67 results about "Type graph" patented technology

The invention provides a house type shape identification method, a device and equipment, and relates to the technical field of image processing. The method comprises steps of obtaining house type graph to be identified; inputting the house type graph into a pre-established house type shape identification model, and outputting a identification result of the house type graph, wherein the house typeshape identification model is established by the following steps: acquiring a set number of house type graph samples and corresponding identification information; establishing a network structure of the convolutional neural network; inputting the house type graph sample, the house type shape information and the structure information marked in the house type image sample into a convolutional neuralnetwork for training; carrying out adjustion and optimization on trained nerural network to obtain the trained house type shape identification model. The technical problems that in the prior art, when the number of house type graphs is large, an existing manual mode is low in identification efficiency and accuracy are solved, and the house type shape identification efficiency and accuracy are improved.

The invention discloses a scene roaming method, system and device based on three-dimensional modeling and a storage medium. The method comprises the steps: scanning to obtain three-dimensional data ofa scene, and uploading the three-dimensional data to a server; according to the three-dimensional data, generating a three-dimensional model of the scene and point cloud data of the scene by adoptingan artificial intelligence algorithm through a server; according to the point cloud data, generating a two-dimensional house type graph of the scene by adopting an artificial intelligence algorithm through a server; and roaming a scene according to the three-dimensional model, and displaying the two-dimensional house type graph. According to the invention, automatic construction of a three-dimensional model and automatic generation of a two-dimensional house type graph are realized by adopting an artificial intelligence algorithm through the server, so that application in the real estate industry is facilitated; according to the method, 360-degree immersive roaming experience without dead angles can be provided through three-dimensional roaming in the three-dimensional model, the automatically generated two-dimensional house type image can be displayed, a user can conveniently identify and understand the displayed house type, and the method can be widely applied to the field of three-dimensional modeling application.

The invention provides an indoor design method and device. The indoor design method comprises the steps that S1, a user enters a client for design; S2, a user draws a planar house type graph through aplanar graph drawing interface of the client, and after drawing is completed and decoration styles and tones are selected, the system uploads data to the server side; S3, artificial intelligence operation is carried out according to user requirements, an artificial intelligence algorithm is performed by the server side, furniture placement operation is stsarted, and a parameter matching process is performed on the furniture placement operation and preset decoration template data; S4, the client generates an indoor virtual reality scene according to a placement operation result and the matching result; S5, the client receives the input of a user according to the client in the virtual reality scene built in the step S4 according to the scene built in the step S4, carries out the replacement, addition, deletion or movement operation of the model and the material, and further perfects the indoor design; And S6, an effect picture and a 720-degree panorama picture are generated.

The invention provides a barrier identification method and apparatus, a computer device and a readable medium. The method comprises the steps of obtaining information of a to-be-identified barrier of N+1 continuous frames around a current vehicle scanned by a laser radar; according to the information of the to-be-identified barrier in each of the first N frames, obtaining first point cloud projection drawings, on a horizontal plane, of a point cloud layer with at least two heights, a first reflection information projection drawing, on the horizontal plane, of the to-be-identified barrier, and a first duty ratio projection drawing, on the horizontal plane of the to-be-identified barrier in each frame; and according to a pre-trained classifier model, a pre-obtained first barrier type graph, on the horizontal plane, of the to-be-identified barrier in each of the first N frames, and the at least two first point cloud projection drawings, the first reflection information projection drawing and the first duty ratio projection drawing of each of the first N frames, predicting a first barrier type graph of the (N+1)th frame. According to the technical scheme, the identification accuracy and efficiency of the to-be-identified barrier can be effectively improved.

The invention relates to the technical field of modeling, and provides a three-dimensional house type model generation method and device, equipment and a storage medium, and the method comprises the following steps: obtaining key points of a two-dimensional house type graph when the two-dimensional house type graph is received, the key points being turning points, connection points, bifurcations and/or endpoints appearing on the two-dimensional house type graph; determining a to-be-detected area corresponding to the key point, and marking pixel points meeting a preset condition in the to-be-detected area as wall body pixel points; based on a breadth-first search algorithm, determining a wall in the two-dimensional house type image according to the key points and the wall pixel points; andgenerating a three-dimensional house type model from the two-dimensional house type map according to the wall body. The key points in the two-dimensional house type graph are firstly extracted, the walls are gradually identified from the key points as starting points to the periphery, and finally, the three-dimensional house type model is constructed according to the walls, so that conversion fromthe two-dimensional house type graph to the three-dimensional house type model is realized, and a user can conveniently know the house type in detail.

