515 results about "Speckle pattern" patented technology

The invention provides an intelligent real-time detection method and system for PCCP welding quality, and relates to the technical field of online detection and intelligent evaluation of pipeline welding quality through machine learning. Light energy data and multi-light-source images of a spiral weld pool are collected, exposure parameters are dynamically adjusted through the energy difference of visible light near-infrared bands, and the real-time detection of the PCCP welding quality is achieved. Inhibiting strong light interference and generating a weld surface image; a stress concentration area is positioned by scanning a welding seam thermal deformation area and combining speckle pattern change, sound frequency change and the elastic characteristic of the thin-wall steel cylinder; inputting the surface image and the deformation data into a space-time convolutional neural network, fusing light energy change, image details and spatial features to construct a weld joint space structure diagram, and adaptively correcting the position of a sensor; and comparing the sinking depth of the three-dimensional point cloud reconstruction, analyzing the correlation between the sinking degree and the stress, and generating a probability thermodynamic diagram to output the pressure-bearing failure risk level, so that the probabilistic early warning of the pressure-bearing failure risk can be realized.

A dynamic adaptive depth camera occlusion detection method, system and device, and a storage medium, the method comprising: step T1: constructing a three-frame time sequence analysis window, continuously obtaining infrared speckle image sequences at different moments and corresponding dense depth map sequences, and establishing a pixel-level space-time alignment mapping relationship; t2, calculating a motion vector between adjacent frames, and executing the step T3 when the average motion intensity exceeds a dynamic threshold value; wherein the dynamic threshold is determined by the dense depth map; t3, positioning a speckle feature point in the infrared speckle image, generating an analysis window by taking the speckle feature point as a center, calculating a gradient change rate, and when the gradient change rate is smaller than a gradient threshold value, marking the region as a primary candidate region; and T4, carrying out space-time consistency comparison on the primary candidate region and two adjacent frames, and reserving a shielding region of which the variation value is smaller than a variation threshold value in three continuous frames. The method is high in precision, good in real-time performance and high in stability.

The invention provides a surface roughness measuring method and device based on a multispectral laser speckle image, and belongs to the field of surface quality visual inspection. Comprising the following steps: respectively acquiring speckle images formed by irradiating the surface of a workpiece with lasers of N wavelengths at different angles, and constructing a multispectral laser speckle data set; the speckle images of N wavelengths at a single angle are spliced according to channels and subjected to feature fusion, and then a multispectral laser speckle fusion image is generated; the fused images of the multiple angles form an image sequence, and single images in the sequence are subjected to multi-scale feature extraction and fusion to generate single-angle features; the multi-angle features generate global illumination information through a multi-angle feature aggregation module, and finally a roughness value is calculated through a multi-layer perceptron. Multi-spectral laser speckle coding surface features are adopted, multi-angle information is fused through the multi-angle feature aggregation neural network, and the relation between the light source wavelength, the incident angle and the surface three-dimensional shape and roughness is comprehensively considered, so that high-precision measurement is achieved.

The invention discloses a method for testing mechanical properties of a carbon fiber reinforced composite material based on image recognition, and relates to the technical field of material tests.The method comprises the steps that a carbon fiber sample of a fiber laying layer is prepared, the width tolerance is controlled through laser cutting, and speckle patterns are sprayed on the surface; training a U-Net + + neural network based on the collected macro / micro image, and analyzing the dynamic change of the fiber angle in real time; fusing the acquired macro / micro image with the dynamic change of the fiber angle, and constructing a constitutive equation of dynamic fiber orientation distribution; and importing the constitutive equation into a finite element model, iteratively optimizing the fiber-matrix interface strength parameters through inversion calculation, performing blind test verification on the sample by using the optimized finite element model, and outputting a cross-scale correlation test report. According to the method, an interface strength parameter is subjected to inversion iteration optimization through a finite element model, a macro-micro strain field, a fiber slip angle and an interface damage parameter are integrated through a cross-scale correlation test report, and full-chain data support is provided for a composite material structure.

