184 results about "Merge algorithm" patented technology

The application of a watershed algorithm to pixels and their touch values obtained from a scan of a touch sensor panel to determine patches corresponding to images of touch is disclosed. Prior to applying the watershed algorithm, background pixels having little or no touch values can be eliminated. A primary merge algorithm can then merge adjacent patches together when the saddle point between them is shallow as compared to the peak represented by the patches. However, if two candidate patches for merging have a total number of pixels below a certain threshold, these two patches may not be merged under the assumption that the patches might have been caused by different fingertips. Conversely, if two candidate patches for merging have a total number of pixels above a certain threshold, these two patches can be merged under the assumption that the patches were caused by a single thumb or palm.

The invention provides a method and device for detecting road traffic abnormal events in real time. The method includes the steps of monitoring a road, obtaining a plurality of frames of continuous monitor images, extracting bright white segments from the monitor images, obtaining lane lines and lane end points through processing, building a lane model, determining a bidirectional detection area of a lane according to the lane model, detecting a moving object in the bidirectional detection area according to a Gaussian mixture model background subtraction method, determining the position of the moving object, building the mapping relation between the moving target and an actual vehicle according to the position of the moving target in the multiple frames of continuous monitor images by the adoption of a posterior probability splitting and merging algorithm and a feature point matching and tracking method, obtaining the running track and running speed of the actual vehicle, detecting the lane model and the running track and running speed of the actual vehicle according to a prestored road traffic abnormal behavior semantic model, and judging whether the road traffic abnormal events exist or not. The method has the advantages of being intelligent, high in accuracy and the like.

A three-dimensional rebuilding method based on laser and camera data fusion is characterized in that the rebuilding method comprises the following steps: (1) a C++ language is adopted to compile a laser data automatic collecting platform based on a radar principle; (2) a rotation matrix R and a translation vector T are arranged between a camera coordinate system and a laser coordinate system; (3) a motion object is extracted and subdivided by utilizing an optical flow field region merging algorithm, thus further subdividing the point cloud corresponding to the motion object; and (4) guarantee of point cloud and image texture on the precise positioning is provided for the three-dimensional image texture mapping, thus realizing the three-dimensional rebuilding system based on data fusion. The three-dimensional rebuilding method has the advantages of quickly and automatically collecting and processing laser data, quickly completing the external labeling of the laser and CCD, effectively extracting and subdividing effective point clouds of the object in complex environment and exactly completing the three-dimensional rebuilding of the laser and CCD data of the object.

A dynamic packet consolidation algorithm for a wireless transmission regime. At least first and second packets of a plurality of packets are sampled. The respective throughput times of each first and second packet are then calculated. The first and second packets are consolidated into a third packet if the sum of the corresponding throughput times exceeds a predetermined limit.

The invention relates to the field of wireless communication, and in particular relates to a method and device for detecting a system frame number of a long-term evolution (LTE) system. The method comprises the following steps of: equally dividing a scrambling sequence C of a TTI (transmission time interval) into four sections by using the self-decoding characteristic of PBCH (physical broadcast channel) data of a wireless frame, respectively descrambling with the PBCH data generated in the first step, after matching a descrambling rate, selecting whether to adopt a merge algorithm or not according to a condition, and then carrying out channel decoding and CRC (cyclic redundancy check), judging whether to continuously receive the data of a next wireless frame or not according to a CRC result, and determining the system frame number according to the adopted scrambling sequence. The device comprises a receiving module (10), a signal processing module (20) and an adjudicating module (30). The invention can improve the efficiency of a terminal in the blind check of the PBCH, decreases power consumption, and determines the system frame number as quickly as possible, so as to be convenient to the read of a subsequent system message.

