instrument area through an instrument detection model, cutting the instrument area, analyzing scale line and pointer features through an instrument feature extraction model, and extracting a dial center, a main scale line and a pointer position in combination with key point detection and a semantic segmentation technology; inclination correction optimization: calculating a perspective transformation matrix according to the main scale line endpoint coordinates, performing rotation correction on the inclination instrument image, and eliminating a visual angle inclination error; and angle method reading calculation: based on the corrected image, realizing accurate reading by combining the angle relation between the pointer and the adjacent scale lines with OCR identification scale values. According to the method, the recognition accuracy and robustness in a complex scene are remarkably improved, computing resources are reduced, and the recognition precision of instrument reading is improved.

Three-dimensional Gaussian splatting mechanisms that initialize a set of 3D Gaussian distributions, un-project pixels from two-dimensional (2D) planes to 3D space by applying queries to the 3D Gaussians at expected un-projected ray depth positions, and splat the 3D Gaussian distributions on the 2D planes based on the expected un-projected ray depth positions.

The invention discloses a multi-view construction personnel tracking method and system based on attention perception. The method comprises the following steps: giving synchronous images from S cameras, and inputting the synchronous images into an encoder for feature extraction to obtain a multi-view feature map; transforming the multi-view feature map into a unified aerial view space by using perspective projection, and aggregating features after projection transformation of all views by using a convolutional layer; inputting the aggregated aerial view features into a decoder for decoding; a cross attention module is introduced, bird's-eye view features of a current frame and an adjacent frame are processed through 3D position coding, instance tokens are extracted as query, keys and values, an affinity matrix is generated through CNN coding similarity, features are propagated through matrix multiplication, and the bird's-eye view features of the current frame are updated. According to the method, the multi-view feature map is projected to the aerial view to realize early fusion, and a cross-frame attention mechanism is introduced, so that the problem of appearance feature distortion caused by perspective transformation is solved.

The invention provides a core photograph automatic correction and cutting method based on YOLOv8-seg and perspective transformation, and belongs to the technical field of image processing, and the method comprises the steps: S1, collecting a core box image; s2, polygonal point sequence labeling; s3, obtaining an optimal segmentation model; and S4, cutting. Aiming at the problems of much manual intervention, low efficiency and poor adaptability in the existing core photograph processing, the invention provides a set of full-automatic end-to-end solution without preprocessing, realizes rapid and stable correction and cutting of the core photograph in a complex scene, relieves the mechanical labor burden of geological personnel, focuses on the core exploration work, and improves the core exploration efficiency. And the practical application requirements of large-scale geological exploration projects are met.

The embodiment of the invention discloses a video generation method and device, electronic equipment, a computer readable storage medium and a computer program product, and the method comprises the steps: generating a rough three-dimensional world of a target scene formed by a three-dimensional Gaussian ball set, and automatically searching a multi-angle offset view angle at the same coordinate position as a new view angle. According to the method, a multi-view-angle combined repairing mechanism is realized, geometric and semantic consistency of image data among different view angles is ensured, consistency of scene structures, textures and semantic information at different observation angles is realized, and a continuous, coherent and high-quality three-dimensional world is constructed. According to the method, the problem of multi-view inconsistency in a traditional method is solved, the new view image quality under large view conversion is improved, the image quality problems of floating objects, artifacts, structural distortion and the like in the image are eliminated, the generated image is more stable and real, the spatial continuity is improved, and the immersion and exploration experience of a user are enhanced.

The invention provides a core photograph processing method based on perspective transformation and a graphical user interface. The core photograph processing method comprises the following steps: S1, collecting a high-definition core photograph; s2, establishing a graphical user interface; s3, defining and designing a user-defined function for selecting the core photos, and acquiring file paths of all the core photos to be processed in batches by using the function; s4, defining and designing a user-defined function to obtain four angular points of the core picture selected in the S3, performing perspective transformation, and storing the four angular points; s5, creating a'batch selection photo 'button and an'interactive perspective transformation' button, and binding the buttons with the custom functions in S3 and S4; and S6, performing packaging on the steps S2 to S5 by using Anacoda, and creating an exe executable file. According to the method, on the premise that the core box does not need to be preprocessed, a method for rapidly and accurately correcting deformation and cutting a large number of core photos is provided for geologists based on the graphical user interface.

