459 results about "Feature point matching" patented technology

The invention discloses a multi-target tracking method combining camera motion compensation and pseudo depth estimation, belongs to the field of computer vision, and is suitable for a complex automatic driving road environment. The method comprises the following steps: constructing a training set and a test set; detecting the image by using a deep learning detector and extracting features; a Kalman filter is adopted to correct a motion modeling state vector, and the target position and size prediction precision is improved; solving a homography matrix through feature point matching, performing global camera motion compensation, and reducing camera jitter and displacement interference; target pseudo depth information is calculated, hierarchical cascade matching is carried out, and association performance in dense and shielding scenes is optimized; a three-level cascade strategy is adopted to complete high confidence degree, low confidence degree and residual target matching in sequence; and finally, outputting a tracking result with a detection frame and identity information to obtain a trained model. According to the invention, accurate detection and stable tracking of multi-category targets can be realized in a complex environment, and identity switching is effectively reduced.

The invention discloses an intelligent change detection method for non-aligned two-time-phase images of an inspection unmanned aerial vehicle, and relates to the technical field of unmanned aerial vehicle image processing and change detection, and the method comprises the steps: obtaining two-phase images collected by a low-altitude unmanned aerial vehicle under a fixed route and same sensor parameters; a lightweight registration model is constructed and trained, feature point matching is utilized to predict matching point pairs, a homography matrix is calculated, and accurate registration of non-aligned images is achieved; and constructing and training a change detection model based on image pair interaction feature fusion, analyzing the aligned image after registration, and outputting a change information binary image. The method can effectively solve the problem of non-alignment caused by position and angle differences during two-time-phase image acquisition of an unmanned aerial vehicle, and the technical problems of low precision, poor robustness and insufficient calculation efficiency of a traditional method in change detection, effectively improves the automation level of low-altitude safety monitoring of ground highways and railways, reduces the maintenance cost, and improves the safety of the unmanned aerial vehicle. The important application value is realized.

The invention provides a bridge stay cable video multi-target identification tracking and vibration extraction method based on an unmanned aerial vehicle. The method comprises the following steps: 1, constructing a bridge stay cable refined inclined slender target detection model based on a YOLOv11 model; 2, providing a multi-target tracking algorithm fusing the inclined slender target detection model and a StrongSORT algorithm; step 3, improving a displacement extraction method combining SIFT / ORB feature point matching and a sub-pixel refinement technology; 4, designing an unmanned aerial vehicle motion correction algorithm based on variational mode decomposition and time-frequency domain combined screening; and 5, constructing a joint working modal analysis algorithm for realizing combination of a natural excitation technology and a random subspace recognition algorithm. According to the method, high-precision extraction of the vibration signals of the stay cable and identification of modal parameters of the stay cable are realized, and technical support is provided for health monitoring of the large-span cable-stayed bridge.

The invention relates to the technical field of hydraulic engineering safety monitoring, and discloses a dam safety studying and judging method based on monitoring data multi-physics field simulation, which comprises the following steps: S1, multi-source data acquisition and time-space alignment; s2, dynamically updating the numerical model; s3, safety evaluation and early warning decision making; s4, performing multi-source risk coupling analysis; and S5, issuing the early warning information in a multi-mode manner. According to the method, time-space reference unification of multi-source monitoring data is achieved through feature point matching and sliding window cross-correlation analysis, a self-adaptive Kalman filtering algorithm is adopted to dynamically invert permeability coefficients and elastic modulus parameters, and boundary conditions of a finite element model are adjusted in combination with real-time water level changes; the dynamic simulation precision of a seepage field-displacement field-stress field coupling model is improved, the static evaluation limitation of a fixed threshold value method is broken through through a three-dimensional time-varying safety envelope surface and a Bayesian network grading early warning decision tree, and the risk prediction capability in the flood routing process is enhanced through multi-parameter joint probabilistic reasoning.