The invention relates to a micro-strapdown altitude heading reference system and a working method of the micro-strapdown altitude heading reference system. The system comprises a data acquisition module, a data processing resolving module and a peripheral communication interface module, wherein the data acquisition module integrates a three-axis gyroscope, a three-axis accelerometer, a three-axis magnetic resistance sensor and an SPI (Serial Peripheral Interface) into an integrated micro-electronic mechanical system sensor; the data processing resolving module adopts a 32-bit ARM (Advanced RISC Machine) microcontroller; and the peripheral communication interface module consists of a serial port level switching chip, an RS-232 interface, a CAN (Controller Area Network) transceiver and a CAN interface. The working method of the system comprises the working steps as follows: the system is electrified and initialized; the system is self-tested; a magnetic field is self-calibrated; initial alignment is conducted; measured attitudes are combined: resolving the information of attitude headings according to initial three-dimensional altitude and altitude resolving algorithm of the strapdown altitude heading reference system, and resolving the three-dimensional magnetic altitude angle according to the compensated magnetic field information and acceleration information, conducting the information merging algorithm and finally obtaining the stable altitude heading reference information; and the altitude heading reference information is sent to application equipment.

The invention discloses a method for automatically extracting interesting image regions based on a human visual attention system, which mainly solves the problem that the prior method for extracting interesting regions cannot extract a plurality of interesting regions and edge information. The method comprises the following steps: calculating the local brightness contrast, global brightness contrast and edge of an input image; fusing characteristic graphs corresponding to the three characteristics through global nonlinear normalization merging algorithm so as to generate a contrast graph; calculating the position characteristics of the input image so as to establish a weighting graph; establishing a saliency graph corresponding to the input image through the contrast graph and the weighting graph; and segmenting out the interesting regions of the input image according to the saliency graph. The method can effectively extract a plurality of interesting regions in the input image, and is used in the technical field of image analysis and image compression.

A method and system for constructing a single topic-connected overlay network are disclosed. A link contribution array, which stores sets of edges in an order according contribution values, is provided. A highest contribution index indicates a highest element in the link contribution array. The method includes performing, at every iteration, a Greedy Merge (GM) algorithm for selecting an edge from the highest element in the link contribution array, removing the selected edge from the link contribution array, and adding the selected edge to a set of overlay edges. After the selected edge is added to the set of overlay edges, contribution values of other edges are updated. The GM algorithm terminates when all elements in the link contribution array become empty. As an output, the GM algorithm generates a single topic-connected overlay network for all topics. A Greedy Join (GJ) and Greedy Leave (GL) functions are also disclosed.

The invention provides a recognition method for English words and digits in a natural scene image. According to the method, recognition of the English words and the digits in a natural scene is divided into a feature extraction step, a feature focusing step and a feature recognition step, wherein a convolutional neural network is utilized to perform feature extraction on an input image, an attention mechanism is utilized to focus on useful information in a feature sequence, a long and short-time memory network is utilized to recognize feature vectors, therefore, a deep neural network and the attention mechanisms are combined, and a final recognition result can be obtained directly when the image is input into the deep neural network. Through the method, it is not needed to perform winding sliding operation on the input image or recognize characters in a window; and meanwhile, an output string is the final recognition result, so that it is not needed to integrate recognized strings through a merging algorithm.

The invention discloses a highway route automatic extraction method based on a remote-sensing digital image. The highway route automatic extraction method comprises the following steps of converting a remote-sensing image of a road of a confirmed region into a vector polygon image of a binary image of a spectral region of the road; removing a small object and conducting noise reduction processing; confirming a main length and a main direction of a polygon region; confirming sectional widths of all sides, confirming a central line of the road, and according to the central line and the width of the road, using a polygon overlapping merge algorithm in a buffer region to conduct equal-width expanding to obtain a road region; and finally generating a specification road network. The highway route automatic extraction method based on the remote-sensing digital image can effectively distinguish a road outline, efficiently extract information, and effectively conduct line processing on the road.

