55 results about "Three Dimensional Size" patented technology

The invention relates to the technical field of large-size part three-dimensional measurement, and discloses a large-size part three-dimensional size measurement method and device based on machine vision, and the method comprises the steps: obtaining a material reflection characteristic reference data set and original point cloud data; according to the reference data set, laser power and gain are adjusted in a self-adaptive mode, and optimized scanning parameters are obtained; and performing denoising, feature extraction, region segmentation and plane fitting according to the optimized parameters and the original point cloud to obtain a high-precision geometric dimensioning result. According to the invention, high-precision, high-stability and high-efficiency three-dimensional size measurement can be realized on a large-size part made of a complex material and having a curved surface, and the three-dimensional size measurement method is obviously superior to a traditional scheme.

The invention discloses a high-strength steel welding defect nondestructive detection and identification method based on multi-modal data fusion. The method comprises the following steps: welding data acquisition: acquiring a two-dimensional image and three-dimensional point cloud data of a high-strength steel welding part to form a data pair; performing data space alignment: generating a space-aligned image-depth map data pair; feature extraction and fusion: obtaining fusion features with spatial geometric information and two-dimensional visual information through feature extraction and fusion; defect identification and classification: identifying defect pixels, decoding and recovering space information of a defect area, calculating the three-dimensional size of the defect, taking the feature information associated with a connected domain of each pixel with the defect and the three-dimensional size as input, and automatically classifying defect categories through a pre-trained full-connection neural network classifier. According to the method, the welding defects of the high-strength steel can be accurately identified and accurately and quantitatively analyzed.

The present invention relates to the technical field of industrial process control and detection. Disclosed is a method for real-time measurement of length-width size distribution of a crystal population in a crystallization reactor using binocular telecentric cameras. In-situ environment calibration is performed for a binocular telecentric stereo vision system, a calibration rod capable of probing into an in-situ environment (a reactor / glass tube) is designed, a binocular telecentric stereo vision imaging model is established, and a simple calibration method of rotating the calibration rod suitable for an in-situ limited space and a calibration plate design scheme are proposed. In order to improve binocular matching efficiency, a simplified telecentric stereo epipolar rectification method is further provided, so as to ensure accurate matching of on-site snapshot image pairs. On the basis of the matched image pairs, a three-dimensional reconstruction method using analytical solution-based ray intersection is provided to measure the three-dimensional pose of particles within a crystallizer. Finally, the three-dimensional length and width of a crystal are quantitatively evaluated by means of statistical data of Euclidean distances of pairs of length and width feature points of the crystal. The present invention has high operability, and can achieve the effect of automatically measuring the three-dimensional size of crystals.

The invention discloses a void disease three-dimensional reconstruction method based on a radar signal time domain peak value, and belongs to the field of pavement disease nondestructive testing. Aiming at the problems that the existing ground penetrating radar (GPR) depends on manual interpretation, the resolution of a grey-scale map is low, and the three-dimensional space form of the void disease cannot be obtained, the method comprises the following steps: firstly, processing GPR data and utilizing time domain signal characteristics: identifying zero crossing points corresponding to the first peak value and the last peak value of a single-channel A-scan signal, and accurately positioning the starting depth and the ending depth of a void interface; based on the transverse continuity of the B-scan image, the relevance of adjacent multiple channels of signals is analyzed, the transverse start-stop channel number and the vertical depth range of the void area are determined, and a horizontal plane positioning bounding box is generated; and finally, fusing coordinate information of a multi-channel B-scan positioning frame, and constructing a void disease geometric model containing three-dimensional sizes of a long (X-axis), a wide (Y-axis) and a deep (Z-axis) through space conversion and superposition. According to the method, the depth is directly positioned from the time domain peak value, complex frequency domain calculation is avoided, and the noise immunity is high; horizontal boundary automatic extraction and disease three-dimensional visualization are realized, and accurate geometric parameters are provided for structural safety assessment; the method can be extensively applied to other pavement disease detection.

This invention discloses a multi-vision method for measuring the dimensions of small workpieces based on a binocular vision system, belonging to the field of workpiece dimension measurement technology. Addressing the difficulty in achieving accurate three-dimensional dimension measurement using existing methods, the following solution is proposed, comprising the following steps: S1: Chessboard corner point extraction; S2: Calibration plate removal; S3: Plane fitting; S4: Vertex extraction; S5: Stereo matching algorithm; S6: Dimension calculation. The average absolute error of this invention at different angles is 0.33 mm, the total relative error is 0.90%, and the variance is less than 0.01 mm. 2 This enables relatively accurate measurement of the three-dimensional dimensions of small workpieces from multiple perspectives.