The invention relates to a house type graph generation method and device, computer equipment and a storage medium. The method comprises the steps of acquiring the boundary of a target building and thelayout constraint of the target building; outputting a plurality of first house type graphs according to the layout constraint of the target building; screening a plurality of second house type images from the plurality of first house type images; for each second house type graph, applying layout constraints of the second house type graph to the boundary of the target building to obtain a layoutgraph of the target building corresponding to each second house type graph; and for the layout diagram of each target building, inputting the layout diagram of the target building and the boundary ofthe target building into a house type diagram generation network to obtain a predicted house type diagram of the target building output by the house type diagram generation network. By adopting the method, the quality of the generated house type graph can be improved.

The invention relates to the field of image detection, and provides a house type image recognition method and device based on a neural network, equipment and a storage medium, and the method comprisesthe steps: obtaining a required house type image when receiving a house type image recognition instruction sent by a query terminal, and carrying out the feature extraction of the required house typeimage based on a preset composite network model, and obtaining a required feature vector; performing house type graph query according to the demand feature vector to judge whether a matched target house type graph exists in a preset house type graph library or not; if yes, returning the target house type image to the query terminal; and if not, returning a corresponding house type graph non-existence prompt to the query terminal. According to the invention, the characteristics of the house type image are extracted through the composite network model, and the house type image is detected and matched through the extracted characteristics, so that the intelligent identification of the house type image is realized, the identification efficiency is improved, the manual workload is reduced, andthe labor cost is saved.

The invention provides a method for automatically identifying a wall body in a house type graph, which comprises the following steps of: carrying out position analysis on interference information in the house type graph, and eliminating the interference information in the house type graph by utilizing the maximum outer contour of the wall body; setting a segmentation threshold value according to bimodal characteristics of a wall gray histogram in the house type image, segmenting a wall area from the gray scale house type image according to the segmentation threshold value, and filtering an interference item according to the size of a connected domain area; and generating a vector data structure of the wall in a directional straight line fitting manner, so that the wall is identified from the house type graph. According to the method for automatically identifying the wall body in the house type graph, in the house type graph preprocessing stage, removal processing on interference information in the house type graph is added, the difficulty of follow-up wall recognition is reduced, and the accuracy of wall recognition is improved; the improved histogram double-peak method is utilizedto segment the wall body, and the situation of uneven gray scale caused by edge information damage is considered during threshold value selection, so that the wall body and the background can be distinguished better.

The invention discloses an AR glasses-based indoor decoration design method and system, and the method comprises the following steps: S1, obtaining a plane house type graph of a to-be-decorated houseand parameter characteristics of the plane house type graph, and generating a three-dimensional model of the to-be-decorated house; S2, establishing a three-dimensional virtual decoration article model for decoration articles required by house decoration; S3, the generated three-dimensional virtual decoration article model is rendered into a three-dimensional model of a to-be-decorated house, andan indoor decoration design scheme of the to-be-decorated house is generated; and S4, converting the decorated three-dimensional model into an AR scene model by utilizing AR glasses, and outputting and displaying the AR scene model in the AR glasses. According to the method and the system, a client can feel the decoration effect personally on the scene, the decoration design efficiency is improved, and the satisfaction degree of the client is enhanced.

The invention provides a carbon dioxide emission space gridding method, a carbon dioxide emission space gridding system, a carbon dioxide emission space gridding medium and a carbon dioxide emission space gridding device. The carbon dioxide emission space gridding method comprises the following steps: converting regional classified land utilization graphs into graphs in a first preset data format;converting the graph in the first preset data format into a graph in a second preset data format, and extracting to obtain a single land utilization type graph; constructing a regional boundary gridgraph of the regional classification type land utilization graph, and superposing the regional boundary grid graph with the single land utilization type graph to obtain a single land utilization typegrid graph; and calculating to obtain a single land utilization type carbon dioxide emission grid diagram. According to the carbon dioxide emission space gridding method, system, medium and device, non-point source carbon dioxide emission space gridding is conducted on the basis of the corrected regional classification land utilization data, and the accuracy and authenticity of non-point source carbon dioxide emission space distribution information are improved.

