34 results about "Point graph" patented technology

The invention discloses a spatial-domain image steganography method and system based on generative adversarial network. The carrier image is converted into a probability graph through the generation network of a U-shaped structure, and then the probability graph is coded by utilizing a hyperbolic tangent coding module, a tampering point graph is generated, and the carrier image and the tampering point graph are added to generate a secret-carrying image; and then a steganography analysis network is used for distinguishing the carrier image and the secret-carrying image, and the classification result is fed back to the generation network in an error mode; and finally, the trained generation network and the coding module are combined together, as a final spatial-domain image steganography model, the carrier image is input into the whole model, and the secret-carrying image is output. The invention further discloses a space-domain image steganography system based on the generative adversarial network, and the system includes a generation network module, an encoding module and an image steganography module. According to the spatial-domain image steganography method based on the generative adversarial network, the security is obviously improved, and the design is simple.

The invention belongs to the technical field of intelligent teaching, particularly relates to a teaching knowledge point graph system based on network swarm intelligence, and aims to solve the problems of single knowledge point source, poor diversity, repeated knowledge creation and uneven knowledge point quality of an existing teaching knowledge point system. The method comprises the following steps: automatically extracting teaching plan keywords, inputting the teaching plan keywords into a teaching knowledge point map exploration module, and interactively exploring retrieved information bya user: if satisfactory knowledge is found, directly updating a teaching plan; and if not, supplementing knowledge in a man-machine cooperation mode. The man-machine cooperation means that the user and the AI knowledge producer construct new knowledge at the same time and fuse the new knowledge, and the new knowledge is cooperatively created by combining the capabilities of the user and the AI. And the intelligent expert recommendation system recommends experts to participate in verification of the newly constructed knowledge by analyzing the knowledge and the evaluation data of the user, andintegrates the new knowledge with the teaching knowledge point graph. The invention has the advantages of multiple knowledge point sources, high quality and less knowledge repeated creation.

The invention discloses a Mark point graph structure for easy laser marking; the Mark point graph structure comprises Mark points in a spiral form; the Mark point is used for monocrystalline PERC superposing with the selective emitter technology; a preparation method adjusts laser parameters, and lasers spiral Mark points on four corners of a silicon chip. The Mark point graph is a spiral graph, and can prevent problems encountered by a normal concentric circle graph; the spiral new graph can be fully compatible with existing laser doped equipment without increasing any cost, and is good in industrialization production process stability; the making time of the Mark point graph structure is 0.1-0.2s, thus shortening the whole SE structure making time, improving the production power, and increasing enterprise incomes; the Mark point graph structure and the making method are worthy of promotion.

The invention provides a bill form detection method, which comprises the following steps of: inputting a pre-processed bill form picture into a multi-category text detection system, and respectively generating a central point graph, a category graph and a distance graph; searching a central point in the central point graph based on an extreme point detection method so as to determine the positionof a semantically independent field; determining the size of each semantic independent field in the distance map based on the searched central point, thereby determining a candidate box; and based onthe candidate box, determining the category of the candidate box in the category graph by adopting a voting mechanism. Compared with the prior art, the method has the following beneficial effects: based on the idea of center point detection, post-processing of non-maximum suppression (NMS) can be avoided, so that the process is simplified, the prediction process is simple, the speed is high, the detection accuracy is high, and the robustness is good.

The invention discloses a critical rainfall estimation method based on rainfall spatial distribution. The critical rainfall estimation method comprises the steps that firstly, estimating the confluence flow velocity in each grid unit in a hydrological response unit secondly, on the basis of the confluence flow velocity in each grid unit, obtaining a confluence time grid in the drainage basin; performing spatial interpolation on the time-interval rainfall data observed by the station to obtain a rainfall spatial distribution grid, and combining the convergence time grid to obtain a time-interval dynamic rainfall index; counting an accumulated dynamic rainfall index before a flood peak occurs in the flood process; then counting the tension water content saturation, namely the initial soil water content saturation, of the day before the flood process begins; finally, whether the flood is an over-alarm flood is judged, and making a cumulative dynamic rainfall indicator - initial soil watercontent saturation scattering point graph on the basis of multiple floods and i, marking points corresponding to the overwanted flood, and drawing a critical rainfall line in combination with the points corresponding to the overwanted flood.