The invention provides a camera pose and 3D Gaussian point cloud joint optimization method and system, and belongs to the technical field of computer vision, and the method comprises the steps: obtaining a to-be-processed image; processing the obtained image by using a pre-trained optimization model to obtain an optimized image; the optimization model comprises a feature extraction module, a matching module, a generation module and an optimization module; the feature extraction module is used for feature extraction; the matching module is used for carrying out feature point matching; the generation module is used for screening inner points through RANSAC and calculating a basic matrix to generate an initial point cloud; and the optimization module is used for performing nonlinear optimization on the initial pose and the point cloud, and outputting camera parameters and sparse Gaussian point cloud. According to the invention, synchronous high-precision estimation of the camera pose and the 3D Gaussian point cloud is realized; dependence on a pre-calculation pose or a pre-training model is abandoned, and the pose and the point cloud are directly combined and recovered from a disordered image; a large-view-angle jump scene is processed through a progressive optimization strategy, and the stability of the pose and the point cloud under sparse matching is ensured.

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 discloses a gas leakage detection method. The method comprises the following steps: S10, carrying out denoising and image enhancement on an acquired visible light image and an acquired thermal infrared image; s20, carrying out image feature point matching; s30, fusing the visible light image and the thermal infrared image after registration; and S40, gas leakage detection based on the improved YOLOv8 framework. According to the invention, thermal imaging, multispectral and visible light imaging multi-source fusion technology is adopted, so that the problems of efficient detection and accurate identification of multiple types of gases in a broadband range are solved.

The invention discloses a landscape object identification and explanation method and system, a medium and a product, and relates to the technical field of augmented reality. The method comprises the following steps: when the head motion state of a user is stable, triggering a camera to collect and recognize a single-frame image, matching a recognition result with a regional knowledge base, and generating a candidate object list and coarse-grained poses which are sorted according to spatial probability and historical popularity; performing feature point matching on a current frame and an ideal frame in the video stream, and calculating a fine-grained pose of a target landscape in the candidate object list relative to a camera coordinate system according to a matched target feature point; according to the fine-grained pose and the real-time eye movement data, determining a drop point part of the sight line of the user on the target landscape, forming a multi-dimensional query vector according to the drop point part, the intention of the user and the environment context parameters, and generating multimedia explanation content; and carrying out fusion rendering on the multimedia explanation content and the real-time video stream, and outputting and displaying. By implementing the technical scheme, the personalized multimedia explanation content can be generated.

The invention provides a multi-camera-based AI glasses FOV extension method and system, and the method comprises the steps: transmitting a synchronization signal to a plurality of cameras, enabling the plurality of cameras to carry out the exposure at the same time, and obtaining a plurality of frames of original images at the same time; pre-processing the multiple frames of original images to obtain pre-processed images; key feature points of the preprocessed image are extracted through a feature point detection algorithm, and feature descriptors are generated; through a feature matching algorithm, feature point matching is carried out on the adjacent preprocessed images by utilizing the feature descriptors, and a local transformation matrix between the adjacent preprocessed images is calculated; based on the local transformation matrix, fusing the adjacent preprocessed images to obtain a fused image of the adjacent preprocessed images; and all the fused images are spliced to obtain a large-view-field-angle image covering the target view angle, so that the method has the advantages of effectively expanding the shooting view angle, eliminating image distortion and improving the image splicing naturalness.

An image processing device includes: a region dividing unit that divides each of a plurality of captured images captured by an in-vehicle camera into a plurality of divided regions; a feature point detection unit that detects, as a feature point, a portion exceeding a threshold value for each of the divided regions; a threshold value setting unit that sets the threshold value for each of the divided regions so as to evenly distribute the feature points detected in each of the divided regions; a feature point matching unit that performs matching by associating the feature points detected in the plurality of captured images with each other between the plurality of captured images; and a relative parameter calculation unit that calculates a relative parameter on the basis of the feature points associated between the plurality of captured images.

The invention provides a digital pathological image high-color accurate splicing method and device based on feature point matching. The method comprises the following steps: obtaining a target splicing region in a target splicing image and a to-be-spliced region in a to-be-spliced image based on a scanning sequence of digital pathological images; performing affine transformation on the to-be-spliced region by using the affine transformation matrix to obtain an initial correction region; constructing a global color mapping function, and mapping the initial correction image by using the global color mapping function to obtain a color correction image; and splicing the color correction image and the target splicing image based on the scanning sequence of the pathological images to obtain a digital pathological image splicing result. According to the scheme, only the target splicing area and the to-be-spliced area are subjected to feature point matching, the cumulative distribution function of the effective feature points is calculated according to the RGB channels, and the global color mapping function is constructed and optimized, so that accurate color correction is realized, and the visual integrity and consistency of the spliced image are improved.

