1643 results about "Network construction" patented technology

The invention discloses a terminal D2D (device-to-device) cooperation relay communication implementation method in a TD-LTE-A (time division-long term evolution-advanced) system, and relates to a relay communication implementation method in the TD-LTE-A system. The method disclosed by the invention aims at realizing communication coverage on terminals in the TD-LTE-A system under a 4G (4th generation) standard and improving service quality and system throughput. According to the method disclosed by the invention, an on-net idle user in a cell is used as a mobile relay, terminal D2D cooperation relay communication is established with a user on the edge of the cell, relay service is provided for the user on the edge of the cell, the coverage range of the cell, the service quality of the user on the edge and the system throughput are significantly improved by resource allocation and time slot allocation of the cell, and the network construction cost of the system can be further greatly reduced. The method disclosed by the invention is suitable for the TD-LTE-A system.

The invention discloses an image semantic segmentation method based on deep learning. The method comprises four parts of data set processing, deep semantic segmentation network construction, deep semantic segmentation network training and parameter learning, and semantic segmentation on a test image. The RGB image and the gray level image of the input image are used as the input of the network model, the edge information of the gray level image is fully utilized, and the richness degree of input characteristics is effectively increased; a convolutional neural network and a bidirectional threshold recursion unit are combined, and more context dependency relationships and global feature information are captured on the basis of learning image local features; coordinate information is added tothe feature map through the first coordinate channel module and the second coordinate channel module, the coordinate features of the model are enriched, the generalization ability of the model is improved, and a semantic segmentation result with high resolution and accurate boundary is generated.

A method, apparatus, and medium for generating a grammar network for speech recognition and a dialogue speech recognition are provided. A method, apparatus, and medium for employing the same are provided. The apparatus for generating a grammar network for speech recognition includes: a dialogue history storage unit storing a dialogue history between a system and a user; a semantic map formed by clustering words forming each dialogue sentence included in a dialogue sentence corpus depending on semantic correlation, and generating a first candidate group formed of a plurality of words having the semantic correlation extracted for each word forming a dialogue sentence provided from the dialogue history storage unit; a sound map formed by clustering words forming each dialogue sentence included in the dialogue sentence corpus depending on acoustic similarity, and generating a second candidate group formed of a plurality of words having an acoustic similarity extracted for each word forming the dialogue sentence provided from the dialogue history storage unit and each word of the first candidate group; and a grammar network construction unit constructing a grammar network by combining the first candidate group and the second candidate group.

The invention discloses a method, a system and a multimode terminal for the realization of the network selection in a multi-system communication network. When a network selection control entity is configured in the multi-system communication network and the multimode terminal needs services, the network selection control entity controls the multimode terminal to selectively access a network according to the configuring selection policy and the obtained selection reference information, and the multimode terminal carries out services via the selected access network. According to the proposal provided by the invention, a suitable access network can be determined to supply the service connection for the multimode terminal by combining the selection policy with the selection reference information according to the relevant information of the access network such as the load condition, so as to make the appropriate access network bear appropriate subscribers and services and ensure reasonable use of network resources. Accordingly, the multi-system networks can be matched with each other in a cooperation mode so as to economize the network resources and guarantee the economy of the construction of the multi-system communication network.

The invention relates to a ribbed structure which comprises a ribbed composite wall plate, a hidden frame and a floor. Wherein, the ribbed composite wall plate is a network construction element formed by reinforced steel concrete and light material and divided by the reinforced steel concrete beam as rib beam and rib post in small sections with embedding light material stuffing blocks into the grid; the ribbed composite wall plate also comprises the goatee bar extending from the rib beam and rib post, which is four steel bars with certain anchoring length extending from each rib beam and post and is longitude steel bar whose end is a hook in connection to hook hidden frame; or else the goatee bars are two U-shape closed ring extending from each rob beam an post and is inserted with longitude steel bars in connection; the hidden frame is formed by outer frame post, connection post, and hidden beam which are embedded outside the ribbed composite wall plate while using common concrete, profiled bar concrete or steel structural beam and post; the stuffing material is made from light material with certain strength, volume weight and little elastic modulus; and the floor can select on-situ irrigating concrete, on-sit or prefabricated ribbed composite floor, pre-stress layered floor or special-shaped pre-stress hollow floor.