The invention discloses a dense crowd counting method and system based on hidden density distribution and a terminal. The method comprises the steps: obtaining a self-adaptive hidden Gaussian densitygraph through a Gaussian network according to a crowd point graph; according to the counting loss item, the smooth item and the Bayesian item, guiding the hidden Gaussian density graph optimization, so as to enable the generation quality to be higher; taking the hidden Gaussian density map as a training target, combining an adversarial loss function and a Bayesian loss function, and outputting thedense crowd image as a predicted density distribution map; summing the predicted density distribution maps to obtain the predicted density people number. The density predictor, the hidden Gaussian density generator and the discriminator are alternately trained and cooperatively optimized. According to the method, the precision is improved to a large extent, the robustness is good, and the application value is high due to the fact that the parameter quantity and the operand of the inference stage are not increased.

The embodiment of the invention provides a distribution network project management current situation evaluation method and apparatus, and belongs to the data processing field. The distribution network project management current situation evaluation method includes the steps: acquiring a plurality of project data prestored in a local database; based on material quantity and financing volume, drawing a point graph; based on a first preset rule, acquiring the projects, to be analyzed, displayed on the point graph; performing clustering operation on the projects to be analyzed according to a second preset rule, and obtaining a plurality of clusters; acquiring the standard characteristic graph of each cluster; acquiring the characteristic analysis graph of each project to be analyzed in each cluster; and based on the characteristic analysis graph and the standard characteristic graph, acquiring a problem statistical graph. The distribution network project management current situation evaluation method can obtain the differences between the characteristic analysis graph and the standard characteristic graph by comparing the characteristic analysis graph with the standard characteristic graph one by one, so as to evaluate the project to be analyzed by means of the meaning represented by the differences to infer the total trend of the existed problem in the project to be analyzed and provide basis for the decision of modifying problems for a leader layer, a management layer and an executive layer.

The invention discloses a spatial-temporal trajectory-based group movement mobile cluster mode sorting method. The method comprises the following steps of: constructing a mobile cluster-interest pointbipartite graph; and carrying sorting on the basis of a restart type random walk model. According to the method, a problem of carrying out importance sorting on mobile clusters on the basis of the restart type random walk model is converted into a problem of carrying out importance calculation on vertexes in the mobile cluster-interest point graph; a probability value of each vertex represents the importance of the vertex, and when the probability is larger, the vertex is more important; and by utilizing the model, the importance sorting of the mobile clusters is easy to realize, so that users can automatically complete the sorting process without setting excessive parameters.

The invention relates to the technical field of knowledge graphs, and provides a graph data query method and device, computer equipment and a storage medium. The method comprises the steps of obtaining service classification information corresponding to each vertex in a graph data full set, and classifying the graph data full set according to the service classification information to obtain a plurality of graph data subsets; determining a corresponding super point according to each graph data subset to obtain a super point graph data set; when a query request sent by a terminal is received, obtaining a query condition carried in the query request; determining a traversal condition according to a query type corresponding to the query condition, and filtering out superpoints which do not meet the traversal condition from the superpoint graph data set to obtain target superpoints; and obtaining target query data corresponding to the query condition from a graph data subset corresponding to the target super point, and sending the target query data to the terminal. By adopting the method, the graph data query efficiency can be improved.

