388results about How to "Build accurately" patented technology

The invention discloses a personalized commodity recommending method which integrates attributes and structural similarity. In the method, users and commodities are used as nodes with characteristic information to be mapped to a network by integrating the attribute information and structural similarity information, and an information network chart is established according to the purchasing relation between customers and the commodities; and interests and preference among user node pairs are measured by the integrated attributes and structural similarity in the information network chart, and the nearest neighbor is selected by the interests and the preference to improve the accuracy of recommending. On the basis of the recommending method, the invention also discloses a personalized commodity recommending method which integrates the measurement of the attributes and the structural similarity. In the system, the interests and the preference of the users are measured accurately by a computing method of integrating the similarity of the attributes and the similarity of node structure backgrounds in the information network chart, and the generation efficiency of the nearest neighbor is improved by utilizing clustering technology. The method and the system can be applied to electronic commerce, and provide personalized commodity recommending for the users.

A method for providing selected content items to a user. The selection of content items is based on metadata pre-assigned to content items, typically authored content metadata, and on metadata generated and associated afterwards, called derived content metadata. Additionally, the selection of content items can be based also on context metadata, particularly derived context metadata. Derived metadata are automatically generated on the basis of derivation rules corresponding to algorithms to be applied to, e.g., the content of content items, authored content metadata and context metadata. User profiles can be used for improving the selection quality. A method is also disclosed for building and maintaining user profiles based on machine learning techniques.

The invention relates to a device and method for composition based on a small quad-rotor unmanned aerial vehicle and belongs to the technical field of mobile robot positioning and navigation. The small quad-rotor unmanned aerial vehicle can rapidly enter a complex environment which a mobile robot cannot enter, carried laser radar is utilized to construct a two-dimensional map in real time according to an SLAM method, self localization and navigation of the unmanned aerial vehicle can be achieved by combining an IMU device and the like, and efficient exploration of a true complex area is achieved; the height of the small quad-rotor unmanned aerial vehicle can be conveniently adjusted to obtain two-dimensional maps on horizontal planes of different heights; the small quad-rotor unmanned aerial vehicle is rapid in movement speed and more flexible, movement and mapping of the small quad-rotor unmanned aerial vehicle are not subject to disturbance of obstacles on the ground, the range limitation of a detecting robot is broken through, the small quad-rotor unmanned aerial vehicle is extremely high in practical value, and accurate and rapid map construction can be achieved; compared with robot environment composition, the rotor wing robot can perform environment scouting more rapidly and more flexibly, and three-dimensional spatial images can be obtained.

The invention relates to a method of constructing a fine discrete road grid in an urban drainage simulation system and belongs to the cross field of municipal engineering information technology, database technology and geologic information system technology. From the demand on construction of a waterlogging simulation model, under the restraint conditions in road design specifications and on the basis of conventional map measured elevation points, the number and positions of interpolation points are optimized under the drainage simulation system platform by the fitting method of local elevation change curves of cross-longitudinal road section and by road characteristic terrain vector layers according to mathematical model interpolation encryption plane and elevation information of physical road tomography; accordingly, a fine discrete road elevation grid unit is constructed, true tomographic road features are expressed quickly and economically, the demand on precise simulation of waterlogging models is met, especially the dredging and resisting actions of road micro-tomographic features upon waterflow can be correctly expressed, and the method is significant to promotion and application of the drainage simulation technology.

A computer-based detection (e.g., speech recognition) system combines a word decoder and subword decoder to detect words (or phrases) in a spoken input provided by a user into a speaker connected to the detection system. The word decoder detects words by comparing an input pattern (e.g., of hypothetical word matches) to reference patterns (e.g., words). The subword decoder compares an input pattern (e.g., hypothetical words matches based on subword or phoneme recognition) to reference patterns (e.g., words) based on a word pronunciation distance measure that indicates how close each input pattern is to matching each reference pattern. The subword decoder sorts the source set of reference patterns based on a closeness of each reference pattern to correctly matching the input pattern based on generated pattern comparisons. The word decoder and subword decoder each provide an N-best list of hypothetical matches to the spoken input. A list fusion module of the detection system selectively combines the two N-best lists to produce a final or combined N-best list. The final or combined list has a predefined number of matches.

