174results about How to "Realize 3D reconstruction" patented technology

The invention discloses a stereo vision sensor integrates a line array camera and an area array camera and a calibration method. The vision sensor can synchronously acquire gray and depth information of an object image. The sensor mainly comprises the line array camera, the area array camera, a laser and other accessories. The line array camera acquires a clear image through line laser illumination, and forms the stereo vision sensor with the area array camera. The line laser is used as the feature of the area array camera image, matches an epipolar constraint to realize corresponding matching of the line array camera and the area array camera image. Three-dimensional coordinate reconstruction is realized through a stereo vision measurement model. The sensor can simultaneously acquire the image information of the object and the spatial depth information corresponding to each pixel in a push-scan manner. The stereo vision sensor provided by the invention can be widely applied to the fields of object recognition, fault diagnosis and the like.

The invention relates to a bend pipe measuring and allowance locating method facing digital manufacture. The method comprises the following steps that multiple sensors are calibrated globally by meansof a laser tracker and a target, and coordinate systems of the sensors are unified into a global coordinate system; the sensors collect bend-pipe images and carry out edge detection and morphologicalrefinement to obtain a center line of the bend pipe in a strictly synchronous way, and 3D reconstruction of the center line is carried out in a bidirectional polar line matching algorithm; the centerline is dispersed into space point cloud, a control point, straightway end points and an arc center point of an arc segment are calculated, and 3D molding is carried out by taking the center line asa guiding line; and a design model matches a measurement model via a nearest point iteration method, the allowance position is calculated. Thus, self-locating cutting and welding of the bend pipe is realized, and the digital manufacturing process of the bend pipe is accelerated.

The invention relates to a measurement method of smoke concentration, temperature and particle concentration distributions and a device thereof, in particular to a measurement method by using multi-source tomography technology while reconstructing smoke concentration, particle concentration and temperature distributions and a device. The method comprises the following steps of: transmitting a laser signal to an optical fiber collimator; making reciprocating swing movement as soon as the optical fiber collimator emits a light beam; reflecting the light beam by a film-plated columnar reflector to a photodetector after passing through a flue outlet; transmitting the measurement signal to a computer by the photodetector to obtain the projection data of a fault plane at different angles; reconstructing gas concentration and temperature by using the projection data; and reconstructing particle concentration by using the projection data. In the invention, a tunable laser respectively scans a particle and a smoke influencing regions, so that the measurements of the smoke and the particle are independent from each other; the fault plane distributions of smoke, particle concentration and temperature can be reconstructed at the same time to achieve 3-dimensional reconstruction; the measurement of fault plane parameters has short time and dynamic change; and the measurement cost is very low.

The invention relates to a DTI (Diffusion Tensor Imaging)-based cranial nerve fiber bundle three-dimensional rebuilding method and a method of manufacturing a head three-dimensional model comprising nerve fiber bundles based on a 3D printing technology. The method comprises the following steps: magnetic resonance scanning is carried out on a target tissue area and surrounding nerve fiber bundles to acquire MRI image data for the target tissue area comprising the nerve fiber bundles; DTI processing and related processing are carried out on the acquired MRI image to acquire an MRI image with an identifiable nerve fiber bundle; calculation is carried out on X-axis information, Y-axis information and Z-axis information of the MRI image with the identifiable nerve fiber bundle, three-dimensional rebuilding is carried out via mimics software, and a three-dimensional mode comprising the cranial nerve fiber bundle is acquired. Through the three-dimensional model and the 3D printing three-dimensional solid model, the position relation among an anatomic structure, a brain function area and each tissue can be displayed clearly, a solid model is provided for an operation, and the method is used for operation approach design and operation simulation.