The invention provides a font element determining device and a font element determining method. The invention aims at providing a font element determining device capable of determining radicals and constituent parts except radicals. As a solving way, a quantitative data generating part generates quantitative data relative to a contour graph. The quantitative data comprise a size and a central coordinate of a rectangle surrounding the contour graph. A radical configuration mode analysis part determines a configuration mode of radicals and constituent parts except the radicals according to the quantitative data, and divides the contour graph into a first group as the radicals and a second group as the constituent parts except the radicals according to the configuration mode. A transformation part transforms the contour graphs of the first group and the second group to a first point graph and a second point graph. A radical / constituent part recognition part respectively determines the radicals and the constituent parts except the radicals according to OCR recognition of the first point graph and the second point graph.

The invention discloses a knowledge graph personalized learning path recommendation method based on RankNet-transfer. The method comprises the steps of knowledge model construction and model recommendation. In the knowledge model construction, a blind area knowledge point graph is constructed according to the construction of a course knowledge point graph and the analysis result of the knowledge test by the cognitive model. In the recommendation model, a RankNet algorithm and a Transform algorithm are used to realize reordering of blind area knowledge points, and then a topological ordering method is used to traverse a reordering sequence and a blind area knowledge point graph to generate a final knowledge point recommendation sequence. The accuracy of the recommendation result is improved, and the method adapts to the cognitive sequence of the learner.

The invention discloses an electronic chart position point data simplification method, which comprises the following steps of: 1, dividing chart position points into three types, namely point chart layer discrete position points, line chart layer continuous position points and surface chart layer continuous position points, and respectively processing the three types of chart layer position pointdata; 2, for the discrete position point data of the point graph layer, the discrete position points are continued according to a distance or byte rule, a reference position point of each position point is determined, and then each position point is represented by a relative coordinate relative to the reference position point; and 3, for the line and surface layer data, processing the initial position points of all the line segments according to the step (2), taking the midpoint of two adjacent position points as a reference position point for each group of continuous position points, and representing the line and surface layer position points by using the relative coordinates relative to the reference position point. And 4, storing the relative coordinate data, and generating an electronic chart file. The electronic chart precision is maintained, the chart file is greatly simplified, the storage space is reduced, and the processing efficiency is improved.

The invention provides a method of increasing high-aspect ratio photoetching process windows through a dual graph technology. The method comprises the following steps: photoetching process requirements are divided to a combination of twice photoetching; a first photomask is used to carry out first photoetching processing on a process graph; a second photomask is used to carry out second photoetching processing on the process graph, and wherein the process window graph formed by the combination of the first photomask and the second photomask has a pixel size meeting the process requirements andare 1D groove graph photomasks. According to the method of increasing the high-aspect ratio photoetching process windows through the dual graph technology, in view of the high-aspect ratio layer, when the pixel size is reduced gradually and a target linewidth is reduced, the dual graph technology is adopted to split a one-layer 2D point graph to a two-layer 1D groove graph, and in combination ofthe PSM photomask process mode, the high-aspect ratio layer process windows are increased.

The invention discloses a terminal and a system for detecting leakage of a tap water pipeline, and relates to the field of pipeline leakage detection. For mainly aims to solve the problems that the leakage point of the tap water pipeline detected by satellite remote sensing detection is high in false alarm rate and low in checking efficiency of the tap water pipeline, the terminal and the system comprise a PC end system, a mobile phone end system and a leakage point working card system, wherein the PC end system comprises a receiving module, an overlapping processing module and a positioning module, wherein the receiving module is used for receiving a leakage point coordinate information diagram of a leaking region detected by the satellite scanning detection, the overlapping processing module is used for coinciding a regional leakage point graph, a regional high-definition map and a regional pipe network diagram accurately through an image superposition technology to generate a leakage point comprehensive information diagram, and the positioning module is used for positioning the position of a checking worker and the position of a leakage point to be checked. The terminal and thesystem can accurately evaluate the specific position of the leakage, so that the checking range is greatly reduced, the checking efficiency of the tap water pipeline is improved, and the labor cost isreduced.