The invention relates to the technical field of scene matching navigation, and provides a multi-mode scene matching navigation positioning method and device for a self-adaptive variable-structure unmanned aerial vehicle. According to the method, geographic coordinates corresponding to all pixel points in an aerial image are determined according to pose information in the aerial image of the unmanned aerial vehicle, and the geographic position of the unmanned aerial vehicle with errors is corrected based on difference information between the pixel coordinates of the central point of the aerial image of the unmanned aerial vehicle and the geographic position corresponding to the center of a camera. Based on the corrected geographic position of the unmanned aerial vehicle, determining a reference image pixel coordinate corresponding to the angular point pixel coordinate of the aerial image, and obtaining a simulated aerial image under the view angle of the unmanned aerial vehicle through perspective transformation so as to determine a homography transformation matrix; according to the method, the satellite reference image pixel coordinates of the interest point in the aerial image are determined based on the homography transformation matrix and the perspective transformation matrix, and finally the longitude and latitude of the interest point are determined, so that the limitation of the traditional method in view angle difference compensation, computing resource constraint and cross-modal processing is solved, and the universality, robustness and positioning accuracy of the system are improved.

The invention provides an indoor human body positioning and track generation method and device, equipment and a medium, and the method comprises the steps: obtaining target Bluetooth data and position data with timestamps, achieving the data fusion through a data fusion algorithm, and obtaining a first moving track point set; based on a preset data acquisition period and timestamp information of each track point of the first moving track point set, when the track points are missing, detecting by using a computer vision algorithm and determining a corresponding target by using a tracking detection algorithm based on video data acquired by monitoring equipment, and obtaining an image sequence containing the target; determining an actual physical coordinate of a target in each image based on camera calibration and perspective transformation, obtaining an image positioning result and a second moving track point set, and aligning with a timestamp of the first moving track point set and realizing data fusion to obtain a third moving track point; and according to the map information, matching each track point of the third moving track point set to a grid map to form a target track map.

The invention discloses an automatic cattle checking method and system based on two-dimensional panoramic vision, and belongs to the technical field of computer vision and intelligent breeding, and the method comprises the following steps: reasonably arranging a plurality of fixed cameras in a cattle farm area, obtaining the internal reference and distortion coefficient of each camera through employing a camera calibration technology, and obtaining the internal reference and distortion coefficient of each camera; performing real-time distortion correction on the collected video frames; on the basis of side-by-side splicing or homography matrix perspective transformation, all paths of corrected images are fused into a complete two-dimensional panoramic image, and a view blind area of a single camera is eliminated; calling a lightweight target detection model on the panorama, and extracting bounding boxes and confidence coefficients of all cattle at one time; sorting according to the confidence from high to low, performing IOU non-maximum suppression de-duplication on the detection frame, and ensuring that the same cattle only counts once in an overlapping region; and the whole process is completed through the edge computing node. According to the invention, efficient, real-time and accurate on-site cattle checking of a large-scale cattle farm can be realized.

The invention discloses a multi-camera synchronous electronic image stabilization method and device based on virtual camera imaging, and the method comprises the steps: obtaining and splicing images captured by all cameras in a multi-camera cooperative imaging system, and constructing a virtual camera image; calculating the translation offset of the virtual camera according to the perspective transformation matrix between the adjacent cameras, and calculating the internal reference of the virtual camera; converting the pixel coordinate system into a to-be-stabilized virtual camera coordinate system according to the internal reference of the virtual camera; converting the compensation amount to a virtual camera coordinate system to be stabilized; calculating an image stabilization rotation matrix for converting the coordinate system of the virtual camera to be stabilized into the coordinate system of the stable virtual camera; and converting the to-be-stabilized image from the stable virtual camera coordinate system to the pixel coordinate system according to the image stabilization rotation matrix and the internal reference of the virtual camera to obtain a stabilized image. The method can be applied to multi-camera synchronous electronic image stabilization in a multi-camera collaborative imaging system, and high-quality anti-shake video real-time output is realized.

