65 results about "Tomographic reconstruction" patented technology

The invention belongs to the field of CT (Computed Tomography) tomography reconstruction technology and artificial intelligence, and discloses a finite angle CT reconstruction method based on combination of a three-dimensional conditional diffusion model and synchronous iteration. The CT is an imaging technology which utilizes X-rays to irradiate a target from different angles and acquire projection, and obtains an internal three-dimensional structure through reconstruction. Different from traditional CT depending on nearly full-angle scanning, the method only collects limited-angle projection, and achieves fault reconstruction under the limited-angle condition through a three-dimensional condition diffusion model of space domain-frequency domain two-way decoding and by means of structural generality priori of a workpiece. Meanwhile, projection and fault data consistency correction is carried out in combination with a synchronous iteration reconstruction technology, the advantages of an iteration method in the aspect of physical mechanism characterization are exerted, interpretable physical constraints are provided for a deep learning network, and therefore the reliability and precision of a reconstruction result are improved.

The invention relates to the technical field of CT detection, and discloses a plane CT fan-beam projection filtering back projection reconstruction method and CT detection device.The method comprises the steps that S1, tilt geometric configuration parameters are determined according to the distance from a detector to a rotation center and a tilt angle, a sample rotating table is controlled to rotate, and the detector is triggered at each rotation angle to collect first projection data; s2, calculating a global three-dimensional coordinate of a detector pixel point under each rotation angle, and generating second projection data; s3, performing frequency domain filtering according to the thickness of the sample and the second projection data to obtain third projection data; and S4, calculating a ray intersection point coordinate from a reconstructed slice pixel point to a detector plane, and reconstructing the third projection data according to the ray intersection point coordinate to obtain a tomographic reconstruction image.According to the method, high-frequency detail features are reserved while noise is suppressed, the problem that a traditional fixed filtering kernel parameter is poor in adaptability to different samples is avoided, and the method is suitable for being applied to the field of image reconstruction. Therefore, high-precision three-dimensional imaging is realized.

The invention discloses a self-supervised tomographic SAR (Synthetic Aperture Radar) reconstruction method based on orthogonal subspace decomposition. According to the method, a traditional single original space is reconstructed and expanded into a Range-Null-Raw triple space collaborative learning framework, an R2R and N2N principle-based self-supervised loss function is constructed by using an orthogonal subspace decomposition characteristic of a linear measurement operator, direct mapping learning from noise measurement data to a high-quality reconstruction result is realized, and the real-time performance of a real-time reconstruction system is greatly improved, so that the real-time performance of the real-time performance of the real-time performance of the real-time performance of the real-time performance of the real-time performance of the real-time performance is improved. Observable components are learned through Range space, unobservable components are learned through Null space, a cross constraint strategy of global consistency is ensured through Raw space, and the adaptability of the model to real measurement conditions is enhanced. The performance gap between simulation training and actual deployment in deep learning TomoSAR reconstruction is effectively solved, and a new solution is provided for reliable application of an unsupervised deep reconstruction network in a real TomoSAR imaging task.

The invention relates to the technical field of marine cable monitoring, in particular to a marine cable life prediction and health management method and system based on multi-source data fusion, and the method comprises the steps: constructing an electrochemical impedance spectroscopy and optical fiber humidity sensor array, and synchronously collecting multi-source signals in a coupling environment; performing dynamic compensation on the signal by combining a marine environment and working condition parameters, and eliminating interference to generate a state feature set; constructing a space corrosion-aging distribution diagram by using an electro-chemical coupling field chromatography reconstruction algorithm; real defects are determined through space-time adaptive segmentation and defect dynamics screening; and finally, calculating the corrosion depth and the aging rate by adopting a hyperbolic tangent-power law attenuation combined model to realize risk assessment. According to the invention, the accuracy of defect identification and life prediction of the marine cable in a complex environment is improved.