The invention discloses a method and a device for detecting aggregation of defect points, relates to the technical field of automatic optical detection, improves the detection efficiency and accuracyof detecting whether the defect points in a defect point graph have aggregation, and adopts the main technical scheme that obtaining a to-be-detected defect point graph, a partition mode, the number of partitions and a variance threshold value are obtained; carrying out region segmentation processing on the defect point graph to be detected according to the partitioning mode and the partitioning number so as to obtain a plurality of partitions; determining the number of defect points corresponding to each partition, and calculating a normalized proportion value corresponding to each partitionaccording to the number of defect points corresponding to each partition; calculating variance values corresponding to the plurality of normalized proportion values; judging whether the variance valueis greater than the variance threshold or not; and if yes, determining that the defect points in the to-be-detected defect point graph have aggregation. The method is applied to the process of detecting whether the defect points in the defect point graph have aggregation or not.

The invention discloses a high-precision track point display and control method in a space coordinate system, comprising the following steps: (1) inputting track points successively and optimizing the track points according to time, positions and attitudes; (2) computing basic coordinates in the space coordinate system; (3) traversing a preferred track point list, performing coordinate offset on the track point, and successively acquiring the local coordinates of track points with the exception of an initial point; (4) drawing a track point graph by a underlying rendering platform. The high-precision track point display and control method realizes display control of high-precision track points in the world coordinate system effectively, and avoids a high-precision graphic drawing joggle problem.

An image-based digital bamboo-woven picture automatic generation method comprises the following steps: step 1, carrying out fuzzy processing on an input image I by using a Gaussian fuzzy algorithm toobtain a fuzzy image Ir; step 2, regarding every 4 * 4 pixels of the blurred image Ir as a mosaic, and performing gray mosaic to obtain a gray mosaic image Ig; step 3, applying an Otsu method to the grayscale mosaic image Ig; step 4, carrying out region segmentation on the It by using a seed filling algorithm to obtain a region pattern R; step 5, performing morphological gradient operation on theregional pattern R to complete edge detection, and obtaining a mask layer M marked with each regional boundary on the R; step 6, reading the connection point mode of each number from the file; tilingand filling each region with the connection point mode on the region mode graph R by using a corresponding connection point mode to obtain a connection point graph P; and step 7, traversing each mosaic on the binary image It to obtain the digital bamboo-woven picture added with the connection points.

The invention relates to a map segmentation method and device, a pose estimation method and an equipment terminal. The map segmentation method comprises the following steps: segmenting a map; acquiring map point data and constructing a corresponding point graph according to the map point data, wherein the point graph comprises nodes formed by map points and corresponding edges; measuring and estimating a to-be-estimated state variable corresponding to each node in the point diagram, and obtaining corresponding measurement and estimation values, wherein the to-be-estimated state variables comprise homogeneous coordinates of the nodes and corresponding sensor when the nodes are acquired; calculating the inter-point correlation between any two adjacent nodes in the point graph according to the measurement and estimation values; judging whether the inter-point correlation is smaller than a preset correlation threshold value or not; when the correlation between the points is smaller than apreset correlation threshold value, removing the edges corresponding to the corresponding adjacent nodes so that the static points and the dynamic points can be separated to complete map segmentation.The map segmentation method eliminates the influence of the dynamic points, and a foundation is laid for subsequent pose estimation.

The invention relates to a GIS and SWMM model-based sub-catchment area quick search connection method, which comprises the steps of 1, importing original data into GIS software to generate a surface map layer of a sub-catchment area shp, and inputting pipe network node data to generate a point map layer of the shp; 2, converting the graph surface layer of the sub-catchment area shp in the step 1 into a point surface layer, and performing mask analysis on the point surface layer to generate point X and Y coordinates and elevation information; 3, based on GIS software, conducting neighbor analysis on the point graph layer of the pipe network node shp in the step 1 and the point graph layer of the sub-catchment area in the step 2; 4, comparing the elevation information of the two points in the step 3, completing the search connection of the sub catchment areas, and exporting the coordinate information to generate an Inp file; and 5, creating and generating a hydraulic pipe network model on the SWMM platform by using the Inp file.