The invention relates to the technical field of image modeling, and provides a three-dimensional house type model generation method and device, equipment and a storage medium, and the method comprisesthe following steps: determining house type components in a two-dimensional house type graph when the two-dimensional house type graph is received; obtaining three-dimensional house type components corresponding to the house type components; erasing the house type components to obtain a two-dimensional house type graph only reserving a wall body so as to obtain the wall body; stretching the wallbody into a three-dimensional wall body according to the size marking information corresponding to the two-dimensional house type graph; and generating a three-dimensional house type model according to the three-dimensional wall body and the three-dimensional house type components. According to the invention, the house type components are recognized and erased to obtain the wall body, the three-dimensional house type model is generated according to the three-dimensional wall body corresponding to the wall body and the three-dimensional house type components corresponding to the house type components, conversion from the two-dimensional house type graph to the three-dimensional house type model is achieved, and clients can conveniently know the house type in detail.

The invention discloses a method and system for making a two-dimensional house type image based on a touch screen interaction terminal and a panorama. The method comprises the steps that the panoramaof each room of a house type is acquired through an image acquisition module; the image recognition module detects and recognizes the panoramas of the rooms to generate coordinate data of ground closed structure points and door double structure points of the rooms and plane bottom views of the rooms; the data graph conversion module converts the coordinate data of the ground closed structure points and the door double-structure points of the rooms into two-dimensional planar graphs of the rooms and outputs the two-dimensional planar graphs to the touch screen interaction terminal; a touch screen interaction module is used for calibrating ground closed structure points of all rooms and double structure points of doors, two-dimensional plane graph splicing of the adjacent rooms is conducted,splicing data is generated and submitted to a structure splicing module, the structure splicing module conducts splicing work, and a complete two-dimensional house type graph is output through a datagraph conversion module. Compared with the prior art, the calculation amount is small, the drawing process is fast, drawing can be conducted on site, the professional requirement is low, and the threshold of house type drawing can be lowered.

The invention discloses a method and a system for identifying a room type in a house type graph. According to the scheme, the method comprises the steps of 1) identifying door, window and wall information in the house type graph by using mask-rcnn; 2) speculating the boundary of each room according to the door and window wall bbox result of mask-rcnn; 3) combining the mask value of each bbox in mask-rcnn to calculate the probability value of the room (including a bedroom, a living room, a restaurant, a kitchen, a toilet, a balcony and like) to which each pixel point belongs in the house type graph; and 4) calculating the type and boundary of each room according to the boundary and the probability value of the room to which the pixel point belongs.

The invention discloses a house type image generation method based on an indoor image, which comprises the following steps: receiving a house type image generation request, and obtaining an initial image corresponding to the house type image generation request; identifying the initial image through a preset image identification model, and judging whether the initial image is an indoor image or not; when the initial image is an indoor image, detecting the indoor image through the preset image recognition model to obtain a house type generation assembly in the indoor image; and generating a three-dimensional house type graph according to the house type generation assembly and the position size information of the house type generation assembly. The invention further discloses a house type image generation device based on the indoor image, equipment and a storage medium. The terminal carries out image detection and recognition on the initial image, when the initial image is the indoor image, the terminal converts the indoor image into the standard house type image, and automatic generation of the standard three-dimensional house type image is achieved.

The invention belongs to the technical field of artificial intelligence, and discloses a house type graph intelligent identification method, which comprises the steps of obtaining a two-dimensional house type graph to be identified; identifying the two-dimensional house type image based on a preset wall body identification model, and outputting the wall body information in the two-dimensional house type image; and generating a three-dimensional house type model graph corresponding to the two-dimensional house type graph based on the wall body information, and outputting the three-dimensional house type model graph. The invention also discloses a house type graph intelligent identification device, an apparatus and a computer readable storage medium. According to the invention, the accuracyof wall body identification in the house type graph is improved.