The embodiment of the invention discloses an LSM-tree (The Log-Structured Merge-Tree) index optimization method and an LSM-tree index optimization system. The method includes the following steps: data written into a memory are stored according to an LSM-tree memory index structure; when the memory data which are stored on the basis of the memory index structure meet a writing threshold condition, according to a first merging algorithm, the memory data are merged; the merged memory data are written into a disk file according to an LSM-tree disk index structure. The technical scheme of the invention solves the technical problems of excessive CPU (central processing unit) consumption, decrease in the outward throughput of a database, excessive I / O (input / output) operation, memory cache usage and the like exiting in the prior art which are caused in the scenario that same Key data items are updated frequently, consequently, CPU consumption is reduced remarkably, the outward throughput of the database is increased, and moreover, the problems of extra memory overhead, performance degradation and the like caused by the usage of a memory cache are prevented.

The invention discloses a three-dimensional model search method based on mesh segmentation. The method comprises the following step of analyzing and constructing a segmentation field through a hierarchical spectrum, and particularly comprises the steps of judgment of concave vertexes, constructing of a Laplacian matrix, matrix decomposition, selection of low-frequency feature vectors, generation of sub feature vectors, weight calculation of the sub feature vectors, and constructing of an edge symbol matrix. Contour lines are sampled in the segmentation field and are grouped and merged through a grouping-merging algorithm to obtain a plurality of candidate contour sets, the final segmentation boundary is determined according to the weight of each contour line in the candidate contour sets, and three-dimensional models are automatically segmented. Three-dimensional model mixing feature description sub-matrixes are obtained by calculating the feature description sub-matrix of each segmentation block of the three-dimensional models, the similarity of the mixing feature description sub-matrixes of each three-dimensional model in a three-dimensional model database to be searched and a target three-dimensional model to be searched is calculated, the similarity values of the three-dimensional models are ranked and output from low to high, and the three-dimensional model searching is achieved.

The invention relates to a mass airborne LiDAR point cloud Delaunay triangulation network parallel construction method. The method comprises the following steps: carrying out data block division and task distribution on mass airborne LiDAR point cloud data; applying a pointwise insertion algorithm to each data block respectively so as to construct a sub-triangulation network in parallel, and acquiring a sub-triangulation network set; using a merging algorithm to connect the adjacent two sub-triangulation networks so as to generate the merged triangulation network; using an optimization algorithm to carry out optimization on a newly generated triangle and the adjacent triangle so as to obtain the merged triangle which accords with a Delaunay rule; repeating the merging and the optimization until all the sub-triangulation networks are merged and the complete Delaunay triangulation network can be obtained. The invention also provides a corresponding apparatus for executing the above method. By using the method and the apparatus of the invention, a post-treatment speed of the mass airborne LiDAR point cloud can be substantially increased. The method and the apparatus of the invention have great significance for practical application.

The invention relates to a watershed algorithm-based high spatial resolution multi-spectral remote sensing image segmentation method, and aims to solve the segmentation problem of a high spatial resolution multi-spectral remote sensing image. The method comprises the following steps of: solving the gradient of the multi-spectral image by using a multi-spectral gradient algorithm to obtain a gradient image, then segmenting the gradient image by using a watershed algorithm, and finally performing region merging according to a merging algorithm by using the similarity between regions to obtain a final segmentation result.

The invention relates to a cooperation system framework and a consistency maintenance method thereof. The cooperation system framework is characterized by comprising a server unit and a plurality of client units, wherein the server unit is used for storing data sources, partitioning and merging areas of documents and meanwhile serving for Session management; the client units are communicated with the server unit, control the uploading / downloading process of the documents and perform area cooperative editing on the documents. The cooperation system framework and the consistency maintenance method thereof can partition the documents into mutually-disjoint task areas according to the editing requests of different teams, support the consistency of subdocuments by correcting local operation to execute a broadcasting process and obtaining an execution process in a long distance, and after the task is finished and obtain overall-consistency shared documents from partitioned documents through a document merging algorithm, thereby effectively reducing the interactive interference of editing among users, eliminating will conflict and improving the system operation efficiency.