An optical speckle receiver for receiving a speckle signal from a sample, the optical speckle receiver comprising an optical detector and an aperture and / or lens array. The aperture and array respectively comprise a plurality of apertures or lenses and is located between the sample and the optical detector such that the received speckle pattern is obtained from multiple discrete sample locations.

The invention discloses a device and method for measuring dust concentration distribution at an inlet of an aero-engine, and belongs to the field of aero-engine sand swallowing tests, the device comprises a pulse laser light source, a shaping module, an optical image acquisition module and a data post-processing module, the pulse laser light source is arranged at a preset position around the aero-engine and is used for emitting laser beams, and the shaping module is used for shaping the laser beams; the aero-engine is located in a sand dust environment to be detected; the shaping module is used for receiving the laser beam and outputting a sheet-shaped beam; the optical image acquisition module is used for receiving coherent scattered light emitted by the grit in the irradiation area of the flaky light beam to form a laser speckle image, and transmitting the obtained laser speckle image to the data post-processing module; and the data post-processing module processes the laser speckle image by adopting an inlet sand dust concentration measurement algorithm to obtain the concentration distribution of the inlet sand dust of the aero-engine. The device is simple in structure, and is suitable for measuring the sand dust concentration distribution at the inlet in the sand swallowing experiments of aero-engines with different sizes or stages.

The invention provides an indoor crustal stress test device and method based on optical strain measurement, the test device comprises a stress simulation device, the stress simulation device can perform real three-dimensional crustal stress environment simulation on a test rock mass as far as possible; the drilling releasing device is provided with a hollow drill bit used for drilling the optical strain measurer installed in the test rock mass, and the drilling hole and the hollow drill bit are coaxially arranged. Thus, according to the method, pressure is applied to the test rock mass from multiple directions through the multiple hydraulic loading columns, the real three-dimensional crustal stress environment can be more accurately simulated, speckle image changes in the speckle coating can be continuously recorded in real time through the optical strain measurer, and hole wall strain data are obtained; and converting the strain data into stress by applying an elastic theoretical formula, and accurately calculating the stress state of the experimental rock mass.

The invention relates to the field of data transmission, in particular to a multimode optical fiber image transmission method, system and device and a medium. The method comprises the steps of acquiring an original image of a target area through image acquisition equipment of the power transmission line operation robot; converting the original image into an optical signal and transmitting the optical signal through a multimode optical fiber to generate a speckle image with spatial disturbance characteristics; inputting the speckle image into a neural network model constructed based on a sliding window self-attention mechanism for image reconstruction; the reconstructed image is output to a power transmission line state analysis system for real-time monitoring; the sliding window self-attention block calculates local self-attention features in windows with fixed sizes through window multi-head self-attention and sliding window multi-head self-attention mechanisms which are alternately arranged, and feature interaction between adjacent windows is achieved through a sliding window mechanism. According to the method, the anti-interference capability and the dynamic environment adaptability are improved, the calculation complexity is reduced, and the model generalization capability is improved.

The invention discloses a space-time speckle projection three-dimensional imaging method based on multi-frame optical flow alignment. Firstly, a projector based on DLP is used for projecting a space-time speckle pattern to a measured scene, and a binocular camera synchronously collects a three-dimensional space-time speckle image. The calibration parameters of the binocular camera are used to carry out stereo correction on an acquired original speckle image, and a parallax image is generated frame by frame in combination with a coarse-to-fine single-frame speckle matching strategy. And estimating a two-dimensional inter-frame displacement field between continuous disparity maps by using an optical flow method by taking an intermediate frame disparity map as a reference, compensating motion artifacts in a space-time speckle image, and ensuring strict space-time registration of a dynamic target. And based on the speckle image after motion correction, a speckle matching strategy is expanded to a time-space domain, and high-precision multi-frame three-dimensional measurement of a complex dynamic scene is realized. The method is suitable for performing rapid and high-precision three-dimensional modeling on a moving target in an unstructured environment, and can perform accurate three-dimensional measurement on a high-speed dynamic target undergoing any translation or rotation motion.