The invention relates to a camera point location offset detection and automatic calibration method suitable for high-voltage equipment monitoring. The method comprises the following steps: constructing current preset bit reference primitive information; constructing a mapping relation between a unit PT step length and an imaging space in the current scene; constructing a multi-layer surrounding type auxiliary reference primitive information matrix; judging whether the camera deviates or not; searching optimal matching meta-information; and calculating a deviation rectification PT value, and carrying out PTZ camera control and deviation rectification. According to the method, high-precision automatic correction of the preset position of the camera in high-voltage equipment monitoring is realized, key feature points of power equipment can be stably extracted in a complex field environment by fusing local feature matching and a dynamic key point selection mechanism, and the matching accuracy is greatly improved; actual deployment shows that the repeated positioning precision of the preset position can be improved to + / -2 pixels from original + / -15 pixels, the manual rechecking frequency is remarkably reduced, and differential requirements of equipment of different voltage grades on the inspection precision are met.

For a cross-modal target prompt matching task, the invention provides a training-free multi-stage matching method based on local feature points and global spatial transformation so as to realize robust positioning of the same target in different modal images. According to the method, firstly, feature points and descriptors of a target prompt image and a to-be-matched image are extracted by utilizing a pre-trained SuperPoint, then, MINIMAGlue finely adjusted on cross-modal data is preferentially adopted to perform feature point matching, and when matching fails, LightGlue is switched to serve as remediation; in the matching process, an optimal angle is selected through a rotation enumeration strategy to enhance matching robustness; in the coarse positioning stage, calculating a minimum enclosing rectangle according to matching points in the ROI, and if the matching points are insufficient, estimating a homography matrix by adopting RANSAC and carrying out geometric correction on a rectangular frame; in the fine positioning stage, the central point of the coarse positioning result is used as a prompt to input MobileSAM for target segmentation, and a final rectangular frame is corrected through area consistency constraint. Through integration and fusion of the multi-stage process, accurate matching and stable positioning of cross-modal target prompt are realized under a training-free condition, and the generalization ability and practicability of the method are improved.

The invention discloses a calibration method and system based on multi-scale feature fusion and an adaptive optimization algorithm, and belongs to the technical field of computer vision. The method comprises the following steps: collecting a plurality of images containing calibration modes; detecting and positioning feature points by adopting a multi-scale feature fusion technology, and distributing a confidence score for each feature point; carrying out feature point matching and removing abnormal values by using an improved RANSAC algorithm based on the prior geometrical relationship and confidence score; preliminarily calculating camera parameters; and a nonlinear optimization model fusing confidence weighting and adaptive regularization is constructed, joint optimization is performed on camera parameters, and a high-precision calibration result is output. According to the method, through an optimization mechanism of multi-scale feature fusion and confidence guidance, the problems of unstable feature point detection, abnormal value sensitivity and insufficient precision of a traditional method in a complex environment are effectively solved, and the accuracy, robustness and automation degree of image calibration are remarkably improved.

The invention discloses a pathological image recognition system and method, and relates to the technical field of image recognition, and the method comprises the steps: continuously adjusting the relative posture between pathological section images through an iterative optimization algorithm until a local deformation parameter combination enabling a target function between two feature point sets to be minimum is found, and positioning a local region with the highest similarity, the method comprises the following steps: selecting a local region, defining local pathological interest regions of upper and lower A rows and left and right B columns by using a central point of the local region, determining a focus matching region and a matching point F of the focus matching region through multiple times of iteration local deformation adjustment in combination with a minimum tolerance threshold, constructing a universal pathological equation and a contact equation of two pathological section images at the matching point F, and realizing multi-region joint recognition. According to the recognition system, accurate matching and multi-region joint recognition of focus regions among a plurality of pathological section images are realized through the technical means of constructing a space mapping matrix, finely matching feature points, fitting local interest regions and three-dimensional form expressions and the like.