The invention discloses a target detection method in a Codebook dynamic scene based on superpixels. The method is characterized by comprising the following steps that (1) a superpixel partition method is used for partitioning video frames, K superpixels are obtained by partitioning; (2) a Codebook background modeling method is used, a Codebook is established for each superpixel partitioned in the step (1), each Codebook comprises one or more Codewords, each Codeword has the maximin threshold values during learning, the maximin threshold values are detected, background modeling is completed; (3) after background modeling is completed, currently-entering video frames are subjected to target detection, if a certain pixel value of the current frames accords with distribution of the background pixel values, the certain pixel value is marked as the background, otherwise, the certain pixel value is marked as the foreground; finally the current video frames are used for updating the background model. The method solves the problems that a traditional Codebook background modeling algorithm is large in calculated amount and high in memory requirement, and established Codewords are not accurate are solved, target detecting accuracy and speed are improved, the requirement for real-time accuracy is met, and accordingly the requirement for intelligent monitoring in real life is met.

The invention provides an airborne laser point cloud classification method based on a high-order conditional random field. The airborne laser point cloud classification method specifically comprises the following steps: (1) point cloud segmentation based on DBSCAN clustering; (2) point cloud over-segmentation based on the K-means cluster; (3) construction of a point set adjacency relation based onthe Meanshift clustering; and (4) construction of a point cloud classification method of a high-order conditional random field based on the multi-level point set. The method has the advantages that:(1) a multi-layer clustering point set structure construction method is provided, and a connection relation between point sets is constructed by introducing a Meanshift point set cluster constrained by category labels, so that the categories of the point sets can be classified more accurately; (2) a multi-level point set of the non-linear point cloud number can be adaptively constructed, and information such as the structure and the shape of a point cloud target can be more completely represented; and (3) a CRF model is constructed by taking the point set as a first-order item, and higher efficiency and a classification effect are achieved, so that a higher framework is integrated, and a better effect is obtained.

The invention discloses a three-dimensional geological model constructing and integrating method for a mountain city area, and belongs to the field of three-dimensional geological models. The construction of the three-dimensional geological model for the mountain city area is achieved through the method; due to the fact that the features of the geological structure of the mountain city area are combined, according to the process of constructing a sectional framework, drawing a section and constructing a geological model and the method for constructing the geological model by utilizing the section, the construction of the three-dimensional geological model for the mountain city area is completed more accurately. According to the three-dimensional geological model constructing and integrating method, visualization integration of the three-dimensional regional geological model is further achieved, and a data basis is provided for application analysis of geological topics. Moreover, with regard to the result data of the formed three-dimensional regional geological model, visual integration simulation on a three-dimensional digital city platform is achieved, and real-time geological analysis of the three-dimensional geological model can be achieved in a three-dimensional scene.

The invention relates to a method for producing a three-dimensional object by solidification of a material solidifiable under the action of electromagnetic radiation by means of energy input via an imaging unit comprising a predetermined number of discrete imaging elements (pixels), The method comprises performing solidification with exposure using bitmap mask. The bitmap mask may be formed from a stack of bitmap data provided by an overlap analysis of a three-dimensional volume completely or partially enclosing a three-dimensional model of at least a part of the three-dimensional object to be produced. Alternatively, bitmap mask may be formed from a two-dimensional data set comprising overlap information. Solidification may be performing with exposure using bitmap mask generated “on the fly”. The invention is also directed to devices, and a computer and a data carrier useful for performing or executing the method.

The invention discloses an indoor map construction method for a mobile robot integrated with multiple sensors. The indoor map construction method comprises the following steps of respectively acquiring distance information of the mobile robot and an anchor point, gesture information of the mobile robot and environmental information by UWB, a speedometer and laser radar; building a top point-constraint map according to the distance information, the gesture information and the environmental information; optimizing the top point-constraint map by a map optimization algorithm, and acquiring optimized track data of the mobile robot; and building a raster map according to the optimized track data of the mobile robot and the environmental information. The speedometer is used for providing accurate gesture change within a short time of the robot, long-time accurate gesture change can be provided by integrating UWB positioning information, map construction is performed by combining the data ofthe laser radar, and the problem of poor accuracy during constructing indoor complicated environmental map by the laser radar is solved.

The invention discloses a three-dimensional model display system, which is divided into a client side, a server side and a data service side, wherein the client side is used for achieving displaying of operation data in a three-dimensional scene; the server side is used for achieving integration of operation data; and the data service side is used for providing a source of the operation data. The invention further correspondingly provides a method for achieving three-dimensional model scheduling. By the display system and the scheduling method disclosed by the invention, high-precision original three-dimensional model files with different formats, which are submitted by different design units can be split and optimized into a model format which can be supported by a three-dimensional model scheduling engine; and three-dimensional models of all production equipment on the display system can be displayed on the same scene.