The invention relates to the technical field of computer vision, e-commerce and artificial intelligence, and discloses a computer vision-based online clothing size intelligent matching and fitting degree prediction method, which comprises the following steps of: firstly, reconstructing a high-precision three-dimensional human body model by processing an image uploaded by a user; meanwhile, three-dimensional sizes and physical attributes are deconstructed from a clothing commodity graph, then, a physical simulation engine is utilized to dynamically wear a clothing model on a user model for virtual fitting, the wearing effect is truly restored by simulating physical effects such as gravity and collision, quantitative fitting indexes such as tightness and oppression are calculated after simulation, and the fitting quality is improved. And based on multi-objective optimization, scoring different sizes, recommending an optimal size, and generating a visual report containing the 3D effect picture and the fitness thermodynamic diagram. According to the invention, automation and precision from the image to fitting degree prediction are realized, the problem of online clothes purchase size selection is effectively solved, the user experience is significantly improved, and the return rate is reduced.

The present disclosure relates to a system and method for automatically determining a three dimensional size of an item. The system may include an optical sensor scanning a surface of the item being transported on a conveyor spaced apart from the optical sensor by a conveyor distance, the surface including a standard size element having a first dimension. The system may also include a memory storing the first dimension and the conveyor distance. The system may also include a processor configured to calculate a second dimension of the surface based on a scanned image of the surface and a scanned image of the standard size element and the stored first dimension. The processor may obtain an item distance between the optical sensor and the surface and determine a height of the item based on the obtained item distance and the stored conveyor distance.

The invention provides a floor template accurate measurement method based on unmanned aerial vehicle photography. The method comprises the following steps: step 1, floor template scene image acquisition and image preprocessing; 2, sparse point clouds and dense point clouds of the floor template scene are generated; 3, extracting a building contour boundary, and segmenting and extracting a single template point cloud; and 4, calculating the three-dimensional size of the template based on plane fitting optimized by a sparrow search algorithm. According to the method, the 3D GS technology is adopted to overcome the reconstruction defect of a traditional MVS in a weak texture area, the problem of complex background interference is solved through multi-feature collaborative segmentation, adaptive adjustment of plane fitting parameters is achieved through an intelligent optimization algorithm, accurate measurement of the plane size and depth of the template is achieved, and the method is suitable for large-scale popularization and application. And the automation degree of floor template measurement is obviously improved.

The application provides a navigation method for a single-degree-of-freedom deformation mobile robot with known three-dimensional size relationship, which considers robot deformation obstacle avoidance; a three-dimensional environment is simplified into a two-dimensional grid map; different footprint sizes of the robot in different deformation states are considered when a cost map is inflated; an improved A* algorithm is adopted to calculate robot height state information according to special cost values of positions of path points, and a global path containing robot height information is obtained; an improved DWA algorithm is adopted to analyze robot height information of the global path, to judge a current motion state of the robot, to add speed space and motion trajectory space to a dynamic window, to set deformation speed sampling, and to analyze safe sizes of the robot in different states in obstacle cost, so that a moving speed and a deformation speed of the robot are obtained, and the robot is controlled to pass through an obstacle channel with height limitation by applying a deformation function.

The embodiment of the invention relates to an object position determination method and device, a controller, a vehicle and a program product. The method includes determining a first depth, a three-dimensional (3D) size, and a 3D center point of an object in a traffic environment based on a two-dimensional (2D) image corresponding to the traffic environment. The method further includes calculating a first position of the object based on the 3D size and the 3D center point of the object, and parameters of a camera capturing the image. The method further includes determining a second position of the object based on the first depth and a parameter of the camera. The method further includes determining a 3D position of the object based on the first position and the second position. In this way, the position of the object in the 3D space can be accurately and flexibly estimated based on the 2D visual information shot by the camera and the constraint of the geometrical relationship under the condition that the number of training data for the object detection model is insufficient or the quality is poor.