Provided are a method and apparatus for constructing a peer-to-peer (P2P) overlay network. The method of constructing a peer-to-peer (P2P) overlay network to obtain ID values of a plurality of nodes using a distributed hash table (DHT) and registering a node wishing to join the P2P overlay network where the plurality of nodes are sequentially located based on the obtained ID values, includes: (a) transmitting a join request message from the node wishing to join the P2P overlay network to a key node of the P2P overlay network; (b) a first node that received the join request message determining whether an ID value of the node wishing to join the P2P overlay network is a between value of an ID value of the first node and an ID value of a second node next to the first node, while sequentially transmitting the join request message to the plurality of nodes; and (c), if the first node determines that the ID value of the node wishing to join the P2P overlay network is the between value, registering the node wishing to join the P2P overlay network between the first and second nodes. Therefore, it is possible to more easily and efficiently search for sharing resources stored by each of the nodes.

A lightweight semantic segmentation method for a high-resolution remote sensing image comprises the steps of network construction, training and testing. Specifically, a deep semantic segmentation network of an encoder-decoder structure is constructed for a pytorch deep learning framework, after network training is carried out based on a remote sensing image data sample set, a to-be-tested remote sensing image serves as network input. A segmentation result of the remote sensing image is obtained. According to the method, on one hand, model parameters are reduced by decomposing depth separable convolution, the calculation complexity is reduced, the semantic segmentation time of the high-resolution remote sensing image is shortened, and the semantic segmentation efficiency of the high-resolution remote sensing image is improved; and on the other hand, semantic segmentation precision is improved through multi-scale feature aggregation, a spatial attention module and gating convolution, sothat the proposed lightweight deep semantic segmentation network can accurately and efficiently realize semantic segmentation of a high-resolution remote sensing image.

The invention discloses a method for realizing live video playback at an HTTP live streaming (HLS) client, which comprises the following steps: sending a program playback request and a program playback time point to a streaming media scheduling gateway media access platform (MAP) by the HLS client; receiving a playlist file dynamically generated by the streaming media scheduling gateway MAP according to the time point and a corresponding index file acquired from a storage module; and requesting video clips from a streaming distribution module according to the playlist file. The invention also discloses a system for realizing the live video playback at the HLS client, which comprises the streaming media scheduling gateway MAP, the streaming distribution module and the storage module. By adopting the method and the system for realizing the live video playback at the HLS client, the video playback can be realized at the HLS client without rewriting a hypertext transport protocol (HTTP) server, thereby, the network construction cost is saved, and the user experience is improved.

The invention discloses an automatic microseismic signal arrival time picking method based on a depth belief neural network. According to the method, each microseismic record is sampled according unified fixed dimensions, the signal arrival time of partial records is manually picked and is taken as the label information of the corresponding records; the information-picked records and labels of the information-picked records are taken as a total data set during network construction, including a training data set, a verification data set and a test data set; through inputting the data to the depth belief neural network for training and testing, the depth belief neural network is constructed; the actually-acquired to-be-processed data is inputted to the trained network model to carry out microseismic signal identification and automatic arrival time picking, and the network output is an arrival time point of the microseismic data.

