55 results about "Marked graph" patented technology

The invention discloses a method for carrying out augmented reality on a pair of virtual reality glasses by utilizing a monocular camera of a mobile phone for mainly solving a problem of single function of the existing technology. The technical scheme comprises the following steps of (1) establishing a geodetic coordinate system according to a marked graph and utilizing a rear camera of the mobile phone to shoot a scene containing the marked graph; (2) solving a homography transformational matrix between a plane of the marked graph and an imaging plane of the camera of the mobile phone and a projection matrix P of the camera; (3) establishing a virtual object three-dimensional model and two virtual cameras C1 and C2; (4) computing imaging I1 and I2 of the three-dimensional model and the scene under the virtual cameras C1 and C2; and (5) respectively displaying the imaging I1 and I2 on a left screen and a right screen of a screen of the mobile phone and taking the screen of the mobile phone as a display screen of the pair of VR glasses to complete an AR stereoscopic display effect. According to the method, through using the monocular camera of the mobile phone, a function of augmented reality stereoscopic display is completed by the pair of virtual reality glasses, and the method has the advantages of low cost and easy installation and can be used for virtual visualization of industry, education and entertainment.

The invention discloses a virtual object display method and device. The method comprises the steps that when a first operation instruction is obtained, rendering is conducted to generate a first interactive interface, wherein the first interactive interface comprises a first display area; a real scene image is obtained and displayed in the first display area; a second operation instruction of a user is monitored, and a marked graph is selected in the real scene image according to the second operation instruction; when a third operation instruction is received, three-dimensional model data corresponding to a virtual object is obtained according to the third operation instruction, and the virtual object is generated according to three-dimensional model data rendering; the virtual object is controlled to move along with the marked graph. Accordingly, the technical problem that in the related technology, obtaining of the virtual object lacks novel interactive modes, and the user experienceis poor is solved.

The invention relates to a laser marking method which applies a galvanometer scanning type laser marking device for laser marking. The laser marking method includes the steps that a first and second dimension coordinate system is built in a plane where a to-be-marked graph is located; the to-be-marked graph is decomposed into a plurality of sub graphs; an area where the to-be-marked graph is located and divided according to the marking range of the galvanometer scanning type laser marking device, and a plurality of marking areas are obtained; whether each marking area has corresponding markinggraphs is judged according to the position of each marking area and each sub graph, wherein the sub graphs or part of sub graphs placed in the marking areas are marking graphs; and the marking areasaccording to the marking graphs corresponding to the marking areas are marked when the marking areas has corresponding marking graphs. The invention also relates to a laser marking system. The laser marking method and thesystem impose no limitations on the size of the to-be-marked graphs, so that the difficulties of marking are reduced, the speed of marking is accelerated and the efficiency of marking is improved.

The invention discloses a photoetching layout, a photoresist graph and a method for measuring an exposure error of the photoresist graph. The photoetching layout comprises a first die graph, four first mark graphs, and a second mark graph, wherein the first die graph is a rectangle; the four first mark graphs are positioned on the outer sides of four top corners of the first die graph, and have L shapes; two edges of each L shape are respectively parallel to two adjacent edges which form one top corner of the first die graph; the second mark graph is positioned on the outer sides of one top corner of the first die graph and the corresponding first mark graph, and has a cross shape; and two edges forming the cross shape are respectively parallel to the two edges of the corresponding first mark graph. In the process of forming the photoresist graph by using the photoetching layout, the exposure error between two times of exposure on the same photoresist layer can be acquired through one-time measurement, so that measuring steps are reduced, and testing time is saved.