The invention relates to the technical field of 3D visual inspection, in particular to a dynamic shielding area multi-view time sequence point cloud completion method and system in 3D visual inspection. The method comprises the following steps: collecting a multi-view time sequence image of a dynamic detection object for providing a data basis for feature point tracking; and carrying out Lucas-Kanade optical flow method tracking on the feature points, associating the feature points of different visual angles, describing a motion track of the feature points through an optical flow vector, carrying out cross-visual-angle and cross-time feature point matching, and establishing time sequence feature association to obtain time sequence feature points. According to the dynamic occlusion area multi-view time sequence point cloud completion method in 3D visual detection, 360-degree time sequence data of a dynamic detection object is acquired through multi-view time sequence image acquisition, and comprehensive basic data support is provided for subsequent feature tracking.

The invention relates to a three-dimensional model and geographic coordinate dynamic matching method based on multi-source data fusion, and relates to the technical field of three-dimensional live-action modeling, and the method comprises the steps: obtaining three-dimensional model data, geographic reference data, sensor combination data and environment data of a target area, carrying out the preprocessing of the data, and carrying out the calculation of the data; the method comprises the following steps: positioning target sensor combination data based on target environment data to obtain a target positioning result, performing dynamic coordinate system conversion on target three-dimensional model data to obtain a converted three-dimensional model, determining feature point matching information by using a preset AI auxiliary matching model, and finally determining a calibrated three-dimensional model. According to the invention, high-precision and high-robustness alignment of the three-dimensional model and the geographic coordinates is realized.

The invention belongs to the technical field of augmented reality, and particularly discloses an offline store inspection authenticity verification method based on AR environment feature point matching, and the method comprises the steps: scanning a specific region of a store through a mobile terminal, collecting environment visual feature points and pose parameters to construct and upload a sparse point cloud map, and generating a uniquely bound AR scene model; setting a virtual punch-in point in the model, and recording an anchor point coordinate under a world coordinate system of the virtual punch-in point; when the inspection is executed, the mobile terminal preloads a corresponding AR scene model based on a geographic position, extracts feature points by collecting images in real time, and matches the feature points with feature points in the model; and after successful matching, calculating a real space distance between the mobile terminal and the virtual clock-in point, and if the real space distance is smaller than a threshold value, unlocking the inspection task. According to the method, a dual verification mechanism based on pose difference and dynamic visual guidance are introduced, so that the defects of low positioning precision and easiness in counterfeiting in the prior art are overcome fundamentally, and high-precision and high-anti-counterfeiting authenticity verification of offline inspection behaviors is realized.

The invention discloses an AI-based injury condition evidence obtaining and processing method and device and electronic equipment, and the method comprises the steps: carrying out the feature point matching of a first infrared image and a first optical image according to the edge contour of a to-be-detected scene, so as to generate a second infrared image and a second optical image; performing feature point matching on the second infrared image and the second optical image according to the edge contour of the wound to generate a third infrared image and a third optical image; wound consistent features and wound difference features in the third infrared image and the third optical image are obtained based on an AR virtual scale; and according to the wound consistent characteristics and the wound difference characteristics, determining the wound type and the wound cause of the to-be-detected target, and based on the wound type and the wound cause of the to-be-detected target, realizing wound condition evidence obtaining and processing of the to-be-detected target. Both the device and the electronic equipment can execute the injury condition evidence obtaining and processing method. Through the arrangement, the efficiency of evidence collection and treatment of injury conditions can be improved.

The invention discloses a large-parallax image splicing method and device, equipment and a storage medium, and belongs to the technical field of image processing. The method comprises the following steps: performing feature point matching on a to-be-spliced feature point set and a reference feature point set to obtain a feature matching point pair set; according to the parallax value of each feature matching point pair in the feature matching point pair set, layering the parallax of an overlapping region intersected with the reference image in the to-be-spliced image to obtain a plurality of parallax layers and a matching feature point pair set corresponding to each parallax layer; according to the matching feature points to be spliced in the matching feature point pair set of each parallax layer, determining each geometric feature region by using a region growing method; and according to the global homography matrix and the local homography matrix of each geometric feature region, performing fusion processing on the to-be-spliced image and the reference image to obtain a target spliced image. By using the technical scheme provided by the invention, the accuracy of large parallax scene image splicing can be improved.