The invention discloses a cost optimization unmanned aerial vehicle base station deployment method based on an improved genetic algorithm, and mainly solves the problem that the unmanned aerial vehicle base station deployment cost is difficult to optimize in the prior art. The realization method comprises the following steps: 1) establishing a ground wireless communication coverage model of the unmanned aerial vehicle base station; 2) calculating the maximum coverage radius and the optimal hovering height of the unmanned aerial vehicle base station in the unmanned aerial vehicle base station ground wireless communication coverage model scene; 3) deploying the unmanned aerial vehicle base stations at the optimal hovering height, enabling the deployment problem to be reduced from three-dimensional dimensionality to a two-dimensional plane, establishing an unmanned aerial vehicle base station deployment optimization model taking unmanned aerial vehicle base station deployment number optimization as a target, and solving the model to obtain an optimal chromosome; 4) converting the optimal chromosome into a corresponding unmanned aerial vehicle base station coordinate set to obtain an optimal unmanned aerial vehicle base station deployment scheme, reducing the complexity of the deployment problem, improving the solution accuracy, and being applicable to communication network deployment planning, temporary communication network construction and disaster area emergency communication.

The utility model discloses a distributed cluster-based network construction method for the wireless sensing network. Via nodes, the method exchanges information in a neighborhood scope and obtains partial network information, on the basis of which the nodes are calculated in a weighed way to get the weighed value for competitive cluster head, and then the cluster head is selected and clusters are formed to put an end to the cluster formation of the network. The invention brings the partial network information that is commanded by the nodes into full play, and obtains the weighed value of the most competitive cluster head in a weighed way, thereby realizing the most cluster-based optimization of the network and the load balance of the system. The nodes weighed value of the invention is a dynamic change in the cluster-based process, and a more optimized cluster-based result is accessible. Compared with the prior cluster-based network construction method, the invention has the advantages of easy realization of load balance, the elongation of network survival period and better performance, meets different requirements of different application circs on network performance, and has relatively good adaptability and extensibility.

The invention discloses a neural network construction method and device and an image processing method and device in the field of computer vision in the artificial intelligence field. The neural network construction method comprises the following steps: determining a search space and a plurality of construction units; stacking the plurality of construction units to obtain a search network, with the search network being a neural network for searching a neural network structure; and optimizing a network structure of a construction unit in the search network in the search space, wherein in the optimization process, the search space is gradually reduced, the number of the construction units is gradually increased, and due to the reduction of the search space and the increase of the number of the construction units, the video memory consumption generated in the optimization process is within a preset range; and establishing the target neural network according to the optimized construction unit. Under the condition that video memory resources are certain, the neural network meeting application requirements well can be constructed.

The invention discloses a human body key point detection method based on deep learning. The method comprises the steps of data acquisition, network construction, model training and evaluation, optimal model prediction and the like. According to the method, the ResNet50 network is improved, an expanded convolution residual network is provided, and a two-stage expanded convolution residual network is adopted to construct a human body key point detection network. During model training, feature extraction is performed on training data by the first-stage network, prediction is performed by using four channels, loss of all key points in a prediction result are calculated, and the loss is returned to adjust network parameters; the input feature map, the output feature map and the prediction result of the first-stage network are added by adopting an intermediate stage, and are transmitted to a second stage; and feature extraction is performed by the second-level network, prediction is performed on the finally obtained feature map after two-layer transposition, key point loss of a prediction result is calculated, the key point loss is sorted according to a descending order, and the first K * B losses are selected to return and adjust network parameters. An optimal training model is selected to predict human body key points of the to-be-detected image, the precision is high, and the practicability is good.

The invention discloses a deep learning network construction method and system applicable to semantic segmentation. According to the invention, based on the deconvolution semantic segmentation, by considering the characteristic that a conditional random field is quite good for edge optimization, the conditional random field is explained to be a recursion network to be fused in a deconvolution network and end to end trainings are performed, so the parameter learning in the convolution network and the recursion network is allowed to act with each other and a better integration network is trained; through combined training of the deconvolution network and the conditional random field, quite accurate detail and shape information is obtained, so a problem of inaccuracy of image edge segmentation is solved; by use of the strategy of combining the multi-scale input and multi-scale pooling, a problem is solve that a big target is excessively segmented or segmentation of a small target is ignored generated by the single receptive field in the semantic segmentation; and by expanding the classic deconvolution network, by use of the united training of the conditional random field and the multi-feature information fusion, accuracy of the semantic segmentation is improved.