The invention discloses a power transmission tower deformation monitoring method and device and a computer program product, and the method comprises the steps: obtaining a multi-angle collection image of a random speckle pattern on the surface of a power transmission tower; constructing an image sequence based on the multi-angle acquired images, reconstructing a three-dimensional displacement field on the surface of the power transmission tower by using a stereoscopic vision reconstruction algorithm, and exporting a strain tensor field; performing time sequence processing on the strain tensor field under a plurality of time nodes through a strain evolution analysis model of a continuous time sequence, and extracting a strain change trend of each preset structure part; carrying out space consistency mapping and weighted interpolation processing according to the strain change trends of the plurality of preset structure parts, and calculating to obtain the overall structural dependent variable of the power transmission tower; and S5, in combination with historical strain data and the current overall structural dependent variable, identifying an abnormal deformation area of the power transmission tower by comparing the spatio-temporal evolution mode of the strain tensor field. The method has the effect of improving the accuracy of deformation monitoring of the power transmission tower.

The invention belongs to the field of material mechanical property testing and strain measurement, and particularly relates to a local strain measurement method and system based on a four-point bending and digital image correlation (DIC) technology, and the method specifically comprises the following steps: cutting to prepare a sample, carrying out surface electrolytic polishing, preparing speckles on the surface of the sample to be tested, and carrying out a four-point bending experiment; speckle images of the lower surface of the sample are collected, displacement correction is conducted in combination with bending deflection calculation, the bending load and the strain of the upper surface of the sample are obtained in real time, the stress-strain relation and local strain distribution of the lower surface of the sample are obtained through calculation, and finally deformation and failure mechanism analysis is conducted in combination with surface deformation and crack morphology. The method provided by the invention combines four-point bending and DIC plane strain measurement technologies, can realize stretching, compression and dynamic fatigue loading on the surface of the sample, is suitable for performance test and failure analysis of tension, compression and fatigue deformation of brittle metal materials such as intermetallic compounds and amorphous materials, and has the characteristics of convenience, rapidness and low cost.

The invention provides a tunnel point position change automatic observation method and system combined with an AI algorithm, and relates to the technical field of tunnel monitoring, and the method comprises the steps: employing a speckle projector to form an irregular speckle pattern on a tunnel wall surface, and carrying out the multi-angle collection through a polarized light camera array, and obtaining the speckle intensity data; pre-embedding a coding microstructure generated based on quantum state superposition in the speckle pattern; establishing a speckle-stress correlation analysis model according to the speckle intensity data, and calculating three-dimensional displacement data in combination with the deformation data; and inputting the environmental monitoring data, the stress distribution data and the displacement data into the self-evolution heterogeneous feature network for analysis and prediction. According to the method, the measurement precision is improved by adopting the quantum coding speckles, the dynamic correlation analysis of multi-source data is realized through the self-evolution heterogeneous feature network, and the environmental adaptability and prediction accuracy of the system are enhanced.

The invention relates to a method for dynamically measuring a composite material under an ultra-high-temperature working condition, which comprises the following steps of: constructing a composite material high-temperature strain dynamic detection platform, and respectively acquiring speckle images of a tested piece under a cold state condition and an ultra-high-temperature working condition as data sets; constructing a supervised network model for eliminating speckle image thermal radiation and thermal disturbance influence, wherein the supervised network model comprises a multi-scale feature extraction module, a dual convolutional neural network and a polarization self-attention mechanism; constructing a joint loss function L for the supervised image restoration network model; training the model to obtain an optimal image restoration network model; an ultra-high-temperature speckle image serves as input, a speckle image restoration result is output through the image restoration network model, and a displacement field is obtained through DIC calculation; according to the invention, the influence of thermal radiation, thermal disturbance and haze phenomena in an ultra-high temperature environment on image quality is effectively eliminated, and the dynamic strain measurement precision of the composite material under an ultra-high temperature working condition is remarkably improved.