The invention provides a visible light image insulator identification method and solves problems of poor universality and a non-high identification rate existing in a visible light image insulator under the complex background in the prior art. The method comprises steps that a training sample data set is constructed according to an original image; according to the training sample data set, the original image is detected through employing a BING algorithm, and an insulator candidate window is outputted; the insulator candidate window is identified through a convolutional neural network after training, and an insulator fine extraction window is outputted; the insulator fine extraction window is calculated through a high overlapping degree window iteration weight merging algorithm, and an insulator target window is outputted. The method is advantaged in that the problem of poor universality existing in the visible light image insulator under the complex background in the prior art is solved through the BING algorithm, and the problem of the non-high identification rate is solved through the convolutional neural network and the high overlapping window iteration weight merging algorithm.

The invention discloses a code coverage rate computing analysis method, a code coverage rate computing analysis device, equipment and a storage medium, and belongs to the technical field of computers.The method comprises the following steps of acquiring a code of a current version and commit id corresponding to the code from a code library via a git command; packaging and compiling the code, andstoring the code into a folder named by the commit id corresponding to the code; receiving a coverage rate file named by the commit id of the code, and storing the file into the folder named by the commit id; generating a test coverage rate port of the current code via a code coverage rate tool; generating a full coverage rate report via a combination algorithm; and generating an incremental coverage rate report via a git diff command. According to the method, device equipment and medium provided by the invention, combination of coverage rates between the different codes after the code is frequently changed is supported, and thus the certain version of accurate coverage rate file can be acquired.

Using a novel methodology for organizing information both in non-persistent and persistent computer storage devices, limitations common to traditional data storage models are eliminated. The invention combines two distinct—yet complementary—technologies, one for the organization of non-persistent storage and one for the organization of persistent storage. The invention allows records of unlimited dimensions containing data of any type and size, in any combination, to be constructed, maintained, and utilized in both non-persistent and persistent storage. In the persistent model, the invention also allows a user to define variable length frames, multiple hashing algorithms, and reuse of space caused by deleted records using an automatic gap consolidation algorithm.

Fast reconstruction methods are provided for 3D time-of-flight (TOF) positron emission tomography (PET), based on 2D data re-binning. Starting from pre-corrected 3D TOF data, a re-binning algorithm estimates for each transaxial slice the 2D TOF sinogram. The re-binned sinograms can then be reconstructed using any algorithm for 2D TOF reconstruction. A TOF-FORE (Fourier re-binning of TOF data) algorithm is provided as an approximate re-binning algorithm obtained by extending the Fourier re-binning method for non-TOF data. In addition, two partial differential equations are identified that must be satisfied by consistent 3D TOF data, and are used to derive exact re-binning algorithms and to characterize the degree of the approximation in TOF-FORE. Numerical simulations demonstrate that TOF-FORE is more accurate than two different TOF extensions of the single-slice re-binning method, and suggest that TOF-FORE will be a valuable tool for practical TOF PET in the range of axial apertures and time resolutions typical of current scanners.

The invention discloses a recognition and location method for an apple on the tree based on a TOF camera. According to the technical scheme, the three-dimensional imaging technology of the TOF camera with the resolution of 1280 (h)*1024 (v) and the model of OI-VS-1000 is used to acquire the reflected light intensity image and depth image of the apple, and the image is pre-processed; a selective contrast enhancement algorithm is used to improve the depth image; a morphological filtering algorithm is used to filter out noise in the depth map; a target fruit region and a background region are divided by a regional splitting and merging algorithm; a Canny operator is used to extract the boundary; circle detection is carried out on a boundary image by shape location Hough transform in the image to acquire a fruit circle and the center position; and the location technology of the three-dimensional database of the TOF camera is combined to extract the three-dimensional coordinates of each fruit center to acquire the actual radius value of the fruit, so as to recognize and locate the apple.

The invention discloses a robot spray workpiece surface modeling method. The method includes the steps of performing triangular grid partitioning on a workpiece surface, setting a maximum normal vector threshold, connecting triangular faces into small flat patches according to a triangular face connecting method, creating a flat patch adjacency graph, and merging all the small flat patches to a plurality of new large flat patches according to a flat patch adjacency graph merging algorithm. The method is applicable to surface modeling of various complicated sprayed workpieces, particularly applicable to surface modeling of workpieces large in curvature change and fast in calculation, so that preparations are made for follow-up spraying robot route planning and trajectory optimization so as to meet requirements on working of a spraying robot, and working efficiency of the spraying robot and quality of products can be improved.