The invention discloses an intelligent underwater robot. The intelligent underwater robot comprises a communication module, an image obtaining module, an object identification module, a control module, a space construction module and an execution module; the communication module receives a control command of a remote control terminal; the image obtaining module obtains an underwater environment image; the object identification module identifies underwater objects and barriers from the underwater environment image on the basis of a first deep neural network model; according to identified underwater objects, the control module determines whether there is a target object, and if YES, a pixel coordinate of the target object is obtained; the space construction module processes the underwater environment image on the basis of a second deep neural network model, constructs an underwater environment 3D space of the present position of the underwater robot, and determines position information of the target object; the control module further plans a movement path for tracking the target object of the underwater robot according to the constructed 3D space as well as a preset tracking mode; and the execution module controls the underwater robot to move underwater according to the planned movement path. Thus, the target object is tracked and the barriers are avoided.

Embodiments of the invention disclose a method and a device for information recommendation based on conversation interaction, a server, and a storage medium. The method comprises: according to user behavior characteristics on historical reading interfaces and / or a user model, generating a reading theme inquiry message, the user model including at least one of the followings: user interest, user concerned topics, and user concerned authors; according to the generated reading theme inquiry message, performing conversation interaction with a user; according to contents of the conversation interaction, determining a reading theme of the user; and pushing information matched with the reading theme to the user. The method connects different topic information streams through natural language understanding and artificial intelligence technologies based on conversation interaction, to help users to explore potential interest and improve user duration and viscidity. A user model is established more accurately and controlled, so as to improve user experience.

A system to meter multiple agricultural products according to an independently prescribed rate for each in a variable-ratio blend of the agricultural products from a single opener in a plurality of such systems and openers across an applicator such as a drill. Bulk storage compartments associated with the applicator deliver multiple agricultural products to metering assemblies mounted in clusters or pods across the applicator. The agricultural products are fed from the metering assemblies, via a flow re-director, into a manifold, and then into a corresponding single opener having conduits to transport the agricultural product into the soil. Controllers independently regulate metering by the metering assemblies. The flow re-director and manifold provide blending of combinations of the agricultural products for each opener according to a field prescription.

The invention discloses an image retrieval method based on multiple intelligent algorithms and an image fusion technology. According to the method, on the basis of a reinforced learning and genetic algorithm and a clustering algorithm, an intelligent learning framework for active learning is constructed; the image fusion technology and the genetic algorithm are used during relevant feedback, so that an inquiry vector and a similarity matching model are corrected; and the inquiry precision and the inquiry efficiency are improved. The image retrieval method has the advantages of higher inquiry precision and inquiry efficiency and relatively high robustness for translation, rotation and scale transformation; furthermore, after a certain number of inquiries and learning, intelligent retrieval can be realized; and the inquiry precision and the inquiry efficiency are further improved.

The invention relates to a positioning device, a positioning method, and automatic walking equipment. The positioning device comprises a first positioning module, a sensor module and a processing module, introduces a pedestrian dead reckoning technology independent of the external environment into the positioning process for a boundary, so that the pedestrian dead reckoning technology is integrated with other positioning technologies to construct the virtual boundary, the positioning precision is high, the constructed boundary is precise, a physical boundary does not need to be laid out, and the complexity of user operation is reduced.

The invention provides an event-oriented dynamic knowledge graph construction method. The event-oriented dynamic knowledge graph construction method comprises the following steps: data acquisition; event trigger word extraction, wherein trigger words in the first to-be-processed data are obtained through training of a first extraction model; event element identification: event elements in event sentences are acquired through a second extraction model, and the event sentences comprise trigger words; event relationship extraction: the event elements are processed through a third extraction modelto obtain event relationships among the event elements; the event ontology is constructed, the co-reference relationship in the event relationship is identified, the co-reference events are merged, and in addition, the invention further provides electronic equipment capable of realizing the method. According to the event-oriented dynamic knowledge graph construction method and device provided bythe invention, for event representation and processing, the entity knowledge base is used as a basis and a template, and the the event-oriented dynamic knowledge graph construction method and device have unique constituent elements and architecture and can be associated with the entity knowledge base, so that more accurate dynamic knowledge graph construction is realized.