The invention discloses an image enhancement processing method applied to correlated imaging, and the method comprises the steps: carrying out the correlated calculation of reference arm light field sampling data and signal arm light intensity response data, and obtaining a correlated imaging result graph; calculating a normalized autocorrelation matrix and a full width at half maximum of a resultgraph; subtracting the product of the normalized autocorrelation matrix and the gain coefficient from the result graph to obtain a residual graph; iteratively calculating a maximum value point in theresidual error graph and a corresponding normalized autocorrelation matrix, and subtracting a product of the matrix and a gain coefficient from the residual error graph; counting the position and theintensity of the point into a net point graph; stopping iteration when an iteration termination condition is met; and performing convolution on the Gaussian function and the net point graph through half-height full-width fitting, so that a correlated imaging result graph of an aperiodic structure can be obtained. The method can solve the periodic repeated structure problem caused by insufficientspatial frequency sampling, improves the visibility and the signal-to-noise ratio of the correlated imaging result graph, and is high in anti-noise capability, high in calculation speed and good in practical effect.

The invention provides a face reconstruction method and related equipment. The face reconstruction method comprises the steps of obtaining a first number of face images of a face reconstruction targetobject; obtaining a second number of two-dimensional graphs corresponding to the first number of face images through a three-dimensional origin space calculation model according to the first number of face images; obtaining a three-dimensional point graph according to the second number of two-dimensional graphs; and reconstructing a three-dimensional face according to the three-dimensional pointgraph. According to the face reconstruction method and the related equipment provided by the embodiment of the invention, the problems of poor operability and high cost of an existing three-dimensional reconstruction method can be solved to a certain extent.

The invention provides a cubic object volume calculation method for eliminating noise based on a depth image. The method comprises the following steps: acquiring the depth image and a gradient map ofa cubic object; classifying pixel points in the gradient map according to the gradient map, and obtaining segmented images of an edge point map, an upper surface, a left vertical plane and a right vertical plane of the depth image; carrying out edge line detection on the edge point graph, clustering and merging all detected edge lines to obtain a collinear edge line group set, and extracting angular points according to the longest edge line of each group to carry out pairing; achieving positioning of the cubic object; fitting the plane of the cubic object in a three-dimensional space accordingto the upper surface, the left vertical plane and the right vertical plane of the cubic object; removing noise points on each surface through multiple times of fitting, acquiring an ideal point set corresponding to each surface, acquiring the contour boundary and corner points of each surface according to the ideal point set, acquiring the length, width and height of the cubic object are obtainedaccording to the coordinates of the corner points in the three-dimensional space, and calculating the volume of the cubic object. The volume of the cubic object can be accurately calculated.

The invention relates to a graph-based 3D point cloud object instance segmentation method in an out-of-order scene. The method comprises the following steps: step 1) carrying out downsampling and denoising on original scene point cloud data; 2) extracting key points of the scene point cloud, and forming a key point graph according to the neighbor relation between the key points; 3) completing coarse segmentation of the point cloud based on the key point graph; 4) constructing a direction bounding box based on the roughly segmented region, judging whether the region needs to be finely segmented according to the size of the direction bounding box, and extracting a complete target point cloud from the scene according to the direction bounding box; and in a fine segmentation stage, through combination of graph nodes and elimination of cut points of the graph, after mutually adhered part point clouds are segmented, extracting a complete target point cloud from a scene according to a direction bounding box. The method has the beneficial effect that the adhered parts can be separated more effectively.