The invention discloses a semi-supervised semantic segmentation method based on maximum confidence, and the method comprises the steps: selecting a part of images from an existing training data set asmarked images, and taking the remaining images as unmarked images; constructing a network model, and predicting prediction class probability graphs of the marked images and the unmarked images through a segmentation network in the network model; maximizing the confidence coefficient of the marked image prediction class probability graph by adopting supervised learning and adversarial generation modes; predicting a segmentation error region in the unmarked image prediction class probability graph by adopting an unsupervised learning mode; training the network model by combining loss of supervised learning and loss of unsupervised learning; and in a test stage, inputting an unmarked image to be segmented into the trained network model to obtain a segmented semantic image. According to the scheme provided by the embodiment of the invention, semantic segmentation can be accurately carried out on the unmarked image.

The invention discloses a projection pursuit hyperspectral image segmentation method based on transfer learning, belonging to the technical field of image processing. The technical key points are as follows: the multiband characteristic of the hyperspectral image data is utilized to regard each wave band as a gray level image, each image is analyzed and researched; the characteristic that the images of different wave bands are similar but not the same to introduce transfer learning in the projection pursuit clustering algorithm; and a ground object marked graph is used to obtain the label of the source domain image data, and the known label knowledge is utilized to guide the image data without label in the object domain and obtain the optimal projection direction and the optimal subspace, thus increasing the segmenting precision. The method has the advantage that the priori knowledge is utilized to increase the segmenting precision, and can be used for military reconnaissance and in the civil and industrial fields.

The invention provides a detecting method, device and system for cutting precision of a laser, and relates to the technical field of cutting equipment detection. The method comprises the following steps of: obtaining a preset marking and detecting graph; controlling a marking pin assembly to mark on a detecting board according to the preset marking and detecting graph, thereby forming a marked graph; movably mounting the marking pin assembly on a portal frame of the laser cutter; arranging the detecting plate on a bed body surface of the laser cutter; controlling length of a marking and detecting graph practical feature of a projection detector; and determining cutting precision of the laser cutter according to errors between the length of the practical feature and length of a preset feature of the preset marking and detecting graph. According to the embodiment of the invention, the method avoids the problem that output efficiency of laser cutting equipment is severely affected as mounting and dismounting time for fittings such as a laser head, a laser unit and a water cooler is relatively long; and meanwhile, material cost is reduced.

The invention provides laser marking equipment and a focusing method thereof. The method comprises the following steps of roughly adjusting the distance between a focusing lens and a positioning zeropoint according to a first correction amount and acquiring a first marked graph; finely adjusting the distance between the focusing lens and the positioning zero point according to a second correctingamount when confirming that a first graph parameter of the first marked graph corresponding to a to-be-determined focus position is optimal; determining a second compensation amount corresponding tothe to-be-determined focus position as a focus compensation amount when confirming that the difference value between second pattern parameters of each of second marked graphs corresponding to each ofpattern distribution positions is smaller than a preset difference value threshold value and the second graph parameters of the second marked graphs corresponding to the to-be-determined focus position is optimal; and adjusting the focusing lens to the focus point position of the to-be-determined focus position according to the focus compensation amount. According to the equipment and the method,the accuracy of determining the focus point of the laser marking equipment is greatly improved, the operation is simple and convenient, and the operation stability of equipment is improved.

The invention discloses a device based on a variable-color fluorescent drawing board and augmented reality. The device comprises a mobile intelligent terminal with augmented reality, and AR marked graphs of AR information, wherein the AR marked graphs of the AR information are respectively a stick figure card and a color painting card which are arranged on two side surfaces of the variable-color fluorescent drawing board separately; the mobile intelligent terminal comprises a camera, a data terminal and a touch screen; the camera is used for scanning and acquiring the AR marked graphs of the AR information; the data terminal is used for running an APP augmented reality system, processing the AR marked graphs acquired by the camera to generate a virtual model of augmented reality, performing unfolding UV on the virtual model, and adhering a planar to-be-displayed graph onto the virtual model to obtain a virtual model to-be-displayed graph; and the touch screen is used for subjecting the generated virtual model of augmented reality and a real scene to overlapping display. The device enhances the flexibility and creativity of the AR experience, is conducive to enhancing balanced development of left and right brains, and improves children's thinking and judgment capabilities for spatial three-dimensional objects.

