89results about How to "Wide range of recognition" patented technology

The invention provides an intelligent recognition and early warning processing information platform, which is characterized in that a plurality of field video collecting devices collect video monitoring images of located environment scenes; a data center stores and backups the received video monitoring images; an intelligent recognition and video analysis system performs streaming media processing on the video monitoring images and carries out intelligent recognition and analysis; and an application platform sends an early warning event processing instruction to terminal equipment of corresponding management personnel for video images conforming to an early warning condition according to an early warning processing instruction so as to enable the corresponding management personnel to take early warning processing measures in time. The intelligent recognition and early warning processing information platform performs intelligent recognition on various video images and scenes such as human faces, the crowd, flames and objects, is wide in recognition range and high in recognition accuracy, timely notifies corresponding platforms and terminals to take measures, and can improve the early warning processing efficiency.

The invention discloses a method for identifying internet application based on dynamic deep packet detection and relates to a method for identifying the internet application. in the method, an access control list module (10) and a deep packet detection module (20) are comprised; a dynamic deep packet detection strategy generating module (30) is arranged; the access control list module (10), the deep packet detection module (20) and the dynamic deep packet detection strategy generating module (30) are executed sequentially; the dynamic deep packet detection strategy generating module (30) can obtain the characteristic of the internet application dynamically by characteristic statistics for the internet application; the characteristic generates a deep packet detection strategy; furthermore, the deep packet detection module (20) is used for identifying. The invention has wide identifying range, high accuracy, high processing efficiency, flexible operation, good expansibility and strong realizability and can be used for realizing not only network device function but also network analysis system function.

The invention relates to an integrated analysis device for leakage monitoring of water supply pipe network and a method of the analysis device. The device comprises a flow monitoring device with a flow recorder, a third-party communication transmission medium and a main station system; the main station system communicates with the flow monitoring device via the third-party communication transmission medium in transmission protocols; and the flow monitoring device also comprises a first communication module. The installation manner is simple, the integrated degree is high, and the caliber testing range is wide. The method of the analysis device comprises the steps that a water supply pipe network system is divided into blocks, zero pressure test is carried out, the water inlet flow and pressure of a water inlet end of the pipe network block area are monitored in real time and uploaded to the main station system, and the main station system receives the real-time pressure and flow data, and displays, stores, analyzes and reports the data. Early indication of leakage of the water supply pipe network can be discovered timely, the identification range is wide, leakage points are located rapidly and effectively, an analysis result is verified rapidly and effectively, basic cost of the water supply system is reduced, and comprehensive benefits including the economical, cost, resource, management and service aspects of the water supply system are improved.

The invention relates to a classification and identification method for a communication signal modulating mode based on ART2A-DWNN, belonging to the field of classification and identification of communication signal modulating modes and solving the problem that single neural network has long period and low accuracy for classifying and identifying communication signals. In the method, an ART2A-E algorithm based on an ART2A network is taken as a first layer of a combined neural network, and similar modes is roughly classified by selecting relatively smaller vigilance parameters; a DWNN is directly connected with the output layer with the corresponding type of the ART2A network, Morlet mother wavelet Phi(x) with higher resolution in frequency domain and time domain are adopted, learning is carried out by utilizing error back-propagation algorithm, a synaptic weight can be modified with a conjugate gradient method till output is within the error range, and the number of modes in each typeis reduced after rough ART2-E classification, so that the DWNN can quickly converge. The invention is used for classification and identification of communication signals.