The invention provides a carbon fiber three-dimensional density measurement method and system, and relates to the technical field of material detection. Comprising the following steps: irradiating a to-be-detected sample with X-rays, and collecting a plurality of interference images formed after the X-rays penetrate through the to-be-detected sample at different angles of the to-be-detected sample; calculating two-dimensional density distribution of the to-be-detected sample at each angle according to the interference image; and carrying out tomographic reconstruction on the two-dimensional density distribution at the plurality of angles to obtain a three-dimensional density distribution image of the sample to be detected. According to the carbon fiber three-dimensional density measurement method, a multi-angle X-ray interference image is collected, and the density change of the carbon fiber reinforced composite material is reflected by using phase information, so that the density resolution of the carbon fiber reinforced composite material is improved; the method is non-destructive, and is suitable for efficient and reliable analysis of the internal structure of a complex material.

The present application belongs to the field of CT tomographic reconstruction technology and artificial intelligence, and discloses a limited-angle CT reconstruction artifact suppression method based on a multi-domain feature fusion network. In view of the problem that the reconstruction result of traditional CT scanning under limited-angle conditions is prone to artifacts and structural distortion, thereby affecting the image quality and defect detection accuracy, the present application constructs a multi-domain feature fusion artifact suppression network, takes the limited-angle reconstruction result as input, and realizes artifact suppression and detail recovery through the synergistic effect of the encoder part, the decoder part, the feature enhancement part and the feature conversion part. The present application can obtain high-quality tomographic images under limited-angle conditions, effectively reduces the scanning angle and time of industrial CT detection, improves the imaging clarity and reliability without increasing the radiation dose, is suitable for industrial detection of complex structure workpieces, and has important industrial application value.

The present application relates to the field of optical sensing technology, and provides a transparent tomographic imaging method based on an edge convex lens array, which receives an incident light source at a first end face to generate a first light signal, and the first light signal is transmitted along the first end face to a circumferential side face. The first light signal is encoded by a convex lens array structure arranged at the circumferential side face, and at least one group of second light signals is obtained after encoding. The second light signal is sensed by at least one optical detection module. The optical detection module transmits the second light signal to a signal reconstruction module for tomographic reconstruction to obtain a third light signal. The present application also discloses a sensing device. The method and the sensing device can transfer the sampling and encoding process of the light signal from a plane to an edge region of an imaging plane, realize angle-related optical encoding of the in-plane propagating light signal, and obtain an imaging result through tomographic reconstruction. Thus, the technical problems of limited transparency and limited resolution in the existing transparent imaging scheme are overcome.

The application discloses a kind of turbulent flame dynamic three-dimensional structure measuring device and method.The measuring device includes: multi-ultraviolet camera imaging system, for capturing the projection image of turbulent flame OH* free radical radiation luminescence in different directions;Optical fixing system, for fixing camera and adjusting the height and angle of viewing angle;A synchronous controller is used to realize the synchronous acquisition of transient flame image;A computer is used to store image and calculate analysis.The measuring method includes: multi-ultraviolet camera imaging system calibration;Shooting and correcting turbulent flame transient image;According to the corrected flame image, the weight coefficient of the target space discrete voxel is calculated using voxel preprocessing technology;Finally, using tomographic reconstruction algorithm, the three-dimensional structure distribution of turbulent flame in target space is reconstructed.The application constructs chemiluminescence tomography system based on multi-ultraviolet camera, realizes the acquisition of low-cost, high-accuracy turbulent flame transient front chemiluminescence information;In the reconstruction process, the calculation amount of weight coefficient is reduced using voxel preprocessing technology, and the measurement accuracy is improved.

The invention discloses an air inlet channel flow field panoramic measurement system based on multiple cameras and background schlieren. The measurement system comprises a multi-camera imaging module, a pulse illumination module, a synchronization system and a computing platform. And a plurality of cameras in the multi-camera imaging module synchronously collect background plate images and output the background plate images to the computing platform. And the calculation platform performs cross-correlation calculation on the two frames of background images with / without flow field disturbance to obtain a displacement field, and performs two-dimensional flow field chromatography reconstruction on the multi-camera view angle displacement field based on the calibrated imaging parameters to obtain a density gradient field of the flow field in the binary supersonic air inlet. A miniature background schlieren system is adopted to replace a single-point measurement pressure measurement system, large-view two-dimensional reconstruction of a flow field in an air inlet channel is achieved, the system is simple in structure, high in reliability and high in expansibility, and a plurality of cameras are arranged in an optical access channel so that measurement of flow in a larger range of the flow field can be achieved.