The invention relates to a pilot head pose detection method based on a video camera. A video camera with an upward lens is fixed on an earphone worn on the pilot head; and a marked graph is mounted inside a cockpit above the lens of the video camera, the marked graph is a white sign with a black base, and the sign is made of a fluorescent material. When the method is used for detecting the pose ofthe pilot head, firstly, the pilot puts the head straight and then presses an alignment switch, thereby acquiring a standard identification image through the video camera; then, when the pilot swingsthe head, the video camera shoots the marked graph in real time, the shot marked graph and the stored identification image in the standard state are superposed, and characteristic values are extracted; and the current pose of the pilot head is calculated through the conversion relation between the head position of the pilot and the image position, including the pitch angle, roll angle and courseangle of the head. The method provided by the invention has the advantages that the detection result is high in precision, and the detection speed is high; the method is suitable for head positioningof pilots of small general-purpose planes; and avionics systems of the planes do not need to be changed, so that the cost is low.

The invention discloses a cutter state image classification method based on an edge mark graph neural network. The method comprises the steps that S1, a high-definition electronic measurement microscope is used for shooting a cutter abrasion image when a machine tool is shut down during machining; s2, normalization and other processing are carried out on the cutter abrasion image, the image is input into an edge mark graph neural network to extract features, then adjacent edges are established according to the existing features and initialized, a full connection graph is formed, and each edgerefers to the relation type of two nodes connected with the edge; and S3, after the graph initialization is completed, updating the node features and the edge features, further obtaining the final node category prediction from the finally updated edge features, and obtaining the category to which the test set sample belongs by using a weighted voting method in combination with the sample labels inthe training set and the prediction values of the connecting edges of the sample labels and the test sample. The method has the advantage that the wear state of the cutter can be well recognized under the condition of small sample images.

The present invention provides a method and apparatus paper media content expansion in a mobile terminal with augmented reality technology. The system comprises three modules: a mobile terminal module, a server-side database module, and a marked graph of paper publication. The mobile terminal module further comprises three sub-modules: marked graph scanning and uploading, display of expanded content, and downloading of the expanded content. The paper media needs to be embodied with the marked graph in the content expansion area. The identification image ID needs to correspond to a displayed resource ID. The server module further comprises two sub-modules: a matching module of the marked graph, and a content expansion resource database for downloading by the mobile terminal. The method combines the mobile terminal and the augmented reality technology to expand content of traditional paper media. The expanded content comprises textual commentary, picture browsing, relevant video playing, and a three-dimensional model, thereby greatly enhancing the interest and richness of the paper media, and improving user reading experience at the same time.

The embodiment of the invention provides a graph neural network training method and device, and relates to the field of deep learning. The method comprises the steps of obtaining an unmarked graph data set and a first marked graph data set; taking the unmarked graph data set as a training sample, training a preset graph neural network model, and adjusting parameters of the graph neural network model to obtain a first graph neural network model; taking the first marked graph data set as a training sample, training the first graph neural network model, and adjusting parameters of the first graph neural network model to obtain a second graph neural network model; and taking the second marked graph data set of the to-be-applied scene of the graph neural network as a training sample, training a second graph neural network model, adjusting parameters of the second graph neural network model, and obtaining the graph neural network applied to the to-be-applied scene. When the scheme provided by the embodiment is applied to graph neural network training, the efficiency of graph neural network training is improved.

The invention provides a low-error and high-efficiency multi-mark graph augmented reality system, which can bind a plurality of AR virtual models on an AR scene binding object and display the AR virtual models as required. The system comprises AR equipment and an AR model memory connected with a camera module of the AR equipment. An AR virtual model associated with an AR scene binding object and amodel display triggering condition are stored in the AR model memory. Each AR virtual model corresponds to a unique model display triggering condition. The model display triggering condition is thatan AR scene binding object and a triggering object are captured by a camera module of the AR equipment at the same time, and when the AR equipment works, if the model display triggering condition is captured, the AR equipment demonstrates an AR virtual model corresponding to the triggering object. The system is suitable for an occasion where multiple pieces of virtual information need to be stacked on an object in sequence, can effectively reduce virtual-real registration errors caused by repeated pasting of the mark graph, greatly reduces the workload, improves the efficiency, and is convenient to use.