The invention is applicable for the technical field of distributed information processing, and provides an asynchronous production line type graph query method which is applied to a distributed graph database. The asynchronous production line type graph query method comprises the steps of establishing a production line containing multiple querying steps based on a query request; detecting a buffer queue of each querying step and waking up multiple pre-fetching threads based on the capacity of rest elements of the buffer queues; establishing a thread pool for asynchronously managing connection and query of a back-end data layer, configuring the capacity of the thread pool for the production line, and obtaining a query result by the tail end of the production line. Whereby, according to the synchronous production line type graph query method, the query speed and the throughput of a client are increased.

The invention discloses a house type graph-based house point cloud model generation method and apparatus. The method comprises the steps of obtaining a planar house type graph of a building; traversing each image pixel point of the plane house type image, and determining an image category to which each image pixel point belongs according to the RGB color value of each image pixel point; acquiringpoint cloud data of the image category according to each image pixel point under the image category and a preset height value corresponding to the image category; and generating a house point cloud model according to the point cloud data of each image category. By adopting an existing planar graph file, the relationship between the planar graph and the space is distinguished by utilizing geometricfeature statistical learning, so that the point cloud model of a house can be obtained without any hardware, cost is lower and efficiency is higher. The point cloud model generated through the pointcloud data has more advantages in later rendering and detail expression compared with a body block model, and the design and construction efficiency can be further improved.

The invention discloses a project information display method. The method includes: reading project data of a project from a database; showing a stage name of each stage of the project and key indexesof each stage on a showing page; and in response to a first key index selected by a user on the showing page, reading, from the database, statistic data corresponding to the first key index at each stage of the project, and generating and showing one or more preset-type graphs on the basis of the statistic data of the first key index at each stage of the project, wherein the preset-type graphs contain change information of the statistic data of the first key index at each stage of the project. Therefore, the user can be enabled to intuitively understand change information of the key indexes ateach stage of the project.

The invention provides a house type recognition model training method and device and a house type repetition judgment method and device. The house type repetition judgment method comprises the steps of acquiring at least two house type images to be subjected to repetition judgment; inputting the to-be-duplicated house type graph into a house type recognition model, and obtaining a prediction valueof each to-be-duplicated house type graph output by the house type recognition model; based on the prediction value of each to-be-judged repeated house type graph, obtaining a to-be-judged repeated house type graph, and performing duplicate judgment on the at least two house type images to be subjected to duplicate judgment, thereby realizing extraction of the prediction value of the input housetype image to be subjected to duplicate judgment by using the house type recognition model, and performing duplicate judgment on the at least two house type images to be subjected to duplicate judgment based on the prediction value of the house type image to be subjected to duplicate judgment, thereby replacing manual duplicate judgment and improving the efficiency.

The invention provides a quantitative-evaluation method and device of a house type squareness degree. The method comprises: extracting house outer-wall contours from a house type graph of a to-be-tested house, and acquiring coordinate information of the house outer-wall contours; completing the house outer-wall contours to form a rectangle, and acquiring coordinate information of the rectangle; segmenting the rectangle into nine-square grids, and carrying out labeling of orientation information for each grid in the nine-square grids to form nine orientation grids; calculating the proportion ofa vacant area in each orientation grid according to the coordinate information of the house outer-wall contours and the coordinate information of the rectangle; and scoring a house type squareness degree of the to-be-tested house according to the proportion of the vacant area in each orientation grid. According to the method of the invention, quantitative evaluation of the house type squareness degree is realized, quantitative comparison of squareness degrees can be carried out among different house types, and subjectivity of artificial judgment is avoided.

The invention relates to an image recognition method based on deep learning, and the method comprises the steps: S1, obtaining the corner information and character information of a to-be-recognized house type image, inputting the corner information into a quadratic programming model, and obtaining the connection information of wall lines and doors and windows, wherein the quadratic programming model is pre-established and comprises a plurality of corresponding corner point connection rules and a model of constraint conditions of a position relationship between a wall line and a door and window; S2, inputting the connection information of the wall lines and the doors and windows into a graph theory optimization model, and obtaining region composition information of each room region in the house type graph, wherein the graph theory optimization model is a Graham algorithm model which is established in advance and is improved based on graph theory knowledge; and S3, matching the characterinformation with the region composition information to obtain an recognition result. The problems that in the prior art, the house type image recognition result precision is not high, the recognitionprocess calculation is complex, and the expansibility is poor are solved.