The invention relates to an advanced detection method for focusing chromatography induced polarization of underground engineering. The advanced detection method comprises the following steps of: arranging a circle of shield electrodes A1 on a contour line of an operating surface of the underground engineering by utilizing a focusing operating manner and arranging a plurality of parallel measuringlines on parts with different heights from a base plate to a top plate firstly, wherein the measuring lines are provided with power supply electrodes and receiving electrodes, and the base plate is also provided with power supply electrodes and receiving electrodes; acquiring data layer by layer by utilizing a chromatography method; carrying out inversion on measuring line potential difference data obtained through measurement to obtain a three-dimensional resistivity image of a geologic body in the front of the operating surface; drawing a relation coordinate graph of difference data and horizontal distances of half time by using the half time obtained through measurement, calculating an envelope area St of half time difference and abscissa axis, and drawing a two-dimensional profile mapof the envelope area St to realize estimation of the quantity of water in the front of the operating surface. The advanced detection method for the focusing chromatography induced polarization of theunderground engineering, disclosed by the invention, has excellent advanced detection directivity, realizes three-dimensional location of an aquifer and solves problem of incapability of location in a traditional focusing electric method during single-point measurement.

A binocular three-dimensional reconstruction method includes the following steps: step A, image collection: using two color cameras with the same model and arranging the two color cameras horizontally in a parallel mode, and enabling an target object to be arranged within the collecting range of the two cameras, wherein a image collected by a left camera is an original left image and a image collected by a right camera is an original right image; step B, camera calibration, and image rectification and pretreatment: adopting a chessboard method to demarcate and obtain inner and outer diameters of the two cameras, and then rectifying the original left image and the original right image, and next, pretreating the images; step C, hierarchical matching: conducting hierarchical feature extraction and matching to the images according to strong and week of feature lines of the images; step D, three-dimensional reconstruction: after completing the delamination which reaches required precision, according to a parallax principle, enabling obtained matching information of the two images to be calculated into three-dimensional information of the target object, adopting a method of plane interpolation, and then completing the three-dimensional reconstruction. The method has the advantages of being low in algorithm complexity and capable of achieving the three-dimensional reconstruction well and the like.

The invention relates to an automatic image recognition system and method based on the neural network method. The system comprises an unmanned aerial vehicle, a back-stage control platform and / or a far-end center control device of the aerial vehicle. The system can be used in the fields of equipment maintenance, traffic monitoring, wide-range target location and the like, thereby reducing cost and / or improving efficiency. Images of real-time scenes are collected through the unmanned aerial vehicle, automatic / manual remote control of the unmanned aerial vehicle is performed through the far-end center control device of the aerial vehicle, massive data are processed through a back-stage control platform, image contents are calculated and identified, and a variety of operative goals are completed: scene monitoring, scene early-warning and alarming, scene three-dimensional reconstruction, scene recognition, target object recognition or target object tracking and the like. The system is capable of flexibly deploying according to the scene. In addition, automatic scene scanning and three-dimensional reconstruction without depending / depending on auxiliary data are achieved through the unmanned aerial vehicle.

The invention relates to a three dimensional reconstruction method of human body surface under a complex scene. The invention automatically extracts the three dimensional information of the human bodysurface on the basis of robust image foreground extraction and information fusion theory. Firstly, the foreground image of a human body target is extracted from a multiple-view image collected by each camera, the probability of the foreground and background to which each point belongs is calculated to obtain the foreground similarity probability graph; then, the information of each view is fusedto calculate the occupation probability of space voxel by a human body, so that the three dimensional surface of the human body is reconstructed. The invention has the advantages of easy realization,strong robustness, high precision, high reconstruction speed and the like and can provide quick and reliable three dimensional surface of the human body in relevant application fields, such as man-machine interaction, virtual reality, human body information digitization, human body kinematic analysis and the like.

The invention discloses a visual SLAM-based cognitive point cloud map creation system, and aims at obtaining image pose information by utilizing visual SLAM so as to estimate relative dense reverse depth information of environment, recognizing interested objects in the environment and creating a cognitive point cloud map of the environment. The system comprises a system starting module, a reversedepth map estimation module and a map module, wherein the system starting module is responsible for selecting a first key frame; the reverse depth map estimation module is responsible for estimating areverse depth map of the key frame and selecting a new key frame; and the map module is responsible for maintaining the created cognitive point cloud map, recognizing interested objects in the key frame when the new key frame is added into the map, and visually displaying the map creation. According to the system, the relatively dense cognitive point cloud map of the environment is created by utilizing image information, the map is capable of realizing positioning of mobile robots, obstacle avoidance, autonomous navigation and three-dimensional environment reconstruction, and cognitive information in the map can be used for realizing interaction between the mobile robots and human beings and interaction between the mobile robots and the environment.