The present invention relates to a construction method and a construction system for a stock standard news library. The construction method comprises: acquiring stock news information; constructing a stock standard library Z0; pre-compiling texts in the stock standard library; marking text features, and selecting feature words; computing information gain of the feature words so as to obtain feature values of the feature words; training and supporting an SVM algorithm by use of the feature values so as to obtain a classification module for extracting news features; updating the stock standard library by use of the classification model; updating and adjusting news composition in the stock standard library; and determining whether grabbed news information is positive news or negative news so as to construct the classification model, and continuously updating the news information in a database by virtue of a vector machine algorithm so as to realize the construction of the news library. The data base construction method is simple, is convenient in implementation, and is high in news data processing efficiency, and data information is updated in time and is accurate.

The invention relates to the technical field of compression perception and specifically relates to a segmented self-adaptive regularized matching pursuit reconstruction method based on a threshold. The method includes: setting initial state values of different parameters in a sparse signal reconstruction process; calculating the inner products of an iterative allowance r and each column of a sensing array Phi, that is correlation coefficients; finding atoms which satisfy a condition in the sensing array; storing in a subscript set J; ranking correlation coefficients of atoms corresponding to subscripts in the subscript set J from large to small; updating a support set which represents original signals; performing signal approximation through adoption of a least square method and updating the allowance; and performing iteration determination. The segmented self-adaptive regularized matching pursuit reconstruction method based on the threshold combines segmented self-adaptive atom selecting and regularization ideas. The method does not need sparseness as a prior condition in a signal reconstruction process and is self adaptive to approximation sparseness information and capable of constructing the support set accurately and completing precise signal reconstruction and the precise reconstruction rate is higher than the prior method of the same kinds so that the method is higher in practical applicability.

A method of making an article includes determining the greatest dimension of an article along a selected axis, and manufacturing a first part of the article between an end and a face lying in the laterally extending plane containing the dimension. The method further includes placing the article on the support of direct layer deposition (DLD) apparatus with the face upwards, and forming the remaining part of the article using the DLD apparatus.

A vehicle door structure for suppressing permanent deformation due to thermal distortion of a door is described. A door inner panel 18 and a door outer panel 20 are configured in aluminum alloy or the like. An impact beam 26 and a belt line inner reinforcement 44 are joined to the door inner panel 18. The impact beam 26 and the belt line inner reinforcement 44 are configured from a material, such as steel, with a smaller linear expansion coefficient than the door outer panel 20. Therefore, even if the door 42 as a whole is, for example, placed in a high temperature environment such as a bake oven in electrophoretic coating, the door inner panel 18 side and the door outer panel 20 side are restrained, respectively, by the belt line inner reinforcement 44 and the impact beam 26. Thus, permanent deformation due to thermal distortion at the door outer panel 20 side, due to a difference in the thermal expansion conditions of the door inner panel 18 side and the door outer panel 20 side, may be suppressed.

The invention belongs to the field of medical rehabilitation technology, and discloses a virtual reality cerebral apoplexy upper limb rehabilitation training system and a control method thereof. The system comprises a scene constructing module, an action data acquisition module, a parameter configuring module, a central control module, an expert diagnosis module, a rehabilitation reminding module,a massage module, a VR displaying module and a training data generating module. According to the system of the invention, a rehabilitation plan which is made by medical caring personnel is compared with a stored rehabilitation plan through the rehabilitation reminding module; on the condition of no conflict, the rehabilitation plan is stored and preset reminding time is set, thereby facilitatingaccuracy in making the rehabilitation plan; furthermore through reminding a patient at preset reminding time, executing efficiency and accuracy of the rehabilitation plan can be improved; and furthermore virtual training can be realized through the scene constructing module, the action data acquisition module and the VR displaying module. The virtual reality cerebral apoplexy upper limb rehabilitation training system and the control method thereof have advantages of reducing training space occupation space, enriching training modes and improving rehabilitation training effect.

The invention discloses a floor sweeping robot based on a depth camera. The floor sweeping robot comprises a shell, the depth camera, a sensor device, a walking mechanism, a cleaning device, a drivingassembly and a control device. The depth camera is arranged on the shell. The control device positions the floor sweeping robot and establishes a surrounding map according to color information and depth information of a space environment acquired by the depth camera, and a sweeping route is planned on the basis; the driving assembly controls the floor sweeping robot to move according to the planned sweeping route, and the movement direction of the moving robot is adjusted in real time in the moving process according to surrounding sensing information acquired by a sensor array; the control device is further used for controlling the cleaning assembly to execute corresponding sweeping through the driving assembly according to a sweeping mode set by a user. The floor sweeping robot can capture richer space information, thus, the map is accurately established, the sweeping route is planned, and the cost is greatly reduced.