The invention relates to the technical field of X-ray imaging, and discloses a tomography device and method based on a multi-focus X-ray source, and the device comprises a control system which is used for controlling the multi-focus X-ray source to carry out the longitudinal scanning of a detected object; n ray tubes which are transversely distributed are arranged on the multi-focus X-ray source, each ray tube internally comprises an independent cathode and an independent anode, and the cathode and the anode are used for triggering and generating multi-view X-rays; the detector is used for receiving the X-rays penetrating through the detected object and converting the X-rays into multi-view projection data; and the image processing system is connected with the detector and is used for processing the multi-view projection data based on a tomography reconstruction algorithm to obtain tomography data of the detected object and a three-dimensional reconstruction model. The multi-focus ray source and the detector which are horizontally and transversely arranged synchronously move for scanning, multi-view-angle and high-efficiency whole-body tomography data acquisition is achieved, and the requirement for rapid, accurate and whole-body tomography imaging is met in combination with a tomography reconstruction algorithm.

The application discloses a fluorescence tomography reconstruction imaging method based on a physical model depth network, acquires tissue surface light intensity information through a CCD camera, performs mesh partitioning on a phantom by using Amira software, respectively partitions two meshes with different precisions, respectively solves corresponding system matrices and adjacent matrices of mesh nodes by using the mesh after partitioning and optical parameters of each region in the phantom through a finite element method, sets a transfer matrix corresponding to the power density of different meshes, then transmits surface light intensity data into a depth network, solves in a coarse mesh, and demarcates a feasible area for solving a fine mesh according to the solving result of the coarse mesh, and solves in the feasible area, so that the norm of the obtained solution is small, the result is relatively sparse, the error of the norm and the result forward and the real surface light intensity is considered during solving, the position of a fluorescent group in the tissue can be positioned more accurately, and the application has important significance for studying tumors and early diagnosis of tumors.

This application relates to the field of optical nondestructive testing technology and discloses a real-time detection and classification system for polyester filament defects that integrates machine vision. The system includes: a ring-shaped polarization time-domain dual-encoding light source array for generating a dynamic structured light field containing slow-varying polarization encoding and high-speed time-domain orthogonal encoding; a ring-shaped photoelectric sensor array for acquiring the applied light field and converging it into a single-channel hybrid photoelectric signal; a real-time demodulation and tomographic reconstruction engine for demodulating and reconstructing two dynamic cross-sectional image sequences reflecting the geometry and internal physical properties of the polyester filament from the hybrid signal in parallel; and a spatiotemporal multimodal feature flow classifier for fusing and analyzing the two image sequences using a deep learning model to achieve accurate defect classification. This invention improves the accuracy of identifying various complex defects such as bamboo-like joints, fuzzy fibers, internal voids, and stress anomalies by combining single-point detection with computational tomography and fusing information from both geometric and physical modalities.

The invention relates to a sparse view angle three-dimensional tomography reconstruction method and system based on combination of diffusion model prior and neural representation optimization, and belongs to the technical field of computed tomography and vision. The method comprises the following steps: inputting an axial slice containing artifacts into a SliceFixer model for processing to obtain a latent variable; inputting multi-modal condition data, performing slice restoration through a SliceFixer model, and outputting a restored slice; periodically calling a SliceFixer model to carry out slice repair to obtain a pseudo reference body; and calculating the three-dimensional structure perception loss of the pseudo reference body and the reconstruction body, injecting the three-dimensional structure perception loss into optimization of the neural representation model in a weighting form, and realizing three-dimensional tomography reconstruction based on the optimized neural representation model. And the SliceFixer model is combined with the multi-modal condition data to repair the axial slice of the artifact, so that missing details can be repaired, and the calculation efficiency can be improved.

A computer-implemented method of tomographic reconstruction, the method comprising: receiving an input dataset comprising tomographic projection data of an object; and reconstructing an image of the object using an iterative reconstruction technique, wherein the iterative reconstruction technique seeks to minimize a cost function, the cost function including a first regularizer and a second regularizer, wherein the first regularizer is a first trained machine learning model, trained using image data associated with the object and extracted from a first direction and the second regularizer is a second trained machine learning model, trained using image data associated with the object and extracted from a second direction.