The invention provides a left ventricle nuclear magnetic resonance image segmentation and three-dimensional reconstruction method. According to the method, a variational level set evolution model is established to be used for segmenting the boundary of the inner membrane and the outer membrane of the left ventricle, and meanwhile, a three-dimensional curved surface and entity reconstruction scheme of the left ventricle and a corresponding mesh generation scheme are also provided. Due to the fact that a novel quasi local binary simulation item based on an edge detection operator is adopted, the method can be used for effectively segmenting images which are uneven in gray level; only the area item is considered, therefore, theoretical models involved in the method can be simplified, and the algorithm operation time of the method can also be shortened; the application of the convex hull algorithm enables the method to be capable of making up the influences such as leakage caused by fuzziness of the boundary of the outer membrane. According to the method, the three-dimensional reconstruction scheme based on the contour surface generation algorithm and secondary development and application technologies of relevant commercial software is further provided, and can help a user realize three-dimensional reconstruction of the left ventricle.

The invention belongs to the technical field of computer multimedia. In order to provide a simple and practical multi-view video interpolation and three-dimensional modeling method, the invention adopts the technical scheme that: a time-space jointed multi-view video interpolation and three-dimensional modeling method is provided, a plurality of camera arrays are grouped in a spacing manner, and the three-dimensional model of a scene at every moment is reconstructed. The time-space jointed multi-view video interpolation and three-dimensional modeling method specifically comprises the following steps of: (1) obtaining unacquired frames between two frames through interpolation; (2) obtaining an image of unacquired visual angles at this moment by using a model-assisted weighting method; (3) calculating and extracting key points; (4) describing the extracted key points by using shape context and solving by the Hungarian method; (5) obtaining final interpolated frames through solving the problem of Poisson editing and optimization; and (6) reconstructing and rendering the three-dimensional model of the scene. The time-space jointed multi-view video interpolation and three-dimensional modeling method is mainly applied to the design and the manufacture of antennae.

The invention discloses an absorption and scattering coefficient reconstruction method in optical projection tomography (OPT). Transmitted light intensity is decomposed using the theory of series expansion; the scattering coefficient of biological tissues is reconstructed using a statistical iterative algorithm according to an OPT model and based on the characteristics of ballistic propagation signals and scattering signals in measured signals; the absorption coefficient of biological tissues is calculated based on the relationship of the attenuation coefficient, scattering coefficient and absorption coefficient; and finally, spatial distribution of the absorption coefficient of biological tissues and spatial distribution of the scattering coefficient of biological tissues are reconstructed at the same time. The method is applied to projection tomography reconstruction of weak scattering mediums. 3D reconstruction of the absorption coefficient and scattering coefficient of biological tissues is realized. The problem of scattering in OPT is solved effectively, the quality of OPT is improved, and information provided by the OPT technology is enriched.

The invention discloses point light source welding seam scanning detection method, which comprises the steps of emitting point laser to a preset point on a welding seam of a weldment; acquiring reflected light of the point laser; acquiring the relative distance between the preset point and a laser displacement sensor after measurement processing; moving the weldment towards a first direction in sequence by a first preset distance, and acquiring the relative distance between each detection point in the first direction and the laser displacement sensor; moving the weldment towards a second direction in sequence by a second preset distance after each detection point in the first direction is detected, moving the weldment towards the first direction in sequence by a third preset distance, and acquiring the relative distance between each detection point in the third direction and the laser displacement sensor; and calculating shape parameters of the welding seam according to the relative distance between each detection point in the first direction and the laser displacement sensor and the relative distance between each detection point in the second direction and the laser displacement sensor, and generating corresponding characteristic signals by using the shape parameters. The point light source welding seam scanning detection method effectively realizes automatic detection for forming conditions of a welding seam.