The invention provides a core image registration method based on multi-modal image segmentation and feature matching. The method is characterized in that the imaging difference between a rock core laser confocal image and a CT image is reduced through preprocessing color conversion and a normalization strategy, a marker is segmented by using deep learning, an algorithm target area is optimized in combination with morphological operation, feature points are detected under the guidance of an ROI area by using an ORB algorithm, and a target image is obtained. Matching of feature points is achieved by combining an FLANN feature point matching algorithm and a random K-d tree acceleration algorithm, image alignment is completed through a rigid transformation model, and registration precision is quantified through Dice. According to the method provided by the invention, the registration problem of the multi-modal core image caused by imaging difference is solved, the average Dice reaches 0.9210, and the registration efficiency and accuracy are remarkably improved.

The invention discloses an image matching method and device, equipment, a storage medium and a product, and relates to the technical field of image matching, and the image matching method comprises the steps: obtaining different-source to-be-matched images at different visual angles; sequentially performing feature extraction on the images to be matched through a preset feature point matching model to obtain key points and descriptors in each image to be matched, and performing matching on the key points based on the descriptors to obtain initial matching points obtained through initial matching, the number of channels in a convolutional layer of the preset feature point matching model is negatively correlated with the spatial resolution of the to-be-matched image, the change degree of the number of channels in a first stage is smaller than the change degree of the number of channels in a second stage, and the first stage and the second stage are determined based on the spatial resolution; and performing error filtering on the initial matching point to obtain a filtered target matching point. According to the invention, when an image with high spatial resolution is processed at an early stage, the calculation cost can be significantly reduced by reducing the number of channels.

The application provides an image registration data processing method and system, relates to the technical field of image registration, and obtains a plurality of structural feature images in a plurality of monochrome laser images; according to the plurality of structural feature images, a plurality of ghosting areas with excessive eye movement are obtained, each of the ghosting areas includes a corresponding ghosting floating image and a ghosting reference image; a feature point corresponding relationship is obtained; a plurality of feature point pairs with accuracy greater than a preset threshold are obtained, feature point matching processing is performed, and a processed ghosting floating image is obtained; pixel registration processing is performed, and a region registration image is obtained; a plurality of region registration images are aligned and fused, a reference image and a floating image are combined, and a registration image is obtained. The application solves the technical problems of the prior art, such as inconsistent image offset in the image acquisition process, low whole-image registration accuracy, and the like, and achieves the technical effects of correcting the partial image offset caused by patient eye movement and improving real-time registration accuracy.

Embodiments of the present disclosure disclose image alignment methods, devices, electronic equipment and computer readable media. A specific implementation of the method comprises: respectively performing distortion correction on a first image and a second image to obtain a third image and a fourth image; performing feature point matching on the third image and the fourth image to obtain a feature point information group sequence; determining whether a homography matrix corresponding to a current rotation angle exists; in response to the homography matrix not existing, determining the homography matrix according to the feature point information group sequence; and performing perspective transformation on the fourth image according to the homography matrix, so that the fourth image after the perspective transformation is aligned with the third image. This implementation does not need to set a complex hardware structure, thereby achieving efficient and accurate image alignment.

The invention discloses a method and system for monitoring the start-stop state of a conveying belt based on a traditional visual algorithm, and belongs to the technical field of traditional visual algorithms, and the method comprises the steps: installing an industrial camera at a proper position above the conveying belt; after receiving the original image acquired by the industrial camera, the monitoring host carries out graying processing on the image and carries out de-noising processing on the grayscale image; performing histogram equalization processing on the denoised image to enhance the contrast of the image; selecting a plurality of feature points in the belt area, and recording coordinates and feature descriptors of each feature point; for every two adjacent frames of images, matching the extracted feature points by using a feature point matching algorithm, and calculating an optical flow vector of each pixel point between the adjacent frames by using a feature descriptor-based matching algorithm and an optical flow method; and outputting a state and giving an alarm. Manual intervention is greatly reduced, accuracy is high, cost is low, adaptability is high, and implementation and maintenance are easy.