The application belongs to the field of CT tomographic reconstruction and the field of artificial intelligence, and discloses an industrial battery CT image super-resolution method based on deep learning. The method takes a low-resolution tomographic image as input, constructs an end-to-end battery CT image super-resolution network, and outputs a high-resolution tomographic image. The network uses multi-scale attention module (MSA) and enhanced grouping module (EGM) for parallel calculation, adopts hybrid channel attention feed-forward module (HCFB) for nonlinear enhancement, and combines a frequency modulation module for frequency domain feature modeling. The method can obtain a battery CT tomographic reconstruction result with clearer boundaries, more coherent layer structures and fewer artifacts, and provides a more stable and reliable data basis for subsequent defect detection, structure segmentation and three-dimensional analysis.

The invention relates to an emission spectrum chromatography reconstruction prediction method based on a gated circulation unit and an attention mechanism, which comprises the following steps: taking three-dimensional structure distribution of CH * free radicals in a combustion field reconstructed by a traditional chemiluminescence calculation chromatography algorithm as a true value of a deep learning neural network for comparison; a novel neural network model is designed by combining the advantages that a gating circulation unit processes a long-distance dependency relationship and an attention mechanism adjusts importance weights, and three-dimensional reconstruction and prediction of a combustion field at future moments are achieved according to multi-direction two-dimensional projection at historical moments. The method does not need to design a complex acquisition device and does not need to solve a partial differential equation set. According to the multi-direction projection data at the historical moment, the combustion evolution at the future moment can be quickly and accurately predicted within a certain range. And when a frame insertion problem between continuous time sequences is processed, a multi-frame three-dimensional structure can be accurately predicted by means of multi-direction projection at a front moment and a back moment.

The invention discloses a fast inversion method and device for a divertor visible light camera image based on a neural network, and belongs to the technical field of image tomographic reconstruction, and the method comprises the steps: constructing a preliminary data set; selecting a region mainly reflecting neutral particle and impurity radiation according to the preliminary data set, and cutting the region into a fixed size to serve as a real data set; under the condition of assuming tokamak plasma circumferential symmetry, calculating a weight matrix; dividing the real data set into a training set, a verification set and a test set; putting the training set into a ViT model for training, wherein the output of the ViT model is neutral particle and impurity radiation distribution obtained by inversion; according to the relation between the camera image and the radiation distribution of the neutral particles and the impurities, the radiation distribution of the neutral particles and the impurities is multiplied by the weight matrix to obtain a simulated camera image, and a ViT neural network is trained according to the total loss of the network; according to the method, radiation distribution is obtained through one-time calculation, and operation resources are saved.

The invention relates to the technical field of wireless positioning, and discloses a wireless positioning method based on a hybrid communication mode, which comprises the following steps of: in an environment modeling stage, enabling a plurality of anchor nodes to perform full-grid dual-mode signal interaction, and obtaining a flight time difference and a path quality score based on channel characteristic evaluation; the method comprises the following steps: generating a three-dimensional sound slowness field representing an environment by using a weighted tomography reconstruction algorithm, in a target positioning stage, communicating a moving target with anchor nodes to obtain real-time measurement data, generating a shortest flight time field model for each anchor node based on the sound slowness field, constructing a nonlinear optimization target function taking a path mass fraction as a weight, and calculating a target position. The method further comprises a dynamic updating mechanism of a sound slowness field, chromatography modeling is carried out on a space acoustic environment in advance, path mass weighting is introduced, limitation of a traditional sound velocity constant assumption is overcome, influences of environment non-uniformity and time-varying characteristics caused by factors such as temperature and airflow can be compensated, and positioning accuracy and environment adaptive capacity are improved.

A method is provided for correcting a reconstruction artefact of a three-dimensional tomographic image. The method includes the steps of providing an acquired three-dimensional tomographic image from a cell group, and applying the image to a neural network trained in advance to determine a corrected tomographic image.