The invention discloses a digital watermark registration positioning method based on linear Hough transform, which is applied to positioning and identification of a digital watermark on a digital image, and comprises the following steps: encoding input data, adding random offset, generating basic unit watermark blocks, paving the basic unit watermark blocks, and reserving intervals for a linear detection positioning algorithm. In the extraction algorithm, firstly, preprocessing and a Hough straight line detection algorithm are carried out on a carrier image, detected straight lines are grouped, intersection point information between two groups is calculated, vertex positions of basic unit watermark blocks are positioned by calculating the mutual position relation between intersection points, and then perspective transformation correction is carried out on each basic unit watermark block. And the embedded basic unit watermark block information is obtained. And then extracting coding information from the basic unit watermark block, and sequentially carrying out mask removal, error correction and decoding to recover original embedded information. The invention further discloses a system and application for implementing the method, and the system and the application have wide application value.

The invention provides a suspension bridge main cable wire bulging detection method based on multi-feature fusion, and relates to the field of bridge detection. The method comprises the following steps: detecting the position of a carrier roller by using a deep learning target detection algorithm, constructing a perspective transformation matrix by taking the standard geometric dimension of the carrier roller as a benchmark, converting a squint image into a standard top view, extracting a cable strand area mask so as to construct a width contour vector, carrying out quantitative analysis on drum wire characteristics from six dimensions, and establishing a comprehensive scoring model. Through weighted fusion of multi-dimensional features, intelligent judgment and severity quantitative evaluation of drum wires are realized, the detection efficiency is improved, multi-dimensional feature fusion avoids missing detection of a single feature, good adaptability to a complex field environment is achieved, safety and traceability are improved, manual intervention is reduced through whole-process automatic processing, and the detection efficiency is improved. And the result objectivity is ensured by a multiple judgment mechanism.

The invention discloses a product detection method and device, electronic equipment and a readable storage medium, which are applied to the technical field of image processing, and the method comprises the steps: extracting a plurality of product edge feature points from an image of a circular product, carrying out the fitting of the plurality of product edge feature points to obtain an ellipse, and carrying out the building of a standard circle through the ellipse, then, a perspective transformation matrix is constructed by using an ellipse and a standard circle, perspective transformation is carried out on the multiple product edge feature points and images respectively, non-standard circles presented by qualified circular products in the images due to camera shooting angle or product position inclination are corrected, complete visual angle correction is achieved, and the product quality is improved. And finally, based on the plurality of transformation feature points obtained by perspective transformation and the target image obtained by perspective transformation, carrying out qualification detection on the circular product, so that the over-detection rate can be reduced, and the detection efficiency and precision of the circular product can be improved.

The invention provides a method and a device for mapping image coordinates to desktop coordinates, equipment and a storage medium, which are used for improving the positioning precision of a projection touch system in a real environment. The method comprises the steps of performing distortion correction on original gesture video data according to internal parameters of a camera and a lens distortion coefficient to obtain a corrected image frame sequence, and performing interaction space calibration according to the corrected image frame sequence to obtain a distortionless interaction plane, performing perspective change matrix solving on the distortionless interaction plane according to the target vertex coordinate set to obtain orthogonal mapping data, and when gesture key points are detected in the distortionless interaction plane, performing perspective transformation and normalization processing on gesture coordinates of the gesture key points according to the orthogonal mapping data to generate initial desktop control coordinates; and performing display mapping processing on the initial desktop control coordinates according to the physical attribute parameters of the target display to obtain desktop control coordinates.

The invention relates to a calibration plate circle center detection method, device and equipment and a storage medium, in the application, a perspective transformation matrix is determined by creating a virtual standard square, and the perspective transformation matrix is utilized to map an original calibration plate image into a square calibration plate image, so that the calibration plate circle center detection accuracy is improved. The problem of deformation of the calibration plate during oblique shooting can be effectively solved, so that circle center detection is not influenced by view angle inclination; moreover, the contour of each target circle is combined through a clustering algorithm in the square calibration plate image, and the coordinate of the combined center is determined through confidence, so that noise interference can be effectively suppressed, and the precision of circle center detection is improved.