The invention discloses an fMRI dynamic brain function sub-network construction and parallel connection SVM weighted recognition method which comprises the steps that (1) data are preprocessed; (2) the time series of each brain area is extracted; (3)interested brain areas are selected; (4) dynamic brain function sub-networks of all the brain areas are constructed; (5) all the sub-network classifiers are trained; (6) values are assigned to all sub-classifiers to form parallel connection SVM classifiers; (7) unknown samples are classified. Compared with a traditional static function network, information on the time dimension is added on the constructed dynamic brain function networks; prior knowledge is combined for constructing dynamic sub-networks on different interested brain areas, and the feature dimensions are reduced while useful information is reserved; SVM classifiers of all the sub-networks are trained, the parallel connection SVM classifiers is formed by determining the weight of the sub-classifiers through the recognition rate, the brain areas are integrally weighed and classified, and the classifiers have better robustness.

The invention provides a method and a device for detecting malicious HTTP requests, wherein the device comprises a network construction unit of Web access relationship and a detecting unit of malicious HTTP requests, wherein the network construction unit of Web access relationship is used for constructing a Web access relationship network for a to-be detected Web site, the network of Web access relationship embodies the fixed Web page access order of the Web site, the detecting unit of malicious HTTP requests is used for judging whether the HTTP requests which are sent to the Web site are corresponding with the fixed Web page access order of the Web site, if the requests are not corresponding with the order, the HTTP requests are judged as the malicious HTTP requests. The method and the device of the invention utilize the fixed Web page access order of the Web site to effectively detect the malicious HTTP requests.

The invention discloses a method for improving network layout simulation precision, which comprises: performing network transmission and distribution; processing after testing data; computing data based on path consumption, computing a wireless coverage field intensity of the network on a testing path and the Signal-to-Noise; correcting a cell wireless transmission model based on path consumption computing data, executing wireless network coverage prediction based on the cell wireless transmission model; counting the difference between an outputted network prediction result and a test computing result, and adjusting the network layout based on the difference. The invention is capable of improving the wireless network simulation precision, on-line measuring the correcting effect of the wireless transmission model, further, reducing operation pressure of network optimization, and saving network building cost for operators, the method of the invention is suitable for layout of various wireless cellular networks, and is capable of finding network design faults during the network layout process.

The invention discloses a healthy diet knowledge network construction method based on a neural network and a graph structure. The method comprises the steps that word vector modeling is performed on a text corpus, so that each non-stop word in the text corpus corresponds to one word vector with a fixed length; a cosine similarity between two word vectors is used to measure the relational degree between entities corresponding to the two word vectors; food material entity nodes and symptom entity nodes are extracted, the two types of entity nodes are regarded as entity nodes in a topological structure, edge relations between the entity nodes are constructed to form the graph structure, and all the edge relations between the entity nodes are described by one group of representative words; vector expressions corresponding to each representative word are arranged to obtain a representative matrix of the edge relations between the entity nodes; and a classification framework based on a deep neural network is designed, the representative matrix is input, and polarities of the edge relations between the entity nodes are classified. Through the method, the problems that a traditional healthy diet knowledge base is not high in automation degree and obvious in domain limitation are effectively solved.

The invention discloses a constructing method and a dynamically updating method of a GNSS network difference positioning reference station network. In the method, each GNSS reference station is projected to a two-dimensional plane according to the known and accurate coordinates of the GNSS reference station; a reference station network taking a triangle as a basic structure is formed by the discrete points of the two-dimensional plane according to the network constructing rule of a Delaunay triangle network, and the shape of the network is unique. When a new reference station is increased or a certain reference station is not used, the real-time construction of a new network is finished through operation of point insertion or point deletion. The invention can keep the unique optimized network shape of the reference stations and meet the precision requirement when a movable station is positioned.