The invention relates to a connected domain marking method used for a projection interaction system, and solves the defect of low efficiency of the connected domain marking method in comparison with the method in the prior art. The connected domain marking method comprises the following steps that the connected domain of a first frame of image is marked, an image video is acquired, the connected domain of the first frame of image of the image video is marked and a marked graph corresponding to sub-rows is obtained; new pixels and disappear pixels are acquired; marking result correction is performed on the previous frame of image based on the disappear connected domain; and marking result correction is performed on the current frame of image based on the new connected domain. The marking function of the connected domain in the binary image can be rapidly completed in processing successive frames of images.

The invention provides a batch generating system and a batch generating method for making electronic maps. The batch generating system comprises a foreground terminal, wherein the foreground terminal comprises a batch generating module, preset two-dimensional coordinate system space and a marked graph; the batch generating module is based on an ArcGIS (geographic information system) platform and is used for making the electronic maps; the marked graph is used for vector diagrams of the electronic maps; the foreground terminal is connected with a background server through a network; and the background server comprises a database with vector data of a plurality of vector diagrams with unique marks. By the batch generating system and the batch generating method for making the electronic maps, the shortcomings that the electronic map generation efficiency in the prior art is quite low and requirements on high efficiency of the electronic industry cannot be met are overcome.

The invention discloses a watershed algorithm-based insect body identification counting method and system, and belongs to the technical field of image processing. The method comprises the following steps: collecting an image containing insect bodies, and carrying out binarization processing to obtain a binary image; deleting the outline with the area smaller than the minimum insect body area in the binary image, and filling the outline of the insect body to obtain a filled image; gradually corroding the filled image, marking all insect body contours, and expanding the marked background to obtain a first marked graph; further processing the filled image and the first mark graph by using a watershed algorithm to obtain a second mark graph; and obtaining the contour roundness and the contourarea in the second mark graph, and counting the number of contours to complete counting. Compared with a traditional manual counting mode for counting the number of the insect bodies, the method can automatically count the number of the insect bodies, is higher in efficiency and recognition counting accuracy, and is suitable for counting various insect pests.

The invention discloses a method for improving the accurate focal length of an exposure machine. The method comprises the following steps of: setting a figure into a mark figure of an exposure machineon a glass negative film, projecting the mark figure onto an alignment shape on a product by light of the exposure machine during exposure, and completing alignment by overlapping a mark figure projection center and an alignment shape center on the product. The design of a plurality of corners and the design of a plurality of parallel line segments are adopted in the graph, so that the center ofthe graph can be quickly and accurately grabbed; the number of the parallel line segments can be designed according to the size of the alignment shape on the product, so that the center of the line segment in the middle of the number of the parallel line segments of the precise focal length graph corresponds to the center of the alignment shape on the product, and the center of the focal length graph and the center of the alignment shape of the product can be quickly and accurately overlapped. By adopting the focal length graph, the situation that the alignment precision is reduced due to incomplete graphs caused by etching of the glass negative film graph can be avoided.

The invention provides a network panoramic image marking method. A client obtains a panoramic image from a server and performs display; a user performs marking on any graphic image target on the panoramic image displayed by the client and obtains a marked graph; the client stores description information of the marked graph and sends the description information of the marked graph to the server according to a certain rule or regularly; and the server stores client information and the description information of the marked graph. The marked graph is subjected to the same transformation when the client performs view point transformation and geometric transformation on the panoramic image, and corresponding display effects are presented. The marked graph can be applied to the display of the panoramic image and the sub-images of the panoramic image. The marks, generated by different users, of the panoramic image are stored; the users can view the marked graphs generated by themselves when browsing the panoramic image; and the marked graph control operations of addition, deletion, display, hiding, transformation and the like can be performed to achieve relatively good user expereince.