The invention discloses an intelligent furniture arrangement system and method for interior design, and aims to provide an efficient, visual and personalized home layout solution. The system collects data of room space and furniture through a data acquisition module, including but not limited to the size, shape and door and window position of a room and the three-dimensional size, color, material and other information of furniture. The data storage module is responsible for collecting and storing the core data in details and establishing a professional database. The data processing module can receive and analyze an instruction of a user, extract corresponding data from the database according to a user demand, and convert the data into a virtual graph. And the graph display module outputs the virtual graphs to a user terminal to allow the user to carry out personalized layout adjustment. The effect rendering module generates a highly realistic effect picture and optimizes details such as light, materials and colors. And the user interaction module realizes instant interaction with the user by using a natural language processing technology. And the artificial intelligence optimization module predicts preferences based on user habits and environmental information, automatically optimizes furniture placement, and comprehensively evaluates the scheme. Intelligent, personalized and efficient furniture arrangement is realized in the whole process, and the user experience and the arrangement efficiency are greatly improved.

The utility model discloses a non-filling karst cave edge detection device which comprises a supporting piece (2) and a detection piece (3) arranged on the supporting piece (2), a sleeve (1) is arranged outside the supporting piece (2), and the sleeve (1) is arranged in a hole in the ground; the method comprises the following steps: firstly, installing a sleeve (1), supporting a soil wall by using the sleeve (1) to protect equipment in the sleeve (1), inserting a support piece (2) and a detection piece (3) into the sleeve (1) after the sleeve (1) is installed, pushing the detection piece (3) by the support piece (2) to detect data in a karst cave, recording the data, and drawing a three-dimensional form of the karst cave by using the recorded data. The extending detection piece (3) is used for measuring the distances between the boundaries of the karst caves with different depths and different angles and the edge detection holes, the accurate three-dimensional size of the karst caves can be obtained, the arrangement number of the edge detection holes is reduced, and the edge detection work efficiency is improved; according to the more accurate three-dimensional size of the karst cave, the accurate processing work amount is obtained, and the processing effect is further improved.

The invention discloses a shoe manufacturing method and system, and the method comprises the steps: obtaining the foot three-dimensional size data of a target user, the foot three-dimensional size data of the target user comprises the foot length of the target user, the foot width of the target user, and the foot height of the target user; determining a target shoe tree corresponding to the foot three-dimensional size data based on the foot three-dimensional size data and a target shoe type selected by the target user; based on the foot three-dimensional size data and the three-dimensional size data of the target shoe tree, determining a target foot mold shoe tree corresponding to the foot three-dimensional size data; determining shoe parts related to the target foot mold last; in response to the target shoe component selected by the target user, generating an order for making the target shoe; and splitting the order to obtain a plurality of sub-orders, and distributing the plurality of sub-orders to manufacturers corresponding to the target shoe parts. According to the scheme, foot-fitting and personalized services can be provided for the target user, and more foot-fitting shoes can be provided for the target user.

The invention belongs to the technical field of target detection, and provides a monocular three-dimensional target detection method and system based on a nested reasoning chain, and the method comprises the steps: obtaining a monocular image, and extracting feature representation based on the monocular image; the method comprises the following steps: constructing a multi-stage nested reasoning chain based on feature representation, sequentially predicting a three-dimensional size parameter, a space orientation parameter and a space depth parameter of a target step by step according to a preset geometric dependency relationship, and dynamically aggregating intermediate features generated in each reasoning stage in the multi-stage nested reasoning chain to obtain a multi-stage nested reasoning chain; the fused feature representation is obtained; combining the fused feature representation with the original context information of the monocular image by using an error correction mechanism, and correcting an intermediate prediction result to suppress error propagation; and obtaining a three-dimensional detection result of the target based on the corrected prediction result. According to the method, the geometric consistency, accuracy and robustness of the monocular three-dimensional target detection result are improved.

The invention discloses an intelligent material scattering method and an intelligent material scattering system, and belongs to the technical field of material treatment. The material scattering method comprises the steps that S10, material data are collected, wherein the material data comprise the maximum size L in the three-dimensional sizes of materials; s20, based on the maximum size L in the step S10, the distance between the storage bin and the first vibration bin is controlled to be D, and 1.2 L is smaller than or equal to Dlt; 1.5 L; s30, starting a first vibration machine on the basis of detecting that the materials exist in the first vibration bin; s40, starting a second vibration machine on the basis of detecting that the materials exist in the second vibration bin; and step S50, starting a conveyor belt motor on the basis of detecting that the materials exist on the conveyor belt, and keeping the channel smooth through the arrangement of the distance D, so that blockage is avoided, and the working efficiency is improved. According to the material dispersing method, the materials can be fully dispersed, the material subpackaging efficiency is improved, the production cost is further reduced, and the material dispersing method has great application prospects.