The invention discloses a method for detecting the phthalic acid ester compound concentration based on optical fiber immunosense. The method includes the steps that firstly, an optical fiber immunosense detection system is established; secondly, a standard inhibition curve is established, wherein (2.1), a fluorescently-labeled antibody solution and a phthalic acid ester compound standard solution with the known concentration are mixed, data obtained after pre-reaction are input to the optical fiber immunosense detection system, and a response signal is obtained; (2.2), the concentration of phthalic acid ester compounds is changed, the step of (2.1) is repeated, and response signals under different concentrations are obtained; (2.3) the standard inhibition curve is drawn according to the concentration logarithm value and the inhibition ratio; (3.1), the step of (2.1) is repeated on the phthalic acid ester compounds to be detected, and a response signal and the inhibition ratio are obtained; (3.2) according to the standard inhibition curve, the concentration of the phthalic acid ester compounds to be detected can be obtained. The method can detect various kinds of PAEs at the same time and is high in sensitivity and precision and fast in detection.

The invention relates to a deep learning-based quick dynamic human body action extraction and identification method. At present, existing human body identification technology and application have the following deficiencies in multiple aspects: a human skeleton is a complicated structure body, and different human action habits correspond to different human action modes, so that human body identification is difficult generally. The method comprises the following steps of firstly describing whole information of size, color, edge, contour, shape and depth of a human body target; providing useful clues for action identification; extracting effective motion features from a video sequence; performing trace analysis by utilizing a motion trace of the target under the long-distance condition; and performing two-dimensional or three-dimensional modeling on four limbs and trunk of the target by utilizing information extracted from an image sequence under the short-distance condition. The method is used for extracting and identifying quick dynamic human body actions based on deep learning.

The invention discloses a monoclonal antibody capable of identifying a beta-receptor stimulant. The monoclonal antibody is secreted by a hybridoma cell line Sal, which is preserved in China Center for Type Culture Collection, and has a preservation number of CCTCC NO:C201187. The invention also discloses an enzyme-linked immunosorbent assay method and a kit for detecting the beta-receptor stimulant, and application thereof to detection of the beta-receptor stimulant. Compared with prior art, the monoclonal antibody prepared by the invention can identify 8 beta-receptor stimulant drugs, has a wide range of identification; and the monoclonal antibody, the ELISA kit and the ELISA method provided by the invention have high sensitivity, and can detect beta-receptor stimulant residues with lower concentration.

The invention discloses an application identification method and application identification equipment for user definition. The method comprises the following steps of: A, performing packet sniffing on application to be analyzed; B, extracting application signature characteristics from a data packet; C, compiling the application signature characteristics; and D, loading the compiled signature characteristics. The equipment comprises a packet sniffing module, a signature extraction module, a signature compiling module, a signature loading module and an application identification testing module, wherein the packet sniffing module is used for acquiring the data packet; the signature extraction module is used for extracting a signature in a data packet load; the signature compiling module is used for compiling the signature characteristics; the signature loading module is used for loading the compiled signature; and the application identification testing module is used for testing whether the application can be identified or not. The method and the equipment are wide in identification range and highly accurate and extensible, and can be implemented as a network equipment function.

The invention discloses a work-piece crack detection method based on an RFID (Radio Frequency Identification Device). The work-piece crack detection method comprises the following steps: installing a fragile RFID electronic tag at a crack position to be detected of a detected work-piece in advance, so that the fragile RFID electronic tag is respectively connected and fixed with two sides of the crack position to be detected on the detected work-piece; and reading the installed fragile RFID electronic tag at the crack position to be detected on the detected work-piece by adopting an RFID electronic tag reader in need of detecting the crack of the detected work-piece, if reading successfully, judging that the crack is not generated at the crack position to be detected on the detected work-piece, and if not, judging that the crack is generated at the crack position to be detected. According to the invention, a work-piece is unnecessary to take down from equipment; and the work-piece crack detection method has the advantages of being rapid in identification speed, wide in identification range, easy to operate and control, simple, practical, small in equipment volume, flexible in shape, convenient in data format updating, good in safety and wide in application range and is capable of realizing penetrability and non-barrier detection.