The invention discloses an onset time automatic picking method of a microseismic signal on the basis of a time-recursive neural network. Each microseismic record is sampled according to a uniform and fixed dimension; then, the onset time behaviors of parts of record are manually picked to serve as the label information of a corresponding record; the record of picked information and the label of the record are used as a total dataset during network construction, wherein the total dataset is divided into three parts: a training dataset, a verification dataset and a test dataset; the data is input into a deep belief neural network to be trained and tested, and the time-recursive neural network is constructed; and the data which is not subjected to onset time picking is input into a trained network model, and the network outputs the data as a sequence corresponding to input data, wherein a first point which is not zero in the sequence is the onset time point of microseismic data.

The invention discloses a stacked noise reduction self-coding motor fault diagnosis method based on vibration and current signals, and the method comprises the following steps: 1, obtaining the time domain signals of the vibration and current of the motor during different faults, carrying out the preprocessing, and taking the processed signals as network input; 2, determining network parameters; 3, carrying out the layer by layer training, taking a hiding layer of an AE (Auto encoder) at an upper level as the input layer of an AE at a lower level, thereby obtaining a final feature code which is used for training a Softmax network; 4, carrying out the fine tuning of the whole network, judging whether the expected precision is met or not: finishing the training of the network if the expectedprecision is met, or else adjusting the network parameters, and repeatedly carrying out the step 3; 5, finishing the network construction. According to the invention, the multilayer SDAE network is constructed, and the vibration frequency domain signal and the current time domain signal are combined as the input. The SDAE network and a classifier are sequentially trained, and the supervised finetuning of the whole network is carried, thereby achieving the precise diagnosis of the fault of the motor.

The invention discloses a road longitudinal and lateral section obtaining method based on a LiDAR point cloud, and relates to the technical field of measurement. The method comprises the following steps: road data obtaining: generating filtered road surface point cloud data through the steps of field control, data collection, data preprocessing, coordinate conversion and point cloud filtering; data organization blocking: carrying out the block management of point cloud data according to a proper distance grid; TIN network construction through data: carrying out the buffering and constructing a TIN network according to an existing road central axis and a specific distance; section accomplishment calculation: generating a longitudinal and lateral section file according to a mileage file. The method is advantageous in that the method organizes the point cloud data through an engineering file, manages the point cloud data through the engineering file, carries out the automatic call of ground point cloud data in a corresponding range of the central axis of a road, and enables the data to be used for constructing the TIN network after format conversion. The seamless blocking method ingeniously solves a problem that point cloud data cannot be read in a unified manner because the size of the point cloud data is large.

The invention discloses a safety helmet wearing detection method based on depth features and video object detection, which comprises the following steps of video data acquisition; data marking of manual marking for the data collected by Step 1; data set preparation, wherein the data set consists of divided training set, test set and verification set, and each set contains pictures corresponding tothe original video, and the special training set and verification set also contain annotation data corresponding to each picture; the network construction and training of extracting features of key frames from input video and transferring them to different neighboring frames; transferring and multiplexing the key frame features to the features of the current frame by optical flow; the target classification and location frame prediction; network training, wherein the loss function of each ROI is the sum of cross entropy loss and boundary box regression loss.

Disclosed is an electricity consumption information acquisition system based on a home gateway. The system comprises an intelligent electricity meter, the home gateway, and a remote primary station, and is characterized in that the home gateway integrates electricity consumption information acquisition and an EOC functional module. Disclosed is also an electricity consumption information acquisition method, which comprises that: an acquisition instruction is sent by the primary station to the intelligent electricity meter through the home gateway; acquisition data is updated to a home gateway acquisition module by the intelligent electricity meter; the PLC, or micro-power wireless, or 485 data is converted to Ethernet data, and is converted to a coaxial signal through the EOC module; the coaxial signal is sent to an electricity primary station server through a broadcast television HFC internet, and data acquisition and analysis are finished. Flat design is adopted in the system and the method; a full-IP electricity consumption information acquisition system is built based on the home gateway; and a solution that a convention scheme of "intelligent electricity meter+collector+concentrator+primary station" is replaced with "intelligent electricity meter+home gateway+primary station", so that the acquisition framework is simplified; problems such as non-uniform acquisition standards are solved; and network construction becomes convenient and fast. Registered broadcast television network transmission is adopted, so that re-wiring is omitted and cost is saved.