The invention discloses a self-adaptive method and device for eliminating interference of multiple antennae through diversity and mergence. The method comprises the following steps of: measuring noise power of each receiving antenna; judging whether the ratio of the sum of diagonal elements of an NI (Noise Index) matrix of a data carrier to be detected to an average value of the noise power of all receiving antennae is larger than a preset threshold value; if so, carrying out interference suppression and mergence detection on the data carrier to be detected; and if not, carrying out noninterference suppression and mergence detection on the data carrier to be detected. Starting from an NI estimation matrix as the most fundamental factor influencing an IRC (Interference Rejection Combining) merging algorithm, whether the currently obtained NI matrix contains more interference components is judged through the specific numerical value characteristic of the NI matrix, and then, an IRC algorithm or MRC (Mixed Radix Conversion) algorithm is selected, so that the system obtains the optimal performance.

The invention provides a method and a system for detecting pedestrians in video streams. The method comprises the steps of obtaining multiple paths of video streams; performing moving target detection on each path of video stream; segmenting out a moving target region of each picture; splicing all pictures into a single large picture through a divide-and-conquer-based combination algorithm; obtaining a spliced image; and performing neural network prediction on the spliced image through a convolutional neural network to obtain a detection result. In the whole process, a moving target detection algorithm and a convolutional neural network pedestrian detection algorithm are efficiently combined together through the divide-and-conquer region combination algorithm, so that the pedestrians in the video streams can be efficiently and accurately detected.

The workload profiler and performance interference (WPPI) system uses a test suite of recognized workloads, a resource estimation profiler and influence matrix to characterize un-profiled workloads, and affiliation rules to identify optimal and sub-optimal workload assignments to achieve consumer Quality of Service (QoS) guarantees and / or provider revenue goals. The WPPI system uses a performance interference model to forecast the performance impact to workloads of various consolidation schemes usable to achieve cloud provider and / or cloud consumer goals, and uses the test suite of recognized workloads, the resource estimation profiler and influence matrix, affiliation rules, and performance interference model to perform modeling to determine the initial assignment selections and consolidation strategy to use to deploy the workloads. The WPPI system uses an online consolidation algorithm, the offline models, and online monitoring to determine virtual machine to physical host assignments responsive to real-time conditions to meet cloud provider and / or cloud consumer goals.

The invention provides an information processing method and system, and an auxiliary system. The information processing method comprises the following steps of obtaining description information and carrying out word segmentation operation on the description information to obtain key words of the description information; inquiring a first classification represented by the key words in an appointed search engine and obtaining a first score of the first classification; carrying out automatic text classification on the description information to obtain a second classification of the description information and a second score of the second classification; and adopting a preset merging algorithm to carry out consolidation on the first classification and the second classification to obtain a second classification after consolidation, and selecting the second classification with the highest score in consolidation as a classification result of the description information. By utilizing each embodiment in the invention, the accuracy for system classification and the business processing efficiency of the system can be improved.

The invention belongs to the technical field of electronic equipment identification method or device, and discloses a multi-spectral image change detection method based on probability segmentation andGaussian mixture clustering. The method comprises the steps of firstly, inputting two original multi-spectral images of the same region and different time, and constructing the mixed difference imageHDS by CVA and SAM; secondly, segmenting the difference image into super-pixel space by using statistical region merging algorithm, and finally, using a K-Means algorithm to initialize Gaussian mixture model to overcome its shortcomings of convergence to local optimal solution, fitting the probability distribution of super-pixel feature space, and obtaining the change detection results by using Bayesian discriminant rule based on minimum error rate. The invention makes better use of the amplitude change information and the angle change information of the spectral vector, obtains the local structure characteristic of the image by using the super pixel segmentation, and effectively improves the detection accuracy of the change region in the SAR image.