A quantum mechanical arithmetic operation method for machine vision, based on quantum dots is performed by a quantum processing processor. The quantum mechanical arithmetic operation method comprises, obtaining a first labeled graph connecting between feature points of the first image and a second labeled graph connecting feature points of the second image, generating a point-to-point combination by matching the feature points of the first image with the feature points the second image, generating a conflict graph by adding the largest point-to-point combination by comparing the point-to-point combinations with the threshold, generating non-constrained binary optimization equation for finding a maximum independent set of conflict graphs, converting the non-constrained binary optimization equation into Ising model of the quantum system, and calculating the Hamiltonian of Ising model based on the quantum dots to obtain solution of the non-constrained binary optimization equation.

A quantum mechanical arithmetic operation method for machine vision, based on orbital qubit is performed by a quantum processing processor. The quantum mechanical arithmetic operation method comprises, obtaining a first labeled graph connecting between feature points of the first image and a second labeled graph connecting feature points of the second image, generating a point-to-point combination by matching the feature points of the first image with the feature points the second image, generating a conflict graph by adding the largest point-to-point combination by comparing the point-to-point combinations with the threshold, generating non-constrained binary optimization equation for finding a maximum independent set of conflict graphs, converting the non-constrained binary optimization equation into Ising model of the quantum system, and calculating the Hamiltonian of Ising model based on an orbital qubit to obtain solution of the non-constrained binary optimization equation.

The embodiment of the invention relates to the field of advertising products, and provides a method for manufacturing advertisement used for publicity. The method comprises the following steps of: manufacturing images of each part conforming with the advertisement by using thread stands made of synthetic resins such as PVC, PE, PET and the like, and printing marked graphs or characters on the outer surfaces of the thread stands by a wire-mesh screen; manufacturing each thread stand into an outline frame of the advertisement; combining each thread stand conforming with the images of each part of the advertisement together to form the overall image of the advertisement; pasting connectors at the connecting positions of each thread stand; and connecting and fixing light-emitting devices which emit lights outwards in the connectors. The invention also provides the corresponding advertisement. The connecting parts of the advertisement are easy to be connected tightly and simply, and the product image is truly represented, thus facilitating the work of product publicity or improving the product publicity performance.

The invention relates to a marking model training method, device and system, a marking method, device and system, a storage medium and equipment. The marking model training method comprises the steps of obtaining an initial model of a marking model; executing an iterative operation; the iterative operation comprises the steps of training an initial model based on an initial training sample to obtain a preprocessing model; inputting the current to-be-marked image into the preprocessing model, and outputting a current preprocessing marking result; taking the preprocessing model as a new initial model, taking the initial training sample and the current training sample as new initial training samples, returning to execute the iterative operation until a preset condition is met, and taking the preprocessing model meeting the preset condition as a marking model; wherein the current training sample is the current to-be-marked graph and the current preprocessing marking result conforming to the preset standard. By adopting the technical scheme provided by the invention, the generalization ability and / or robustness of the marking model can be improved without additionally increasing the marking cost of excessive training samples.

The invention discloses a layout method and device of a mask and a mask. The layout method of the mask comprises the following steps that: chip patterns arranged in an array are formed on the mask; a cutting channel is formed between every two adjacent chip patterns, and the cutting channels are used for setting mark patterns, wherein the mark patterns at least comprise a first mark pattern; a set number of segmentation units of the first mark pattern is acquired according to a measurement alignment requirement of the first mark pattern; the set number of segmentation units are sequentially arranged on the cutting channels, so that the first mark pattern does not cover other mark patterns; and first mark pattern single bodies are set to replace at least two adjacent segmentation units, and the first mark pattern single bodies are arranged on the cutting channels, the first mark pattern single bodies are completely overlapped with a pattern formed by splicing at least two adjacent segmentation units. According to the layout method and device of the mask and the mask provided by the technical schemes provided by the invention, the problems of low efficiency and low accuracy of manual layout and marking of the graph can be solved.