Disclosed is a ring type wavelength division multiplexing (WDM) passive optical network (PON) system using the same wavelength for forward and backward channels while inexpensively implementing a redundancy function, and a switching media converter usable in the ring type WDM PON system. The WDM PON system provided a redundancy function to each node, so that the central office of the system can have a simplified architecture. Accordingly, there is an advantage in that the construction costs of the central office can be reduced. And, the network construction costs can be minimized because it is possible to provide a desired redundancy to a part of the nodes, taking into consideration the significance of each node.

An aggregation node device of a passive optical network (PON) is provided which includes an aggregation optical line terminal (OLT) and an aggregation optical network unit (ONU). The aggregation OLT is connected to a user-side ONU. The aggregation OLT aggregates service data transmitted by a user-side ONU and transmits the aggregated service data to the aggregation ONU. The aggregation ONU is adapted to transmit the received aggregated service data to a network-side OLT. A PON system is further provided. The device and system can not only support the conventional time division multiplexing (TDM) services but also support the services based on variable-length packets and the multicast service. Moreover, it is not necessary to build an equipment room and supply power for an intermediate optical distribution network (ODN) which greatly reduces the network construction and operation costs.

The invention discloses a rapid face detection identification method based on deep learning. The method comprises an offline network construction part and an online process design part. For the offline network construction part, a network structure designed in the method is distinguished from an existing method for using a single-scale template when face detection is carried out since detectors are respectively trained for different scales of faces, and the detectors with specific scales are trained and operate in a multi-task manner by using and constructing an image pyramid. For the online process design part, a real-time face feature buffer pool structure is established by designing a data structure thought of a list using FIFO, the mostly time-consuming real-time feature extraction part in the process is moved to the mostly front end of the whole process, an identity card reader for detecting identity cards is used as a trigger point, and a method for establishing a result matchingmapping table is proposed, innovation is carried out in three aspects, the time of the whole process is saved effectively, rapid face detection identification is realized, and accurate judgment whether persons and cards are unified is obtained.

The invention relates to a ship network construction method based on radio frequency identification and cloud computing. A network structure consists of a cloud platform, a middleware system, a ship management center, a data acquisition and transmission system, a network fixed base station and shipborne radio frequency equipment; a data acquisition system of a ground base station scans a shipborne radio frequency signal in an area and transmits the signal to the ship management center; the ship management center transmits data to a data analysis and storage system of the cloud computing platform in real time through the middleware system; and the cloud computing platform acquires related information of ships, converts the information into data information which can be identified by the ground base station, and returns the data information to the ship management center. An inland ship management information system which is constructed on the basis of radio frequency identification (RFID) technology is in seamless connection with other information systems through a distributed network of the cloud platform, so that distributed real-time data sharing and comprehensive ship information management among ship management departments in different areas and different forms can be realized; meanwhile, the application field of emerging cloud computing technology is widened.

The invention discloses an ultrasonic image super-resolution reconstruction method for improving contour definition based on an attention mechanism. The ultrasonic image super-resolution reconstruction method comprises the steps of S1, data acquisition; S2, network construction; S3, initializing a network; S4, network training; S5: super-resolution image reconstruction. On the basis of an existingfeature extraction reconstruction network, the method builds another level of parallel codes-codes; according to the attention mechanism network of the decoding structure, utilizing common convolution and cavity convolution, better obtaining high-frequency information in an ultrasonic image, combining the two levels of network features, and extracting the final image features by using convolutionto form a super-resolution reconstruction network. Through the two-stage parallel network, the attention mechanism network is used for positioning the specific position of the high-frequency information, the tissue interface and the tissue area in the ultrasonic image can be effectively distinguished, the edge reconstruction definition of the tissue contact surface in the ultrasonic image is improved, and the problem that the contour of the reconstructed ultrasonic image is fuzzy is solved.