The invention discloses a computer vision-based railway tool online and offline checking method, which comprises the following steps of: acquiring online and offline and warehouse tool images of a high-speed railway work section, cutting a target area through an area focusing algorithm to eliminate interference, and combining enhancement technologies such as perspective transformation and multi-transformation splicing with a generation method based on SAM and a diffusion model. Constructing a diversity training data set; constructing a text vision multi-modal rotating target detection model, and guiding the model to pay attention to key categories by enhancing feature extraction and fusing text information and visual features of a job schedule; detecting parameters are adjusted by adopting a self-adaptive threshold optimization algorithm, a few-sample detection technology is combined, and after category sensing scale filtering, outlier frame suppression and cross-category non-maximum suppression processing are performed, a result is compared with an operation plan, and a result is output; and based on a sample cutting and screening mechanism driven by false detection and missing detection, high-value samples are mined and supplemented to a training set. The method has the advantage that the safety and efficiency of railway tool management are improved.

The invention discloses a vehicle target detection tracking and trajectory data extraction method based on deep learning, and relates to the technical field of unmanned aerial vehicle aerial photography. The method comprises the following steps: carrying out stable frame processing on an unmanned aerial vehicle video, extracting feature points and feature vectors through an SURF algorithm, matching and screening reliable matching pairs through an FLANN algorithm, calculating a homography matrix through an RANSAC algorithm when a condition is met, carrying out perspective transformation to eliminate jitter, and outputting a stable video sequence; vehicle target detection: introducing an AIFI module to construct an improved YOLOv5OBB model, and outputting vehicle rotation bounding box parameters and categories after training; vehicle tracking and trajectory extraction are carried out, cross-frame tracking is realized based on a DeepSORT model, original trajectory data are preprocessed, and the speed, the acceleration, the additional lane number and the ID of an adjacent vehicle are calculated. According to the method, the problems of video jitter and insufficient detection precision are effectively solved, and the accuracy and continuity of track data extraction in the highway scene are improved.

The invention discloses a writing process analysis method based on computer vision, which relates to the technical field of handwritten character recognition, and comprises the following steps: carrying out perspective transformation on real-time writing data, obtaining a writing image frame sequence, carrying out motion blur correction and pen point coordinate positioning on the writing image frame sequence by adopting an RAFT (Reversible Addition Fragmentation Transform) optical flow algorithm, and forming a writing track data flow; and performing curvature segmentation on the writing trace data stream to obtain discrete stroke segments, and performing spatial topological correlation and geometric structure mapping on the discrete stroke segments to generate a writing stroke topological graph. Through the RAFT optical flow algorithm and the stroke semantic analysis model, the precision of stroke recognition is improved, and efficient and accurate analysis from dynamic visual input to structured character output is achieved.

The invention is suitable for the technical field of automobile auxiliary driving, and provides a quick panoramic image splicing method based on a star flash wireless communication technology, which comprises the following steps of: S1, calibrating internal parameters and distortion coefficients of a camera by adopting a Zhang Zhengyou calibration method, calibrating external parameters by combining perspective transformation, and mapping each view angle image to a unified overlook coordinate system; s2, acquiring an image through a camera supporting a star flash protocol, and transmitting the image to a vehicle body domain controller by using a star flash technology; and S3, after distortion removal and perspective transformation are performed on the received image, a local pyramid fusion strategy is adopted: binary mask marking and multi-band fusion are performed on an overlapping region, original data are reserved in a non-overlapping region, and a panoramic image is finally output. Through combination of high-precision calibration, local pyramid fusion and a star flash communication technology, high-quality and low-delay splicing of vehicle-mounted panoramic images is realized, and the stability and practicability of the system in a complex environment are improved.

The invention belongs to the technical field of building construction, and particularly relates to a three-dimensional point cloud virtual assembling method based on image recognition. A three-dimensional point cloud virtual splicing method based on image recognition comprises the steps that point cloud data of splicing plates and side cross beam splicing plates at the joints of side main beams, side box beams and steel truss web plates are collected through a base station type scanner, standard size frame lines are constructed by conducting manual segmentation, voxel filtering and Gaussian filtering on the point cloud data, an image is intercepted, and the three-dimensional point cloud virtual splicing method based on image recognition is obtained. Gray processing, threshold segmentation, perspective transformation and size identification are carried out on the image, two-dimensional coordinates of bolt hole groups and frame lines are obtained, a two-dimensional to three-dimensional coordinate system is constructed, three-dimensional coordinates in the bolt hole groups are obtained through conversion, and virtual splicing is carried out according to the corresponding relation between the bolt hole groups of different components. According to the method, the efficiency and the precision of virtual assembly are remarkably improved.