The invention provides a spectrum management method and system based on a smart grid double-layer cognitive network architecture. The method comprises the following steps: receiving a spectrum conversion request, performing spectrum allocation between power information terminals in the smart grid double-layer cognitive network architecture by using a graph coloring part of an interference reduction type graph coloring-simulated annealing fusion algorithm, obtaining real-time spectrum information, establishing an available spectrum matrix and an inter-cell spectrum interference matrix, establishing a local solution list, calculating the optimum spectrum allocation through a simulated annealing part of the interference reduction type graph coloring-simulated annealing fusion algorithm according to the established local solution list and the inter-cell spectrum interference matrix, and feeding back an optimum spectrum allocation result to the power information terminals. During the entire process, by taking the interference reduction type graph coloring-simulated annealing fusion algorithm as a spectrum management allocation core, an optimum spectrum allocation strategy of a smart grid double-layer network can be quickly acquired, and spectrum resources in the network can be fully used.

The invention belongs to the technical field of home decoration data sharing, and particularly relates to an online design data sharing method and system. The online design data sharing method comprises the following steps: 1) obtaining a basic house type graph; 2) transmitting the basic house type graph to a designer side; 3) the designer side drawing the basic house type graph to obtain a decoration drawing; 4) the designer side transmitting the decoration drawing to the client side; and 5) the client confirming the decoration drawing. According to the sharing method, a designer side and a client side can design home decoration drawings on line, and data sharing between the drawn decoration drawings and building materials can be achieved through the mall module; according to the method provided by the invention, different users can realize simultaneous or time-sharing visual modification of the same design data through the same system. In addition, the designer side and the client side can check the accurate engineering quantity list and construction quotation in real time, and the problem of design drawing reworking caused by construction cost is avoided.

The invention discloses a line draft type graph similarity judgment method which comprises the following steps: a preprocessing step: obtaining a vector line draft image to be judged, carrying out graying processing and dual polarization processing on the vector line draft image to be judged, enabling the vector line draft image to be judged to present a black-white-gray state, and adaptively scaling the vector line draft image to be judged to a standard size; a comparison step: dividing the vector line draft image to be judged into n*n square areas to obtain a square matrix corresponding to the vector line draft image to be judged; and projecting the square matrix in the row direction and the column direction to obtain pixel point distribution conditions which are linearly distributed inthe row direction and the column direction, comparing the pixel point distribution conditions with vector line draft images of each standard size stored in an image library, and judging the similarityof the vector line draft images. The invention further discloses electronic equipment and a storage medium. The line draft graph similarity judgment method and device can obtain an ideal and accurateline draft graph similarity judgment result.

The invention relates to a data visualization device and method, and the device comprises: a data management module which enables a user to select to-be-displayed data through the data management module; a chart class library which comprises a chart developer and is used for selecting chart types according to data needing to be displayed; and a chart development module which is used for carrying out visual chart development on the selected chart type, correspondingly generating a sub HTML page for each visual chart, and embedding the HTML page into another HTML page needing data display through Iframe marking. Due to the fact that the data visualization system and method are used for chart development, developers can directly develop the visualization chart in a configuration mode, complexcoding development work is not needed, development efficiency is improved, and the development threshold of data visualization is lowered.

The invention discloses a GraphX data caching method oriented to convergent graph application. The GraphX data caching method comprises nine steps of initialization, vertex calculation and identification updating, active vertex information collection, filtering opportunity judgment, vertex attribute synchronization, vertex active state synchronization, expired data filtering, and judgment of whether iterative calculation is finished or not and whether iterative calculation is finished or not. Aiming at the characteristic that the side scale in the graph is far greater than the vertex scale, the method focuses on cache optimization of side data, defines the side which is not used any more in the application running process as expired data, and identifies the expired data in the applicationrunning process through an indirect identification method taking the vertex as the center. The invention designs a data filtering method based on an active vertex scale change trend. When the scale ofthe active vertexes is remarkably reduced, filtering of expired data is carried out, the performance overhead of data filtering is reduced, meanwhile, it is guaranteed that the system can completelycache graph data needed by calculation into a memory space under the condition that memory configuration of the application is limited, and the execution efficiency of the application is improved.