The invention relates to the technical field of marine observation and aims to provide a multi-node data synchronous acquisition system and method for real-time monitoring of underwater surface shapedeformation. The system comprises at least four array belts, wherein each array belt is formed by a plurality of strip-shaped rigid substrates connected through movable joints; on each rigid substrate, three sensor units are connected to slave station data acquisition units through cables, and the slave station data acquisition units are connected to a central processing unit through cables; the central processing unit comprises an outer pressure shell as well as an embedded controller and a power supply which are arranged inside, and after acquiring data from MEMS attitude sensors, each slavestation data acquisition unit transmits the data to the embedded controller. Synchronous acquisition of underwater and even seabed multi-node data can be realized, three-dimensional reconstruction ofthe surface shape is realized, and the system and the method can be used for improving the marine observation capability. A clock of a master station is taken as a clock source of the system, time synchronization in the system is realized in a form of I / O timing outputting of a clock synchronization instruction, and the problem about time synchronization of the system is effectively solved.

The invention discloses a crucial boiler pipeline online measurement system. The crucial boiler pipeline online measurement system comprises two obviously-identified radial deformation measurement metal blocks, a radial measurement binocular camera, two obviously-identified axial deformation measurement metal blocks, an axial measurement binocular camera, a displacement sensor and at least one stress and strain transducer, wherein the stress and strain transducers are arranged at the bottoms of the metal blocks. The displacement sensor is positioned between two prearranged axial deformation measurement metal blocks, and two sensing heads contact with the metal blocks respectively. The binocular cameras are connected with an industrial switch and are connected to a main control machine via a transmitting network bridge and a receiving network bridge. The crucial boiler pipeline online measurement system is capable of detecting data such as stress, strain and displacement of a boiler pipeline comprehensively and efficiently in real time to reflect boiler pipeline damage degree, and is used for judging remaining life of the boiler pipeline to perform failure early warning, so that safety and stable operation of electric power are guaranteed.

The invention discloses a three-dimensional scanning method for a plant organ form, which is used for establishing a three-dimensional model for the surface of a plant organ. The method specifically comprises the following steps of: (1) acquiring a calibration board image and a plant organ image sequence; (2) processing the calibration board image to acquire a calibration parameter and a radial aberrance coefficient of image acquisition equipment; (3) extracting the outline of the plant organ in each image of the image sequence of the whole body of the plant organ; and acquiring all pixel point coordinates on each outline; and (4) processing each outline and acquiring the point set characterizing a measured object, namely a plant organ shell so as to realize the three-dimensional reconstruction of the plant organ. The method is mainly used in the field of agriculture and forestry; a plurality of scanning points can be obtained at the same time; the scanning speed is high; the scanningprecision is high and reaches a pixel level; the automation of the main operation is basically realized; and the method is simple and convenient to operate.

The invention provides a fruit tree crown layer leaf and fruit three-dimensional reconstruction method and system. The method includes the steps of extracting a three-dimensional branch framework of a fruit tree crown layer, judging whether each branch in the branch framework is connected with a sub-branch or not, enabling leaf cluster branches and fruit branches to grow on the branches connected with the sub-branches, constructing a leaf three-dimensional model and a fruit three-dimensional model, enabling leaves to grow on the branches in the three-dimensional branch framework according to the leaf three-dimensional model, and enabling fruits to be borne on the fruit branches according to the fruit three-dimensional model. The method is combined with a fruit tree branch framework extraction method of three-dimensional point cloud, and therefore rapid three-dimensional model reconstruction of the shape structure of a fruit tree crown layer with leaves and fruits is achieved; the number, the size and the direction of the leaves in the reconstructed crown layer model better conform to the shape characteristics of real fruit trees and meanwhile the model has high universality, and therefore the high-precision foundation model support is provided for application such as structure analysis of the fruit tree crown layer and the evaluation of physiological and ecological characteristics of the fruit tree crown layer.