The invention provides a three-dimensional similar material simulation device and method.The three-dimensional similar material simulation device comprises an excavation simulation device used for simulating the excavation process of a coal mine; the main body structure comprises a test platform, a simulation material is accommodated on the test platform, and the simulation excavation device is buried in the simulation material; the shaping parts are arranged on the test platform, a containing cavity with an opening is defined by the multiple shaping parts, the containing cavity is configured to be used for applying boundary constraint to the simulation material so as to shape the simulation material, and the opening is used for allowing the simulation material to enter the containing cavity; wherein the multiple shaping parts are assembled into the wall face of the containing cavity in a vertical stacking and horizontal column forming mode, the height of the containing cavity is adjusted by adjusting the number of stacked layers of the shaping parts, the length and / or width of the containing cavity is adjusted by adjusting the number of columns of the shaping parts, and therefore the three-dimensional size of the containing cavity can be adjusted. In this way, the practicability and universality of the device are improved.

The present invention discloses an AI visual recognition method and system for identifying and correcting T-beam assembly gaps, and relates to the field of AI visual recognition technology. The AI ​​visual recognition method includes: S01, setting an assembly area and a pre-assembly area in a spatial area according to the T-beam assembly process, and obtaining basic dimension data and real-time image data of the T-beam, diaphragm and bridge deck. The present invention uses the AI ​​visual recognition method, based on the basic dimension data combined with the tracking of boundary feature points on the T-beam, diaphragm and bridge deck, so as to understand the spatial distance data of the T-beam, diaphragm and bridge deck. The spatial distance data can be obtained through the boundary feature points, without the need to process a large amount of point cloud data, and the three-dimensional size of the gap can be recognized through the spatial distance data. The processing process is simple, and the correction of the T-beam is achieved by controlling the movement of the boundary feature points and the boundary feature lines, thereby ensuring the convenience of calculation.

The invention relates to the technical field of three-dimensional size measurement of a battery pole plate flow channel, in particular to a three-dimensional size measurement method of a fuel battery pole plate flow channel, which comprises the following steps of: (1) performing three-dimensional scanning to obtain a point cloud p; (2) performing plane correction on the point cloud to obtain a point cloud; (3) converting the point cloud into a depth map D (x, y) and normalizing and enhancing the depth map D (x, y) into D '(x, y); the method comprises the steps of (1) obtaining a depth map D '(x, y), (2) enhancing the depth map D' (x, y), (3) carrying out flow channel / ridge segmentation on the enhanced depth map D '(x, y), (4) carrying out edge detection and section extraction to obtain a flow channel / ridge region, and (6) respectively calculating the width / depth and ridge width of the flow channel by combining the values of the flow channel / ridge region, and carrying out statistical output.

The invention discloses a laser measurement-based plate size detection device for road and bridge construction, relates to the technical field of road and bridge construction, and aims to solve the problems of strong manual dependence, insufficient intuition and poor collaboration with a construction system in traditional plate detection. The device comprises a mounting frame, a conveying mechanism, a positioning mechanism, a detection mechanism, a console, a laser detection assembly and an equipment control and data processing system, wherein the equipment control and data processing system is integrated with a 3D modeling adaptation module, a plate three-dimensional model can be constructed and optimized based on size parameters of laser measurement, visual interaction of the model is achieved through a console, and butt joint with external construction management software is supported. According to the device, through fusion of laser non-contact measurement, automatic positioning and synchronous conveying, 3D modeling and construction cooperation technologies, high-precision and high-efficiency detection of the plate size is achieved, meanwhile, the visual three-dimensional size display and construction system cooperation capability is provided, and full-process technical support is provided for road and bridge construction plate quality control.