The invention discloses a method for determining crack length based on a displacement field of an end face opening area, which comprises the following steps: spraying speckles on the end face of a CT test piece, applying cyclic load to the CT test piece by using a testing machine, and shooting a speckle image sequence of the end face of the test piece in the whole loading process at high frequency by using an image sensor; the speckle images are calculated, and opening displacement values under different cycles are obtained through changes of the speckle images in the crack propagation process of the test piece; and establishing a calculation model between the end face displacement field and the crack length, and calculating the crack length. A non-contact DIC optical measurement technology is adopted, an opening displacement field of the end face of a CT test piece is captured in real time, a conversion formula of opening displacement and crack length is established, the DIC optical measurement technology is used for measuring the full-field displacement field of the whole end face, the average displacement value of multiple sets of measurement points is used for replacing a single-point value, random errors are eliminated, and the CT test piece is obtained. And the adopted calculation model and formula are improved on the basis of a flexibility method formula, so that the precision is higher.

The invention provides a binocular vision point cloud temperature field rapid reconstruction method based on homography transformation. During calibration, the binocular camera and the thermal infrared imager shoot mark points synchronously; solving a homography matrix between the left camera and the thermal infrared imager according to a mark point visible light image obtained by the left camera and a corresponding point in the thermal image; and based on the homography matrix, resolving an external parameter matrix between the left camera and the thermal infrared imager. When a point cloud temperature field is reconstructed, a binocular camera obtains a surface speckle image of a measured object and calculates a three-dimensional point cloud on the surface of the measured object; updating a homography matrix by using plane features in the three-dimensional point cloud; and according to the internal reference matrix and the updated homography matrix, mapping the temperature information in the thermogram to a point cloud corresponding to the visible light coordinates of the left camera. According to the method, only one-time calibration measurement is needed, operation is easy, precision is high, and measurement of the three-dimensional morphology and the temperature of each point of the material with unobvious surface features is achieved.

The invention discloses a noise speckle real-time imaging method suitable for various complex scenes, and the method comprises the steps: firstly obtaining a speckle image of a sample target after a scattering medium through a camera, and carrying out the preprocessing of the speckle image; secondly, the principal component of the preprocessed speckle image is extracted by using VMD based on a row-column separation strategy; and then noise of the principal component is extracted through a main singular value, and an enhanced speckle image is obtained through differential operation with the speckle image. And finally, autocorrelation operation is carried out on the enhanced speckle image, the amplitude of a target is obtained through AM algorithm optimization, and speckle image reconstruction is realized by using a phase recovery algorithm. The method has better robustness and structure retention capability in an edge region and a high-noise scene, can overcome the influence of a strong light interference environment, and remarkably improves the local contrast of a speckle image.

The invention discloses a support deformation monitoring method combining cross laser and laser speckles, which comprises the following steps: selecting a monitoring point on the surface of a support, and projecting the cross laser and the laser speckles to the monitoring point, so that the cross point of the cross laser and the laser speckles are aligned with the center of the monitoring point; aligning the camera to the monitoring point, so that an included angle theta is formed between the shooting direction of the camera and the direction of projecting the cross laser; carrying out calibration setting and parameter setting on the camera to obtain a set camera; under different load working conditions of the bracket, controlling the camera to collect images of the monitoring points at regular time; wherein the blue speckle pattern is collected by a blue channel, and the red structured light pattern is collected by a red channel; processing and strain calculation are carried out on the collected images to obtain in-plane and out-of-plane displacement fields and strain fields of the support at different moments; according to the invention, high-precision, dynamic and more comprehensive multi-dimensional deformation monitoring of support structures can be realized.