The invention relates to a biological autofluorescence tomography method and a device based on iteration reweighting. The method adopts the scheme that by capturing photon signals emitted by tumor cells of a fluorescent protein gene, a size of a tumor focal zone in an organism can be reconstructed three-dimensionally, and positioning analysis can be performed on the focal zone by fusing organism anatomical structure information provided by Micro-CT (Micro-Computed Tomography). According to the method and the device, a non-homogeneous organism model and a photon transmission model based on a diffusion equation are established by combining function information provided by autofluorescence imaging and the structure information provided by Micro-CT imaging, and three-dimensional reconstruction of an illuminant in the organism is achieved by using a norm regularization and iteration reweighting combined optimization strategy. With the adoption of the scheme, a result closer to an actual solution can be reconstructed by less observation quantity; the computational efficiency of solving can be improved effectively; the robustness of a reconstruction algorithm can be improved; and the method and the device are suitable for practical three-dimensional detection and quantitative analysis of a tumor in the practical organism.

The invention discloses an ultrasonic or CT medical image three-dimensional reconstruction method based on transfer learning, which is characterized in that an unsupervised learning mechanism is adopted, and a three-dimensional reconstruction function of an ultrasonic image is achieved through transfer learning by utilizing a visual method according to the characteristics of ultrasonic or CT image acquisition. By means of the method, three-dimensional reconstruction of ultrasonic or CT images can be effectively achieved, in auxiliary diagnosis of artificial intelligence, the effect of auxiliary diagnosis is fully played, and the efficiency of auxiliary diagnosis can be improved through the 3D visual reconstruction result.

The invention discloses a method for creating a semi-dense cognitive map for binocular SLAM (simultaneous localization and mapping). The method comprises the following steps of (1) localizing and identifying an interested object in a key frame; (2) dividing the key frames by meshes, and further finely dividing the meshes of where the interested object is located; (3) extracting a mesh key point ofeach mesh; (4) using a Gaussian-distributed depth filter to representing the depth evaluation of the mesh key point by using a Gaussian-distributed depth filter; (5) matching the mesh key point of the key frame in the existing frame; (6) calculating and updating the depth filter of the mesh key point; (7) when the binocular SLAM selects the new key frame, using the depth filter of the existing key frame to initializinge the new key frame by using the depth filter of the existing key frame; (8) after the binocular SLAM is finished, converting the depth estimation into the specific map representation. The method has the advantages that the autonomous navigation and obstacle avoidance of a mobile robot, and the three-dimensional environment reconstruction are realized; by creating the cognitive map for the environment, the interaction between human and the mobile robot as well as between the human and the environment is realized.

The invention discloses a method for modeling substation GIS equipment based on point cloud data. The method comprises the steps of: (1) collecting the point cloud data of the substation GIS equipment by adopting a method of multiple scanning; (2) splicing the point cloud data of the substation GIS equipment obtained through multiple scanning; (3) performing de-noising processing on the point cloud data; (4) performing rarefying on the de-noised point cloud data; (5) segmenting the point cloud data based on an actual structure of the substation GIS equipment; (6) classifying the segmented point cloud data according to classification indexes of the point cloud data; (7) carrying out model classification reconstruction based on the classified point cloud data; (8) and splicing the obtain a complete substation GIS equipment model. The method disclosed by the invention can be used for modeling the substation GIS equipment precisely and efficiently so as to realize three-dimensional reconstruction of the substation GIS equipment, and provides a good basis for three-dimensional visual and intelligent supervision of the substation.

The invention discloses a face mesh model reconstruction method, a device, an appartus and a storage medium. The method comprises the following steps: acquiring face scanning data to be reconstructedand a three-dimensional face mesh template; The key feature points in the 3D face mesh template are extracted hierarchically, and the 3D face mesh template is deformed according to the posture matching positions of the key feature points extracted hierarchically in the face scanning data, and the target face mesh model is obtained. Obtain the global feature points of the target face mesh model, and according to the pose matching position of the global feature points in the face scanning data, the target face mesh model is deformed, and the reconstructed face mesh model is obtained. The technical proposal provided by the embodiment of the invention solves the problem that the reconstruction speed of the human face mesh model in the prior art is slow and the smoothness is poor, improves thereconstruction speed and the reconstruction effect of the human face mesh model, and improves the smoothness of the reconstructed human face mesh model.