The invention relates to an attitude and size estimation method, system, medium and device, and the method comprises the steps: building a consistent orientation spacecraft normalized coordinate space, obtaining an observation point cloud, mapping the observation point cloud to the spacecraft normalized coordinate space, and obtaining a normalized observation point cloud; respectively extracting geometric semantic features of the mapped observation point cloud and the geometric reference template; based on the fused and spliced geometric semantic features, learning a specific shape deformation residual error of the target to obtain a geometric reference template after shape deformation, and according to the symmetry features of the spacecraft structure, completing point cloud symmetric complementation by adopting a mirror image or rotation strategy, and reconstructing a complete three-dimensional geometric shape of the target; and constructing an attitude solving frame based on the complete three-dimensional geometrical shape, solving a rotation matrix and a translation vector of the target spacecraft by registering the geometrical reference template before and after deformation, and reversely recovering the three-dimensional size and the actual pose of the target in the real physical space by combining the complemented complete geometrical shape.

The application is suitable for the field of machine vision recognition technology, and provides a rock debris piece identification method in a tunnel excavation process, a terminal device and a medium. Point cloud data of a rock debris piece to be identified is collected, and the point cloud data is voxelized. A quick point feature histogram corresponding to each voxel is constructed, spatial position coordinate values, curvature values and cross kernels of the quick point feature histogram of the voxel are obtained, and the similarity between the voxel and other voxels is calculated. Super voxel segmentation is performed according to the similarity, and a plurality of super voxels are obtained. The plurality of super voxels are coarsely segmented. Local gradient information of each point data in the coarse segmentation result is calculated, and the domain radius thereof is updated based on the local gradient information, and a plurality of point cloud clusters are obtained. The bounding box algorithm is used to obtain a three-dimensional size long axis and a three-dimensional size short axis of each point cloud cluster. The rock debris piece to be identified is identified according to the three-dimensional size long axis and the three-dimensional size short axis. The application can improve the accuracy of rock debris piece identification in the excavation process.

1. Name of the product in this design: Line Laser 3D Camera. 2. Purpose of this design: For the three-dimensional measurement and defect detection of industrial parts. 3. The key design feature of this product is its shape. 4. The image or photograph that best illustrates the design's key points: a 3D model.

A system is provided.SOLUTION: A system comprising: measuring means for measuring an object in three dimensions; transmitting means for transmitting measurement data acquired by the measuring means; generating means for generating three dimensional data based on the received measurement data; and providing means for providing the generated three dimensional data to a user.SELECTED DRAWING: Figure 1

The utility model discloses a three -dimensional measuring apparatu, it includes X regulation subassembly, Y regulation subassembly and Z regulation subassembly, X regulation subassembly includes X axle, and X axle has the X scale along the axial arrangement, Y regulation subassembly includes Y axle and Y connector, and Y connector one end slide setting in X axle, and its other end slide setting in Y axle, and Y axle is perpendicular to X axle, and Y axle has the Y scale along the axial arrangement. Z regulation subassembly includes Z axle and Z connector, and Z connector one end slide setting in X axle or Y axle, and Z axle slide setting in the other end of Z connector, and Z axle is perpendicular to X axle and Y axle, and Z axle has the Z scale along the axial arrangement. Utilize above -mentioned three -dimensional measuring apparatu can detect each position of three -dimensional coordinates of work piece to be measured, and then directly obtains the three -dimensional size of work piece.

The invention belongs to the field of water cannon encapsulation, and particularly relates to a water cannon encapsulation process which comprises the following steps: material pretreatment: when a GRG material is prepared, at least one functional material of a fireproof flame retardant and a sound insulation material is proportionally added into the GRG material according to required functions, and the materials are fully stirred and uniformly mixed; according to the invention, functional materials such as a fireproof flame retardant, a sound insulation material and the like are added into the GRG material, so that the water cannon encapsulation not only has a decorative effect, but also has practical functions such as fire prevention and sound insulation. The fireproof flame retardant can inhibit combustion when a fire occurs, so that the risk of the fire is reduced; the sound insulation material can effectively obstruct sound transmission, create a quiet atmosphere for the surrounding environment, and meet the use requirements of various scenes. Through a three-dimensional laser scanner and computer aided design software, three-dimensional size data of a water cannon installation site are obtained, and an encapsulation model matched with the three-dimensional size data is generated.