This invention provides a dual-aperture multi-pinhole collimator, comprising: a collimator plate; and multiple pinholes of two different apertures spaced apart on the collimator plate. During a single imaging process, radioactive rays passing through the imaging fields of different apertures simultaneously form corresponding and non-overlapping projection responses on the detector. Of the two apertures, the first aperture is smaller than the second aperture. The first aperture is determined based on the desired system spatial resolution, and the second aperture is determined based on the desired detection efficiency. This invention also proposes a SPECT imaging system based on the aforementioned multi-pinhole collimator. Its imaging module obtains imaging results from the data required for tomographic reconstruction using an improved MLEM iterative algorithm. The improved MLEM iterative algorithm introduces weighting factors into the original MLEM iterative algorithm, weighting the projection data of the dual-aperture pinholes and the corresponding system transmission matrix elements respectively. This invention improves detection sensitivity while maintaining spatial resolution.

This invention provides a non-destructive testing device and method for nut quality based on a fully automated micro-CT scanner. The method includes: S1, using a transport mechanism to transport the nuts to a rotating platform; S2, keeping the micro-focal spot X-ray source and the flat panel detector stationary, and controlling the rotating platform to rotate intermittently via a PLC controller to achieve full-angle irradiation of the nuts by the micro-focal spot X-ray source; S3, during the intermittent stops of the rotating platform, acquiring planar array images of the nuts from various angles using the flat panel detector to obtain a full-angle planar array image of the nuts, and transmitting it to a computer; S4, the computer performs tomographic reconstruction and image processing based on the full-angle planar array image to obtain the quality trait parameters of the nuts. This invention enables non-destructive testing of nut quality trait parameters through visual recognition technology, significantly improving testing accuracy and efficiency compared to manual trait testing methods and traditional culture methods for detecting mold infection.

The invention provides an electromagnetic chromatography evaluation method for aging damage of a metal material, which is used for establishing a relationship between the electromagnetic property of the material and the aging damage degree in order to solve the problem of the electromagnetic property of the material caused by environment and load action in the early stage of long-term service of the conductive metal material. A broadband electromagnetic response signal is obtained through pulsed eddy current excitation, equivalent multi-frequency information is extracted through spectral analysis, an electromagnetic field positive problem model under the multi-frequency condition is constructed, and chromatographic reconstruction of the internal conductivity of the material in the depth direction is achieved. And introducing a prior constraint that the conductivity changes in a quadratic function along the depth, and reconstructing a stable conductivity distribution profile by adopting a nonlinear least square inversion method. On the basis, local aging damage degree and weighted aging damage evaluation indexes are constructed, and quantitative evaluation of aging damage distribution of the material in the depth direction and the overall aging damage degree is achieved.

The present application belongs to the technical field of image analysis, and discloses a kind of vehicle lamp thick wall piece image defect intelligent detection system;Including emitting depth direction time coding laser pulse sequence to vehicle lamp thick wall piece, so that different depth defects are excited in turn to generate photoacoustic shock wave, and the photoacoustic shock wave is collected in space-time combination, and multiple layer original correlation intensity matrix is generated;Tomographic reconstruction is executed to multiple layer original correlation intensity matrix, and the complete three-dimensional photoacoustic density field of thick wall piece is generated;Three-dimensional photoacoustic density field is carried out voxel grid division and executes three-dimensional connected domain marking, and all independent defect bodies are obtained;The geometric feature and topological feature of the voxel set of each independent defect body are calculated, and the defect feature vector set is used as input, and the defect type and three-dimensional center coordinates are output through the pre-trained defect mapping model;According to defect type and three-dimensional center coordinates, mechanical hand is controlled to execute sorting or marking;Vehicle lamp thick wall piece image defect intelligent detection is realized.

The present application relates to a sparse view three-dimensional tomographic reconstruction method and system based on diffusion model prior and neural representation optimization, belonging to the field of computer tomography and visual technology. The method comprises: inputting an axial slice containing artifacts into a SliceFixer model for processing to obtain a latent variable; inputting multi-modal conditional data, performing slice repair through the SliceFixer model, and outputting a repaired slice; periodically calling the SliceFixer model for slice repair to obtain a pseudo reference body; calculating a three-dimensional structure perception loss of the pseudo reference body and a reconstructed body and injecting it into the optimization of the neural representation model in a weighted form, and realizing three-dimensional tomographic reconstruction based on the optimized neural representation model. The SliceFixer model is used to repair the axial slice with artifacts in combination with multi-modal conditional data, which can repair the missing details and improve the calculation efficiency.