The invention relates to a elbow reconstruction and allowance calculation method based on point cloud data. The method comprises the following steps: collecting elbow point cloud data; carrying out elbow point cloud normal vector calculation; using an MSAC algorithm to detect and extract the straight line segment of the elbow, and achieving elbow segmentation; calculating characteristic parametersof the bent section; extracting the geometric center of the end face; reconstructing the elbow by utilizing the characteristic parameters and the control points; carrying out 3D characterization of the flange; and finally, establishing a catheter cutting allowance energy function, and calculating and determining the cutting allowance and the cutting position of the catheter. According to the elbow reconstruction and allowance calculation method based on the point cloud data, the assembly precision and the assembly efficiency of the guide pipe and the flange can be effectively improved, data support is provided for a guide pipe machining system, and it is guaranteed that a guide pipe allowance removal platform is finally and smoothly built.

The present invention discloses a method and apparatus for estimating depth by using a focus stack based on the density of feature points. The method includes the following steps that: the feature points of each image in a focus stack are extracted, a feature point density-based focus measure is established; a depth estimation model based on a feature point density-introduced weighted focus measure is established; and with an SML focus measure adopted as an instance, an SML and feature point density-based weighted linear mixed focus measure is adopted as the objective function of depth estimation, so that the estimation of the depth and full-focus image of a scene can be realized. According to the method of the invention, the focus measure about the density of the feature points and the linear weighted focus measure are established, and the focus measure-based depth estimation model is constructed, and the depth information of the scene is obtained, and therefore, the full-focus and three-dimensional reconstruction of the scene can be realized. With the method adopted, precise depth information can be provided for the realization of three-dimensional reconstruction, and the full-focus image can be obtained.

The invention discloses a multi-scale cone-beam CT image quick three-dimensional reconstruction method based on wavelet transformation, and the method comprises the following steps: (1), carrying out the circumferential cone-beam CT scanning of a test piece, and collecting a group of projection images for reconstruction; (2), carrying out the wavelet transformation of the collected projection images at the corresponding scales according to the difference of requirements for the resolution of reconstruction data, and obtaining a wavelet decomposition coefficient of each scale; (3), selecting the wavelet decomposition coefficient of the corresponding scale for FDK reconstruction, and continuing the operation if the FDK reconstruction of each component needs to be carried out after wavelet first-order decomposition; (4), respectively extracting cross-sectional images of the obtained reconstruction data along the y-axis, and carrying out the corresponding wavelet inverse transformation; (5), extracting the cross-sectional images of three-dimensional data, obtained after wavelet inverse transformation, along the z-axis, carrying out the linear interpolation, and obtaining a final reconstruction image. The method can quickly achieve the three-dimensional reconstruction within a certain precision range based on the FDK algorithm, and can present three-dimensional data with different resolution in the process according to different requirements.

The invention discloses an intravascular ultrasound image three-dimensional reconstruction method and a system based on deep learning, and the method comprises the steps: collecting a bifurcated bloodvessel image and a normal blood vessel image in an IVUS image, carrying out the labeling, and making a first data set; classifying the first data set by using a classification network to obtain a bifurcated blood vessel image and a normal blood vessel image; respectively marking the inner and outer membranes of the bifurcated blood vessel image and the inner and outer membranes of the normal blood vessel image to form a second data set and a third data set; respectively segmenting the second data set and the third data set by utilizing a semantic segmentation network to respectively obtain inner and outer membrane images of the bifurcated blood vessel and the normal blood vessel; and performing three-dimensional reconstruction on the obtained inner and outer membrane images of the bifurcated blood vessel and the normal blood vessel. According to the method, bifurcated blood vessels and normal blood vessels are classified by using a deep learning method, so that the speed and accuracyof three-dimensional reconstruction can be improved, the reconstruction result is more accurate, intuitive judgment of doctors is facilitated, and the method is of great significance to auxiliary diagnosis of diseases.