The invention relates to a mark generation method and device, a model training method and device, a marking method and device, a storage medium and equipment. The mark generation method comprises the following steps: acquiring a to-be-marked graph of a target object; obtaining a conversion graph corresponding to the to-be-marked graph; acquiring a reference map of the target object; associating the conversion graph with the reference graph to obtain a conversion relation or a transformation relation; and mapping the fiducial mark based on the conversion relation or the transformation relation to obtain a mark. According to the technical scheme, automation of image marking is achieved.

The embodiment of the invention provides a graph-based efficient SPARQL query response method and device and equipment, and the method comprises the steps of: modeling a triple of an RDF data set into a directed marked graph, grouping and merging entity vertexes of the directed marked graph, and forming a multi-level abstract graph, wherein an object and a main body in the triple are the entity vertexes of the directed marked graph; constructing an auxiliary data structure, and constructing an abstract graph as an index in combination with entity vertexes connected with the same class in the directed marked graph, wherein the auxiliary data structure comprises a set of semantic inclusion relationships in a directed marked graph mode, and the directed marked graph mode is used for describing features of the directed marked graph; querying by accessing the abstract graph, and giving a response result. The query response effect of the method is far better than that of an existing query method.

The invention relates to a mine environment monitoring method, device and system, and the method comprises the steps: obtaining an aerial image of a to-be-monitored mine region collected by the flight of an unmanned plane, carrying out the brightness segmentation and color segmentation of the aerial image, and obtaining a plurality of brightness channel images and color channel images; further, feature graphs of the brightness channel image and the color channel image are extracted, graph contours of the feature graphs are determined, environment information collected by the environment monitoring equipment is marked for the graph contours, and marked graph contours are obtained; and finally, according to the brightness channel images and the color channel images corresponding to the marked graph contours, combining the marked graph contours to obtain a three-dimensional model of the mine. On the basis, related monitoring personnel can rapidly and visually know the mine environment through the three-dimensional model, and compared with a two-dimensional image, the three-dimensional model is more beneficial to overall monitoring and feedback of the three-dimensional environment such as the mine.

The invention discloses a printed circuit board design drawing generation method, a printed circuit board and a manufacturing method of the printed circuit board, and belongs to the field of printed circuit board design. The design drawing generated by the method for generating the design drawing of the printed circuit board comprises a first network pattern, a second network pattern, a first bonding pad pattern, a second bonding pad pattern, a first wire pattern, a second wire pattern and a first mark pattern. The simulation circuit of the design drawing is generated according to other graphs except the marked graph in the design drawing. In the simulation circuit of the design drawing, the first bonding pad pattern and the second bonding pad pattern are spaced from each other, so the first network pattern and the second network pattern are not connected together, the situation that two signal networks on the printed circuit board design drawing are connected together and cannot pass circuit detection can be avoided, and the accuracy of circuit detection is improved.

The invention provides a checkerboard corner automatic extraction method, system and device, and a medium, and the method comprises the steps of obtaining a checkerboard image which comprises at leastone marking pattern, and the marking pattern is located in a grid region of the checkerboard image; identifying a marked graph area in the checkerboard image, and further extracting position information of a marked graph in the checkerboard image; determining the positions of each black grid region and each white grid region according to the position information of the mark pattern; and determining the position of each checkerboard corner point in the checkerboard image according to the position information of each black checkerboard area and each white checkerboard area. According to the invention, the mark graph is arranged on the checkerboard image, and the positions of each black grid region and each white grid region are sequentially determined by extracting the mark graph, so that the extraction of the position of each checkerboard corner point is realized, the problem that when the checkerboard image is full of the whole camera field of view, the recognition of the corner points is often wrong is avoided, and thus realizing implementation of fast calibration of camera.