The invention discloses a self-adaptive temperature field measuring device and method for a dynamic arc. The device comprises an arc generating mechanism, an image collecting device, a reflecting mirror set, an angle reference component and a processing unit. The angle reference component is installed on the arc generation mechanism, the central axis is coaxial with the geometric installation axis of the arc generation mechanism, and an angle scale structure is arranged on the surface and used for establishing the mapping relation between the pixel position and the observation angle. The reflector group is arranged on the side of the arc generation mechanism and is used for reflecting radiation light from different directions of the dynamic arc to the same image acquisition device so as to realize multi-view imaging; the processing unit determines the projection angle of each visual angle in real time in the dynamic discharge process, performs tomography reconstruction based on multi-visual-angle radiation information, obtains emission coefficient distribution of an arc section, adaptively distinguishes a high-temperature region from a low-temperature region according to isoline topological characteristics, performs temperature mapping, and realizes accurate reconstruction of a dynamic arc temperature field. And the stability and reliability of measurement under the dynamic working condition are improved.

Methods, systems, and apparatus for performing scanning spectral tomographic reconstruction of an object. The imaging system includes a power source that is configured to provide an alternating high voltage. The imaging system includes an X-ray source. The X-ray source includes an array of X-ray emitters that allow fast switching “ON” and “OFF” using a grid electrode. The source is configured to generate an X-ray beam with an energy spectrum based on the alternating high voltage and uses X-ray filters. The imaging system includes a controller configured to operate synchronously with the alternating high voltage. The controller is also configured to drive an actuator to position the X-ray source with respect to an object and drive the source in a pre-defined trajectory about the object. At each position in the trajectory, the controller is configured to control the exposure timing of the emitters based on a predefined firing pattern.

The invention belongs to the field of CT (Computed Tomography) tomography reconstruction and the field of artificial intelligence, and discloses an industrial battery CT image super-resolution method based on deep learning. According to the method, a low-resolution cross-sectional image is used as input, and an end-to-end battery CT image super-resolution network is constructed to output a high-resolution cross-sectional image. The network uses a multi-scale attention module MSA and an enhanced packet Mama module EGM to carry out parallel computing, uses a mixed channel attention feedforward module HCFB to carry out nonlinear enhancement, and combines a frequency modulation module to carry out frequency domain feature modeling. According to the method, a battery CT fault reconstruction result with clearer boundary, more coherent layered structure and fewer artifacts can be obtained, and a more stable and reliable data basis is provided for subsequent defect detection, structure segmentation and three-dimensional analysis.

A computer-implemented process for electromagnetic imaging, the process including the steps of: accessing scattering data representing at least a two-dimensional array of measurements of electromagnetic wave scattering by internal features of an object, wherein each said measurement represents scattering of electromagnetic waves emitted by a corresponding antenna of an array of antennas disposed about the object as measured by a corresponding antenna of the array of antennas; and processing the scattering data to generate image data representing a spatial distribution of internal features of the object, wherein the generation of the image data does not involve tomographic reconstruction but is in accordance with a weighted mapping to directly map the measurements of electromagnetic wave scattering to a corresponding spatial distribution of electromagnetic wave scattering by the internal features of the object that corresponds to the physical shape of the object to enable the detection, localization, size estimation, shape estimation and classification of one or more features of interest of the object.

The application discloses a kind of based on neural network's quick inversion method and device of filter visible light camera image, belong to image tomographic reconstruction technical field, this method includes: constructing preliminary data set;According to preliminary data set, select the area mainly reflecting neutral particle and impurity radiation, and cut fixed size as real data set;In the case where assuming tokamak plasma is circumferentially symmetric, weight matrix is calculated;Real data set is divided into training set, verification set and test set;Training set is put into ViT model and is trained, and the output of ViT model is the neutral particle and impurity radiation distribution obtained by inversion;Then according to the relationship between camera image and the radiation distribution of neutral particle and impurity, the radiation distribution of neutral particle and impurity is multiplied by weight matrix to obtain simulated camera image, and ViT neural network is trained according to the overall loss of network.The application realizes that radiation distribution is obtained once calculation, and operation resource is saved.