The invention relates to a method and a system for realizing fast and uniform focusing of emergent light spots of a multi-mode optical fiber based on superpixels, belongs to the technical field of novel optical imaging and light field regulation and control, and integrates online speckle acquisition, offline amplitude optimization and an online speckle acquisition stage. When a certain modulation area is in an activated state, respectively capturing a non-interference speckle image of the emergent end of the multimode optical fiber and an interference speckle image generated when the area and the reference area are activated at the same time, and traversing all the modulation areas to totally obtain a speckle image; in the off-line amplitude optimization stage, a focusing light spot is preset, and calculation is carried out by utilizing intensity information of a corresponding focusing position in the speckle image. And analyzing all the focusing light spots one by one so as to construct a binary amplitude modulation mask. And the masks are loaded to the spatial light modulator one by one, so that accurate point-by-point focusing of N emergent light spots of the multimode optical fiber is realized, and the focusing efficiency and uniformity are greatly improved.

According to an aspect there is provided a device for analyzing a speckle pattern from biological tissue, the device comprising:a light source configured to generate light for illuminating the biological tissue, such that the light is scattered by the biological tissue and forms the speckle pattern;a Fourier transform element configured to collect the light scattered by the biological tissue and to transform the collected light into a spatial Fourier domain at a Fourier plane of the Fourier transform element;a spatial bandpass filter arranged at the Fourier plane, the spatial bandpass filter being configured to selectively transmit light at a spatial position of the Fourier plane, the spatial position corresponding to a band of speckle sizes of the speckle pattern;a photosensitive detector configured to detect the light transmitted though the bandpass filter, wherein a signal output from the photosensitive detector comprises information on contrast of the speckle pattern.

The invention provides a three-dimensional calibration method using a speckle image as a calibration object and combining weight parameters. The method comprises the following steps: acquiring a standard grid image and the speckle image through a camera; determining angular point positions in the standard grid image through an angular point detection algorithm; calibrating the stereoscopic vision system by using the obtained angular points to obtain an initial calibration result; calculating 3D coordinates of the camera calibration board in the world by using the disparity map; determining a subset of the reference image corresponding to the checkerboard angular points in the target image through a digital image cross-correlation algorithm; calculating a weight parameter of each calibration point based on the re-projection error; solving internal and external parameters of the camera by using the weight parameters, radial alignment constraint and a least square method; optimizing the main point of the camera by using the weight parameters; and if the re-projection error or the number of iterations does not meet the requirement, repeatedly calculating the weight parameter of the calibration point and the subsequent steps, otherwise, obtaining a camera calibration result, and obtaining the distortion curved surface of the camera lens.

The invention discloses a computational imaging method and system based on multimode optical fiber mode modulation, and the method comprises the steps: designing and building a multimode optical fiber image transmission system, and collecting a natural scene speckle image data set under different transmission conditions through replacing multimode optical fibers modulated in different modes; an optimized network architecture based on a UNET network is constructed, and a deep learning network adopting a series double-subnet structure is constructed for performing feature extraction and image reconstruction from a speckle image to an SLM image and from the SLM image to a real image, and learning complex speckle image features; training a deep learning network based on natural scene speckle image data sets in different mode modulation states, and introducing a joint loss function to optimize the reconstruction performance of the network; the speckle image verification model under different transmission conditions is utilized to reconstruct the quality of the image; according to the invention, the target information in the speckle image is enhanced through mode modulation, and high-quality image reconstruction is realized by using the deep learning network.

A system is provided. The system includes a light source configured to emit a probing beam to illuminate an optical element. The system also includes an image sensor configured to be rotatable around the optical element within a predetermined rotation range. The system also includes a controller configured to control the image senor to move to a plurality of angular sub-ranges of the predetermined rotation range to receive a plurality of scattered beams output from the optical element. The image sensor is configured to generate a plurality of sets of speckle pattern image data based on the received scattered beams. The sets of speckle pattern image data provide two-dimensional (“2D”) spatial information of speckles.