The invention relates to a transformer substation robot inspection system and an inspection method thereof. The transformer substation robot inspection system comprises a background control server andan inspection robot which establish network connection based on a 5G communication technology, the problems of communication speed and stability between the robot and the background server are solvedthrough the use of the fifth-generation communication technology, the performance of the background server can be brought into full play, and the effect of analyzing and processing the picture in real time is achieved. The inspection method comprises task scheduling, path planning, robot autonomous navigation, holder camera control, picture file real-time return and real-time recognition technologies, continuous inspection of the robot in the automatic inspection process can be achieved, the number of times of stopping is reduced, the inspection efficiency and the inspection range are improved under the same battery capacity; and meanwhile, under action of the high-performance industrial personal computer, the inspection method is short in recognition time, high in accuracy and wide in recognition range, the intelligence of the robot is brought into full play, the inspection efficiency is improved, and the user experience is improved.

The invention discloses a music score image recognition method and a playing control method, and the method comprises the steps: obtaining a to-be-processed music score: removing useless points on themusic score through a K-nearest neighbor algorithm, and carrying out the Gaussian-based processing of the music score. Edge detection of the Laplacian operator is carried out to obtain music score content; Obtaining the position of a music note by using an LSD algorithm, and judging the category of the music score according to the linear position and the relative distance; Carrying out classification through a music score classifier, and carrying out image segmentation on the classified music score to obtain an image of an independent note; The obtained image is divided into a surface layer and an inner layer, the surface layer identifies notes, the inner layer identifies continuous notes, the surface layer and the inner layer are respectively input into a convolutional neural network, the identified results are combined to obtain complete notes, and the notes are associated with corresponding motor control feeding, so that the purpose of autonomously identifying the music score and the musical instrument by the numerical control platform is achieved. Compared with other existing similar methods, the music score recognition method and device have the advantages of being wide in recognition range, high in recognition precision and robustness and the like.

The invention discloses a medium material recognition method for robot autonomous environment perception, and belongs to the field of material recognition. The medium material recognition method comprises: driving a metal electrode and a measured material to perform contact and separation a plurality of times; detecting an electrostatic sensing signal by an electrostatic measurement system to obtain the electrostatic signals of multi-class materials. Wherein the metal electrode senses charge change so as to produce the electrostatic sensing signal; extracting three feature quantities such as an electrostatic signal release factor, an effective negative peak number, and a positive and negative peak variance sum within a time window; carrying out recognition classification by using the combined feature vector and constructing a classification device by using a k nearest neighbor algorithm; and re-collecting the electrostatic signals of a certain type of the materials, and carrying out recognition classification on the signal by using the trained classification device so as to achieve the automatic recognition classification on the material by the robot. The invention relates to the medium material recognition method for robot autonomous environment perception, wherein the characteristics of no influence by the illumination condition, reliable principle and simple structure are provided due to no requirement of the external excitation measurement method.

The invention discloses an intelligent robot based on deep learning and with an improved image recognition technology. The intelligent robot comprises a moving seat, a lifting steering mechanism, a multi-directional alignment head and an image recognition system. A four-drive drive steering walking mechanism is arranged in the moving seat and is used for supporting moving of the moving seat and four-drive synchronous steering. The lifting steering mechanism comprises a servo motor embedded in the end face of the top of the moving seat. The tail end of the output end of the top of the servo motor is fixedly connected with an electric telescopic rod. The side face of the tail end of the top of the electric telescopic rod is fixedly connected with the multi-directional alignment head. The image recognition system is arranged in the multi-directional alignment head. The multi-directional alignment head comprises a supporting machine body, a multi-dimensional steering mechanism fixedly connected to the right end of the supporting machine body, a probe detection mechanism and a plane scanning mechanism, wherein the probe detection mechanism and the plane scanning mechanism are used for scanning after steering alignment.

The invention discloses a method and a device for confirming a search time scope in a search engine. The method comprises the following steps of receiving the information inputted by a user and initializing the search time scope; identifying the time information meeting the preset time format in the received information, wherein the preset time format comprises a first time format and a second time format; adopting the identified time information meeting the preset time format for adjusting the initialized search time scope, thereby acquiring a final search time scope. According to the embodiment of the invention, the user need not input time according to a fixed format, the user also need not select time from a time selecting frame and the user only need input a search term containing time information according to a normal time presentation mode, so that the experience of the user is increased.