The invention provides a safe anti-collision method and system for a table top of a laser cutting machine, and belongs to the technical field of visual inspection of the laser cutting machine. The method comprises the steps that visual devices are arranged on the two sides of the table top movement direction of the laser cutting machine, and the heights and angles of the visual devices are adjusted; calibrating physical coordinates and image coordinates of four corners of the table top of the laser cutting machine, and establishing a mapping relation between the physical coordinates and the image coordinates; the method comprises the following steps: respectively shooting table top images through two cameras, performing image preprocessing, performing dense perspective transformation according to a mapping relation between physical coordinates and image coordinates, performing image splicing on the table top images, and performing obstacle recognition; whether the laser cutting machine stops moving is judged according to the obstacle recognition result. Visual devices are arranged on the two sides to cover the full view of the table top, the physical and image coordinate mapping relation is calibrated, whether shutdown is conducted or not is judged through image processing and obstacle recognition, collision is effectively avoided, safety is guaranteed, and efficiency is improved.

The invention provides a panoramic image generation method and device, equipment and a storage medium, and the method comprises the steps: carrying out the single-lens calibration of a wide-angle camera according to a checkerboard calibration method, and obtaining a distortion correction matrix; controlling the wide-angle camera and the array camera to perform time-space synchronous acquisition to obtain an initial panoramic image and a sub-image set; performing nonlinear distortion correction on the initial panoramic image according to the distortion correction matrix to generate a distortionless standardized panoramic image; acquiring sub-image feature information of each sub-image in the sub-image set, and acquiring a projection position, a perspective transformation matrix and a scale transformation matrix corresponding to each sub-image in the standardized panoramic image according to the sub-image feature information; and establishing a panoramic canvas coordinate system according to the standardized panoramic image, projecting each sub-image into the panoramic canvas coordinate system according to the corresponding perspective transformation matrix and the projection position, adjusting the size corresponding to the sub-image in the panoramic canvas coordinate system according to the scale transformation matrix, and generating a panoramic image according to the adjusted panoramic canvas coordinate system.

The invention belongs to the technical field of computer vision, and mainly provides a wheel target image recognition method and device, electronic equipment and a readable storage medium, and the method comprises the steps: carrying out the perspective transformation of an original wheel target image, and obtaining the center coordinates of a concentric circle and a common circle in the wheel target image after the perspective transformation; the circle center coordinates of the concentric circles and the common circles in the wheel target image after perspective transformation are mapped back to the original wheel target image, and the circle center coordinates of the concentric circles and the common circles after eccentric compensation in the original wheel target image are obtained; therefore, correction of the circle center coordinates of the concentric circle and the common circle in the original wheel target image is realized, so that the recognition accuracy of the wheel target image can be improved when the recognition result of the original wheel target image is obtained based on the circle center coordinates subjected to eccentric compensation; the technical problem that the recognition accuracy of an existing wheel target image is low is solved.

Systems and methods for generating, adjusting, and augmenting views of an environment of a host vehicle are disclosed. In one implementation, a system includes a processor programmed to receive a plurality of images captured by at least one camera onboard the host vehicle, wherein each of the plurality of images includes a representation of at least a portion of the environment of the host vehicle; determine a set of points associated with each of the plurality of images; determine, for each set of points determined for the plurality of images, a perspective transformation that maps each set of points to a common coordinate system; apply the perspective transformation determined for each set of points to a corresponding one of the plurality of images; generate a merged image based on the perspective transformation applied to the plurality of images; and output the merged image to a display device.

The invention relates to the technical field of image processing, and discloses an image stitching method, device and equipment, and the method comprises the steps: recognizing a plurality of target structures in a to-be-stitched image; determining a homography matrix by using the target feature points of the plurality of target structures in different to-be-spliced images; the target feature point is determined based on the geometrical shape of the target structure; performing perspective transformation on one to-be-stitched image in the to-be-stitched images by using the homography matrix to obtain a distorted image; and splicing the distorted image with another to-be-spliced image serving as the reference image. According to the invention, the recognition and registration accuracy of the feature points can be improved, so that the image splicing quality is ensured.