The invention discloses a deformation measuring device and method based on structured light illumination digital speckle interferometry, the device comprises a structured light generation and phase shift unit and an imaging unit, the structured light generation and phase shift unit comprises a laser, a spatial light modulator and a spatial mask plate, the spatial light modulator and the spatial mask plate are arranged on a light path of the laser, the laser is used for generating laser, and the spatial light modulator is used for generating light; the spatial light modulator is used for loading stripe modulation patterns to form multiple beams of diffracted light propagating in different directions and realize phase shift operation, and the spatial mask plate is used for only selecting 0-level and + 1-level diffracted light beams for interference to generate structural illumination light; the imaging unit is used for recording a digital speckle image of the sample. According to the invention, the fringe modulation pattern is loaded on the spatial light modulator, the 0-level diffraction beam and the + 1-level diffraction beam are selected for interference in combination with the spatial mask plate to generate rapid structured light illumination and phase shift, the imaging speed is high, the phase shift repetition precision is high, and the accuracy of a reconstruction result is ensured.

The invention discloses an optical fiber coupling real-time correction method in a single-mode and multi-mode optical fiber endoscopic imaging system, which comprises the following steps of: firstly, manually adjusting reflectors M1 and M2 to an optimal multi-mode optical fiber coupling state, acquiring images of light spots at the near end and the far end of a laser beam in the state by using a CCD (charge coupled device) 1 and a CCD 2, and calculating the position of a central coordinate of the images; and collecting a speckle image formed by the light spot through the scattering medium by using a CCD3 (charge coupled device 3). The system monitors the center coordinates of near-end and far-end light spots at the current moment and the correlation coefficient of the current speckle image and the collimation state image in real time. The piezoelectric mirror bracket is fed back and adjusted in real time through the center coordinates of the two spot light spots and the correlation coefficient of the speckles, so that the correlation coefficient of the speckles is kept above a set threshold value, and automatic calibration of light beams is achieved. On the basis of the traditional thinking that two points determine one straight line, the correlation coefficient of using speckles is increased, the alignment precision is improved, and the method has certain practicability and application prospects.

The invention belongs to the technical field of three-dimensional vision measurement, and relates to a deformation measurement method and system of a composite vision measurement unit. The internal and external parameters of the binocular measurement unit are obtained according to the image of the thrust vectoring nozzle target area calibration plate; obtaining coordinates of all binocular measurement units in the same reference coordinate system according to the internal and external parameters and the images of the thrust vectoring nozzles with the first coded mark points, and obtaining coordinates of all binocular measurement units in a unified coordinate system; and acquiring full-field deformation of all target areas of the thrust vectoring nozzle according to the speckle images of the target areas of the thrust vectoring nozzle in the working state and the coordinates of the unified coordinate system of all binocular measurement units. According to the invention, a digital image correlation method and a close-range photogrammetry technology are utilized to realize coordinate system unification and full-field deformation measurement. Compared with a traditional method, the measurement times and data processing steps are reduced, and the measurement efficiency is improved.

A single-frame speckle deflection measurement method based on subarea wavefront correction belongs to the technical field of optical measurement, and comprises the following steps: based on a monocular deflection measurement system, projecting a standard speckle image to a to-be-measured mirror surface from a screen, and collecting a single-frame deformed speckle image reflected by the to-be-measured mirror surface; processing the reflected image by using a Zernike polynomial projection distortion correction method based on a mark point, and correcting the projection distortion of the reflected image caused by a non-orthogonal light path; obtaining sub-pixel mapping of the reflection image and the standard image based on a sub-region wavefront fitting correction and iterative registration algorithm; and based on system calibration and camera calibration, establishing a spatial incident light model and a spatial reflected light model, obtaining surface gradient distribution of the to-be-measured mirror surface, and reconstructing a three-dimensional shape of the to-be-measured mirror surface. According to the method, a reference plane is not needed, high-reflection surface shape measurement of only collecting a single-frame image can be achieved, the dynamic range and precision of speckle deflection measurement are improved through subarea wavefront correction, and the surface reconstruction precision can reach hundreds of nanometers.