The invention provides a dispensing method. The dispensing method comprises the steps that a dispensing area of a workpiece is determined; measuring three-dimensional size coordinates of the dispensing area, and generating a color point cloud picture of the dispensing area; flexibly fitting a three-dimensional space plane of the dispensing area based on the three-dimensional size coordinates of the dispensing area and the color point cloud picture; calculating the three-dimensional fitting average size of the dispensing area based on the three-dimensional space plane of the dispensing area; according to the three-dimensional fitting average size of the dispensing area, the glue path height of the dispensing area is determined, and the glue path width of the dispensing area is calculated; based on the glue path height and the glue path width of the glue dispensing area, the glue weight and the glue dispensing speed are determined; and dispensing the dispensing area of the workpiece based on the glue weight and the dispensing speed. According to the adhesive dispensing method, different adhesive path heights and adhesive path widths can be obtained for different adhesive dispensing areas, so that the phenomena of adhesive overflow and adhesive shortage of the adhesive dispensing areas of the workpiece after adhesive dispensing are avoided, and the adhesive dispensing precision is improved. The invention further provides a dispensing system.

The invention relates to a method and system for inverting the three-dimensional size of a stratum fracture based on water attack data, and the method comprises the steps: introducing a fracture closing pressure drop equation for describing the internal pressure attenuation of the fracture, outputting the fracture closing pressure drop equation, and enabling the fracture closing pressure drop equation to serve as a time-varying boundary condition of a water attack dynamic equation set for describing pressure fluctuation, thereby constructing a coupling model; and an input variable in the coupling model is automatically adjusted by using a global optimization algorithm, so that a simulated pressure curve output by the model and actually measured water attack data are optimally fitted, and finally, the three-dimensional size of the crack is analyzed from optimal parameters. According to the method, the crack closing pressure drop equation is introduced into the coupling model to reflect the influence on the pressure drop in the crack dynamic closing process, so that the simulation of the pressure drop in the water attack process can be realized under the condition that the friction coefficient is not artificially increased or a new independent variable is not introduced. According to the method, the three-dimensional size of the stratum fracture after hydrofracture can be quickly inverted, and the hydrofracture evaluation and diagnosis method with low cost is provided.

The invention relates to the technical field of computers, in particular to a model projection method and system, electronic equipment, a storage medium and a program product. The method comprises the steps that a projection light source is monitored, current light source information is obtained, and the current light source information comprises the light source position and the light source angle; determining the projection size of the target model according to the current light source information, the floor position and the three-dimensional size and the model position of the target model; and creating a floor plane corresponding to the projection size in the target scene so as to project the target model on the floor plane. By adopting the scheme, the performance consumption of the model projection can be reduced on the basis of completely displaying the model projection.

The application discloses a cable icing detection method and device based on a laser radar, equipment and a medium. For two frames of point clouds to be registered, key points are extracted from each frame of point clouds to obtain key point clouds, the two frames of key point clouds are input into a registration network for processing, the registration network adaptively learns an overlapping area of the two frames of point clouds based on a spatial attention mechanism, predicts a relative pose of the two frames of point clouds, registers the two frames of point clouds based on the relative pose, obtains fused point clouds, extracts icing cable point clouds representing icing cables from the fused point clouds, determines three-dimensional sizes of the icing cables based on the icing cable point clouds, and calculates icing thickness based on the three-dimensional sizes of the icing cables. The attention weight mechanism enables the network to adaptively learn the overlapping area of continuous point cloud frames, thereby realizing fast and effective end-to-end registration of large-scale point cloud scenes, providing a good foundation for a fine registration algorithm, improving registration efficiency and accuracy, and further improving icing detection accuracy.

The invention provides an aero-engine component key defect three-dimensional size quantification method and system, and the method comprises the steps: collecting a video sequence when a monocular endoscope moves in an engine, inputting the video sequence with continuous frames into a depth estimation network and a pose estimation network, obtaining a single-frame image depth map and inter-frame pose transformation respectively; calculating luminosity re-projection loss according to the single-frame image depth map and inter-frame pose transformation; deriving a surface normal graph based on the depth map of the single-frame image, generating a composite image through a differentiable physical renderer in combination with the predicted reflectivity graph and light source parameters, and calculating physical rendering loss; and taking the weighted sum of the luminosity reprojection loss and the physical rendering loss as a total loss function, and performing forward propagation on a new input image based on the trained depth estimation network to obtain a depth map. According to the method, a real 3D label is not needed, strong reflection and illumination interference can be overcome, and high-precision three-dimensional size quantification can be realized only through a single hole detector video stream.