The invention relates to an intelligent terminal-based virtualized fitting method which is applied to a virtualized fitting system formed by an intelligent terminal, a third party fitting platform and at least one apparel design terminal; an apparel factory sends apparel parameters of all apparel samples to the third party fitting platform, the third party fitting platform builds apparel sample virtual fitting models, a client sends own actual posture parameters to the third party fitting platform, and the third party fitting platform builds a human body virtual fitting model of the client after building a human body three dimensional sample model in advance; recommended apparel samples matching with the actual posture parameters of the client, corresponding recommended apparel sample virtual fitting models, all stored apparel sample parameters and the apparel sample virtual fitting models are obtained after screening operation and then sent to the intelligent terminal of the client; the client puts favorite apparel sample virtual fitting models on the human body virtual fitting model of the client on the intelligent terminal, and whether dressing effects are satisfying can be observed.

The invention discloses a building method for a semantic knowledge base based on a dictionary. The method comprises the following steps that a knowledge database based on a dictionary semantic knowledge expression model is built; semantic knowledge paraphrased by semantic items are extracted according to a semantic knowledge extraction rule; the semantic item mapping for extracting the semantic knowledge is realized; the rest semantic knowledge paraphrased by the semantic items are extracted. The method provided by the invention can be used for flexibly and fast building the knowledge base containing rich semantics from the dictionary, and in addition, the knowledge expressed in the knowledge base has universality and authority. The invention simultaneously provides a deep semantic knowledge expression method and semantic similarity calculation method for semantic information processing of natural languages, wide application values are realized in the natural language processing field such as the information retrieval field, and the retrieval based on the semantics instead of keywords can be realized, so that the deep semantic retrieval result closer to the requirements of the user can be obtained.

The invention discloses a carotid artery three-dimensional image building device. The carotid artery three-dimensional image building device comprises a handheld scanning auxiliary device, an ultrasonic probe, an image frame capture module and a processing system. The ultrasonic probe is fixed to the handheld scanning auxiliary device; the handheld scanning auxiliary device drives the ultrasonic probe to move to enable the ultrasonic probe to be capable of continuously and stably collecting carotid artery two-dimensional slice images at different positions at a first speed; the ultrasonic probe is connected with the image frame capture module through a first data connecting line; the image frame capture module collects the carotid artery two-dimensional slice images transmitted by the ultrasonic probe at a second speed and transmits the collected carotid artery two-dimensional slice images to the processing system through a second data connecting line; the processing system extracts the carotid artery two-dimensional slice images collected by the image frame capture module at a third speed; and the carotid artery two-dimensional slice images are built into carotid artery three-dimensional images. By means of the carotid artery three-dimensional image building device, the carotid artery three-dimensional images used for auxiliary diagnosis can be accurately built.

The invention provides a stimulation based detection method for SiGe HBT single event effects. The method comprises establishing a reasonable SiGe HBT device model and network; stimulating semiconductor device characteristics of the established SiGe HBT device model; performing key electrical parameter correction on the SiGe HBT device model; selecting a typical incidence position on the surface of the device model to perform single event effect physical model stimulation; analyzing the relation between current and charge collection of all poles and time at different positions and funnel potential change conditions of different positions to obtain the sensitive position of the SiGe HBT towards the single event effect; selecting denser incidence points nearby the single event effect sensitive position, performing stimulation on single event effect semiconductor device numerical values, and locating HBT single event effect sensitive areas and amplitudes accurately. The method has the advantages that damage of the single event effect is analyzed theoretically and quantitatively, assessment time is shortened, test costs are reduced and the like.

An automated build-by-layer rainbow pallet system comprises a layer build track and a layer build cart that carries a layer build robot and has a number of positions for building rainbow pallets, for storing emptied supply pallets, and for storing empty half-pallets. On the rail-sides are dozens or hundreds of supply pallet conveyors for pallets of products to be picked-from when building the rainbow pallets. These are placed on one or both sides of the layer build track. The layer build robot can position itself in front of the appropriate supply pallet conveyor for picking. Once in position, half or full layers can be picked from the supply pallets and positioned over the target rainbow pallet to be built. The process of picking and building layers as the layer build cart repositions in front of the appropriate supply pallet conveyor is repeated until all of the on-board rainbow pallets have been built. The layer build cart returns to an unloading area where the emptied supply pallets and built rainbow pallets are conveyed off to respective exit conveyors. Several empty pallets are then moved from an empty pallet supply, and placed in the individual rainbow pallet build locations either by robot or conveying. The system is then ready to build more rainbow pallets. The system is controlled by a software program for orchestrating the overall movement and building of pallets.