The invention discloses a multifunctional asphalt mixture rut test method which is implemented by using a device integrating loading and measurement, test piece supporting and temperature control drainage, environment simulation and optical measurement and a central control and data acquisition system. A test piece is arranged in the test piece groove, temperature control is realized through circulating a heat-conducting medium in the temperature-conducting cavity, different gradients are formed through the hoisting mechanism and the angle adjusting assembly, the test wheel is driven to perform one-way circulating loading along a closed track and snub at a preset moment, and rolling and sliding friction forces are synchronously collected; in the test process, rainfall is simulated by adopting the watering assembly, incoming water is recycled by the drainage tank, speckle images are acquired by the high-frequency industrial camera, and surface full-field displacement and strain are obtained through digital image correlation analysis. Through time registration and fusion of temperature, gradient, wheel speed, rainfall and friction-deformation data, rut deformation, skid resistance and apparent structure evolution laws of the asphalt mixture under multi-climate and multi-gradient conditions are obtained.

The invention relates to the technical field of image processing, and discloses a speckle structured light binocular stereo matching method and system based on multi-cost fusion, a computer readable storage medium and a computer program product. The method comprises the following steps: respectively carrying out epipolar correction on an obtained left camera image and an obtained right camera image; performing boundary segmentation on the corrected left camera image to obtain a corresponding boundary image; calculating a Census cost and a BT cost by using the corrected left camera image and right camera image, and carrying out weighted fusion to obtain a matched cost body; performing cost aggregation on the obtained cost body by using the boundary image as prior information; and calculating a parallax image corresponding to the left camera image through the aggregated cost body. According to the invention, abnormal matching points in a speckle image stereo matching process are reduced, and the real-time performance and robustness of speckle image stereo matching are effectively improved.

The invention discloses a laser shearing speckle interference phase unwrapping method based on deep learning. According to the method, eight speckle patterns before and after deformation are used as input, multi-scale feature extraction is carried out through layer-by-layer convolution and down-sampling of an encoder, features are sent to an attention fusion module, response weights of channels and spatial positions are adaptively adjusted through channel attention and position attention, then the response weights are input into a decoder, and jump connection with an encoding end is combined, so that multi-scale feature fusion is realized. And step-by-step reconstruction of multi-scale features is realized, and a phase diagram is output. Performing comprehensive constraint on prediction and reference phases in the aspects of numerical deviation, structural consistency and gradient smoothness and updating network parameters by adopting composite loss formed by mean square error, mean absolute error, structural similarity, gradient loss and out-of-plane displacement calculation items; therefore, an unwrapping result which is globally continuous and has clear and stable local details is obtained, and a feasible way is provided for intelligent processing and automatic analysis of the laser shearing speckle interference image.

The invention discloses a multi-particle three-dimensional shape reconstruction method based on projection consistency. According to the method, a three-angle laser interference particle imaging system is used for collecting out-of-focus images of three visual angles, a center M * N pixel range is cut to obtain an out-of-focus speckle image, phase recovery algorithm reconstruction is carried out, a two-dimensional particle shape is obtained, a gray threshold value above 70 is set to be a target contour area, binarization segmentation is carried out, and a target contour image is obtained. And smoothing the target contour area. And carrying out three-dimensional space projection on contours of the four candidate results in Matlab, generating a three-dimensional structure in an intersection region, re-projecting the three-dimensional structure to each view angle, calculating an error S between a re-projected image and an original input projection, and selecting a final reconstruction result with the minimum S. According to the method, through systematic iterative optimization and an error evaluation mechanism, trivial fuzzy interference in a phase recovery process is effectively eliminated, and high-precision reconstruction of a multi-particle three-dimensional shape is realized.