The invention belongs to the field of moxibustion therapy supplies, and relates to a medical moxibustion for a moxibustion therapy of irregular menstruation. The medical moxibustion is mainly characterized in that the surface of existing moxibustion wrapping paper is laid with moxibustion therapy landmarks and acupuncture-point names of corresponding acupuncture points and is provided with moxibustion therapy acupuncture-point graphs, arbitrary quantities of drugs such as aconitum napellus, galangal, anemarrhena asphodeloides, radix rehmanniae, codonopsis pilosula and cyperus rotundus are added into a core body of each acupuncture-point moxibustion therapy section. Compared with the prior art, the medical moxibustion can help to guide users to accurately grasp corresponding acupuncture points, positions, orders and time of the moxibustion therapy, the moxibustion is convenient to use and has no side effect, and the effect of the moxibustion therapy is significantly improved.

The invention belongs to the crossing field of information security and carrier-free information hiding, and discloses a generation type information hiding and recovering method, equipment and a medium. During hiding, a blindage is converted into a binary noisy point graph as a secret-free blindage; the positions of random secret embedding sample small blocks areobtained from the secret-free blindage by the secret key, candidate random sample small blocks with the number consistent with the number of binary noisy points of the secret embedding position sample small blocks are generated, and the sample small blocks consistent or having the maximum difference are selected as coding sample small blocks; and a secret-containing blindage is generated by replacing and scrambling the secret-embedded position sample small blocks with the coded sample small blocks. During recovery, a coded sample small block and a random secret embedding position are generated by a secret key, and secret information is extracted through comparison of the secret-containing blindage graph small block, the secret-free blindage graph small block and the coded sample small block. Compared with an existing method, the method of the invention has the advantages that modification embedding is not involved, splicing traces are not generated, a secret-containing blindage repetition mode is eliminated, a secret information coding space is hidden, and good anti-attack capacity and visual quality are achieved.

The invention provides a project progress visual management method and system, and the method comprises the steps: predicting the project data of each construction region at a future moment, and the project data comprises the number of construction people, a predicted completion construction period, and the number of remaining construction projects, wherein, according to the key point graph of the constructors at the current moment, the key point graph of the constructors at the future moment, and the predicted completion period and the remaining construction project number of each construction area at the current moment, predicting the predicted completion period and the remaining construction project number of each construction area at the future moment; and monitoring whether the project progress is abnormal or not according to the project data of each construction area at the current moment and the future moment. The predicted completion construction period and the remaining construction project number of each construction area at the future moment are predicted, the construction progress of constructors in the period from the current moment to the future moment is considered, and the predicted predicted completion construction period and the remaining construction project number are more accurate.

The invention discloses a sub-resolution graph adding method and device, equipment and a medium. The adding method comprises the following steps of intercepting a plurality of design layouts containing weak-point graphs from an original layout, conducting model-assisted rule-based SRAF graph addition on a region allowing addition of a sub-resolution graph in each design layout, and adjusting the shape of the SRAF graph according to a simulation result of mass production. According to the method, a model-assisted rule-based SRAF adding mode is adopted for part of weak-point graphs of which the process windows cannot meet the mass production requirement, so that the optical environment of the target hole graph in the weak-point graphs can tend to be consistent with that of the surrounding hole graphs, the process windows are enlarged, and the production requirement is met.

The invention relates to an interest point duplicate removal method and device based on a graph neural network, computer equipment and a storage medium. The method comprises the following steps: acquiring all interest points in a target territorial range to be subjected to duplicate removal, and constructing an interest point diagram based on a geographic position according to all the interest points; screening out multiple pairs of interest point repeated pairs from all interest points, and labeling the multiple pairs of interest point repeated pairs to obtain multiple pairs of seed interest point repeated pairs; iteratively training the graph neural network model according to the interest point graph and the multiple pairs of seed interest point repetitive pairs to obtain a trained graph neural network model; processing the interest point diagram through a trained graph neural network model, and determining all interest point repeated pairs in all interest points according to a processing result; and deleting any one interest point in each interest point repeated pair. According to the embodiment of the invention, the method can improve the deduplication effect of the points of interest, is efficient and reasonable, and is wide in application range.