The invention discloses a flexible cable identification and three-dimensional reconstruction method based on a point cloud. The method comprises the following steps: firstly, carrying out preprocessing and filtering on the point cloud and removing outliers; then, based on a deep learning-based point cloud semantic segmentation algorithm, obtaining a trained model through pre-training and fine adjustment, outputting a semantic category label of each point, and completing point cloud segmentation; segmenting the cable point cloud into a plurality of point cloud clusters by using an improved K-Means clustering algorithm, and performing cylinder fitting on each point cloud cluster, wherein the center of each cylinder is a discrete point on a cable center line; finally, adopting a sorting algorithm based on PCA and octree direction constraint, on this basis, fitting a center line equation of the cable through a B spline interpolation algorithm, and achieving three-dimensional reconstruction of the cable in the space. The real-time performance is high, the accuracy is high, and the three-dimensional reconstruction effect of the cable can meet the precision requirement of subsequent mechanical arm grabbing.

The invention belongs to the technical field of cell differentiation, and discloses an image detection system and detection method in cell differentiation. The image detection system in cell differentiation comprises a marker gene introduction module, an induced differentiation module, an image acquisition module, a main control module, an image identification module, an image segmentation module,a data storage module and a display module. The invention improves the accuracy and the recognition speed of the cell image recognition through the image recognition module, can realize the intelligent operation and judgment, can quickly identify the cell type state, and can help the doctor to quickly diagnose whether the cell is in the healthy state or not. At the same time, the image segmentation module screens the result of threshold segmentation, the incomplete cell nucleus is sent into adaptive segmentation for secondary segmentation, the complete cell nucleus image is directly output, and then the results of adaptive segmentation and blocking threshold segmentation are combined to form the final ideal output image.

A non-vision-field imaging device comprises a near-infrared pulse laser used for emitting infrared pulse laser, a double-telescope module used for emitting the infrared pulse laser to an intermediatewall after beam expansion and collimation, and receiving the infrared pulse laser scattered back through the intermediate wall, and the double-telescope module is used for emitting the infrared pulselaser to the intermediate wall before the infrared pulse laser is scattered back by the intermediate wall.Wherein the single-photon detector is used for detecting the number of photons of infrared pulse laser scattered back through the intermediate wall, and the computer is used for recording the number of photons and the propagation time of the infrared pulse laser, counting the distribution of the number of photons in time and achieving three-dimensional reconstruction of the target on the basis of the distribution of the number of photons in time. According to the non-vision-field imaging device, a double-telescope system is adopted, coaxial receiving and transmitting of infrared pulse laser can be achieved, the difficulty of a reconstruction algorithm is lowered, and real-time trackingand positioning of a non-vision-field detection target can be achieved.

A holographic SAR subaperture three-dimensional reconstruction method based on an FOCUSS algorithm comprises the following steps: designing a holographic SAR flight path, and initializing radar parameters; dividing the circumferential flight path of the holographic SAR into a plurality of sub-apertures, and determining the observation geometry of the sub-apertures; performing azimuth two-dimensional imaging processing on radar echoes of different incident angle tracks by adopting a ground distance backward projection algorithm to obtain a two-dimensional projection imaging result of the three-dimensional target; selecting a two-dimensional imaging result of a certain track as a main image, taking two-dimensional imaging results of other tracks as auxiliary images, and performing registration processing on the auxiliary images and the main image; stacking the registered multi-track two-dimensional image sets along the elevation direction according to an incident angle sequence to obtaina to-be-processed three-dimensional matrix, and performing compressed sensing processing on a high-order sequence of the three-dimensional matrix by adopting an FOCUSS algorithm to obtain a three-dimensional reconstruction result; and performing coordinate transformation on the obtained three-dimensional reconstruction result to obtain a three-dimensional reconstruction result with relatively small elevation reconstruction and reconstruction position errors under a geodetic coordinate system.