The application discloses a kind of earphone anti-misoperation control method and system based on fingerprint identification, it is related to the field of fingerprint identification, including, user static fingerprint data and dynamic fingerprint data are collected and preprocessed and stored in fingerprint database;Based on the real-time acquisition user dynamic fingerprint data, generate elliptical dynamic fingerprint window;Elliptical dynamic fingerprint window is converted into rectangular dynamic fingerprint window, and fingerprint feature point matching is carried out;User earphone use data is collected, and touch scene analysis model is constructed, and earphone use scene is obtained;According to earphone use scene, dynamically adjust the operation logic of earphone.The application generates rectangular dynamic fingerprint window, accurately captures the dynamic change of user finger in the touch process, can adapt to the change of finger in different operation, realizes the effective tracking and identification of real-time dynamic fingerprint, improves the accuracy of fingerprint matching, reduces misoperation, and enhances the adaptive capacity of system to complex touch operation.

The embodiment of the invention provides an unmanned aerial vehicle real-time visual positioning method and system, electronic equipment and a storage medium, and belongs to the technical field of visual positioning. According to the scheme, a flight detection sequence of a takeoff stage is obtained; controlling the unmanned aerial vehicle to descend in the target landing area according to the current position information of the landing stage and a preset landing point; performing visual reference detection according to the current image information, and judging whether a landing marker is detected or not; if the landing marker is not detected, whether reference marker information exists or not is judged according to the current flight height; if the reference marker information exists, performing feature point matching according to the current image information based on the reference marker information, and determining a target positioning point; and controlling the unmanned aerial vehicle to land on the ground according to the target positioning point. The positioning error caused by environmental interference can be reduced, and the positioning precision is improved.

The application discloses a label defect control method and system, which cuts the to-be-detected image by using the ROI region through presetting the ROI region; uses a Swish excitation function to improve a YoloV5 algorithm training model to obtain an AI model more suitable for cloth defects; matches corresponding logos, letters and symbols in the to-be-detected image through a deep learning neural network algorithm; scales the length and width of the to-be-detected image in a proportional manner, and shields the symbols matched by the deep learning neural network algorithm, then uses a SIFT feature point matching algorithm to match feature points of the to-be-detected cloth and the set standard cloth, finds out feature differences between the to-be-detected cloth and the standard cloth, judges difference values and marks difference regions to obtain defect positions and sizes. The label defect control method and system have high automation degree and save labor cost, and have fast detection speed and high detection precision.

The invention relates to the technical field of optical engineering, and discloses a multispectral fusion optical coherence tomography imaging method, which adopts a multispectral light source to cooperatively collect OCT signals, fully utilizes the absorption / scattering characteristic difference of different spectrums on different components of a sample, obtains richer sample structure and component information, and improves the imaging accuracy. A data basis is provided for follow-up accurate identification of sample components, synchronism and wavelength accuracy of multispectral signals are ensured through synchronous driving control and accurate spectrum calibration, and information distortion caused by spectrum deviation is avoided; a registration algorithm based on SIFT feature point matching is adopted, high-precision spatial alignment of multispectral OCT images is achieved, the problem of dislocation between different spectral images is solved, the whole imaging process gives consideration to imaging efficiency and imaging quality, the imaging time is shortened through a synchronous collection mode, the method is suitable for dynamic sample imaging, parameter adjustability of all modules is high, and the method is easy to implement. Flexible adjustment can be performed according to different sample characteristics, and the application range of the method is expanded.

The invention belongs to the technical field of screen detection, and discloses a screen shell size detection method and device, equipment and a storage medium. The method comprises the steps that a mechanical arm drives a laser scanner to move to scan reflective mark points around a to-be-detected screen shell according to a first preset path, and the reflective mark points are standard round points of a fixed size; driving a laser scanner to move through a mechanical arm to scan the outer surface of the to-be-detected screen shell according to a second preset path, and obtaining original point cloud data including the to-be-detected screen shell; the standard model of the screen shell is converted into a point cloud format, and a standard point cloud is obtained after downsampling; obtaining a to-be-measured point cloud based on the reflective mark point and the original point cloud data; and performing rigid registration on the standard point cloud and the to-be-detected point cloud to realize detection of the size of the screen shell. According to the mode, size detection is carried out through three-dimensional model registration and feature point matching, manual scanning is not needed, the automation degree is higher, and the detection precision is higher.