The invention provides a human face mask wearing detection method based on ResNet and Canny. The method comprises the following steps: carrying out data preprocessing firstly, and using then a Canny edge detection algorithm for extracting an edge information vector, wherein the weight coefficient of the edge information vector is alpha, the weight coefficient of an image vector is 1-alpha, and theedge information vector and the image vector are spliced to serve as network input after being multiplied by corresponding weights respectively; then establishing a ResNet convolutional neural network, and determining a network structure; finally, inputting a training set image into the ResNet network to carry out feature learning, determining a loss function and a parameter updating mode, carrying out adjustment and continuous iterative training on each layer of network parameters by using error back propagation, and finally realizing convergence to obtain a network model; and testing the trained network model through the test set image. The method is wide in recognition range and high in adaptability; and for the condition of congestion of large-flow people, whether people wear masks ornot can be quickly recognized, and the labor cost and the time cost are saved.

The invention provides an FnCpf1 or FnCpf1 mutant mediated in-vitro DNA editing kit. Transcription of crRNA and digestion-ligation reactions are combined in the same reaction mixture, and assembly ofa carrier and a target fragment is realized by one step. Furthermore, an FnCpf1 mutant with wider PAM recognition range is obtained, PAM can be recognized as 60/64(PAM-NNN), YN, TAA, TAC, TGC and CAAwhich can be applied to the editing scheme are selected, and the recognition range of FnCpf1 is expanded four times. Due to the advantages of Cpf1, one digestive enzyme is just required in the methodand can replace various types of restriction endonuclease at present without sequence dependency, and meanwhile, the digestion and ligation reactions are performed by one step. A CRISPR in-vitro editing system with wide DNA target range is provided and can be taken as an effective one-step reaction seamless editing tool.

The invention discloses a camera based motion sensing recognition system which comprises a camera and a recognition module. The recognition module is used for recognizing types of motions captured by the camera. According to the technical scheme, the problem about how to realize effective fusion of the motion sensing system and android and IOS platforms can be solved effectively.

The invention discloses a traffic object identification method based on a 5G portable intelligent terminal and an MEC; according to the method, information interaction is carried out through mutual communication between the 5G portable intelligent terminal and a 5G MEC server; behavior state sensors in the 5G portable intelligent terminal are carried by pedestrians to acquire the state data of thepedestrians, and the data can be uploaded to the 5G MEC server for pedestrian state recognition through 5G communication, and the state identification result can be sent to the 5G MEC server throughthe local resources of the portable terminal. The 5G MEC server is used for carrying out state clustering analysis on individuals and groups in a traffic scene according to state information such as the position, speed, posture and the like of pedestrians, and different behavior state models of the pedestrian walking, running, riding bicycles, sitting riding, turning and sitting buses, for identifying different traffic objects in the traffic scene. The identification result can provide data with relatively low cost, high real-time performance and high accuracy for vehicle anti-collision earlywarning, real-time scheduling of traffic intersection and the like.

The invention discloses a face recognition method based on 3D imaging and a terminal. The face recognition method comprises the steps that a 3D model of an icon is acquired; three-dimensional featurepoints are acquired from the 3D model; a skin and hair area is recognized on the icon of a target image; target feature points corresponding to the three-dimensional feature points in position are acquired from the skin and hair area; the three-dimensional feature points and the target feature points are compared to acquire a similarity value of the icon in the target image; and the identity of the icon in the target image is judged according to the similarity value. Through the face recognition method and the terminal, recognition of the identity in the image is more accurate, more types of recognition images are available, the recognition range is wider, and more application environments are available for identity recognition.

The invention discloses a method and system for matching an express company name based on an express number, relates to the technical field of logistics information processing, and can quickly and accurately match the express company name to which the target express number belongs by pre-establishing a path model and then inputting the collected target express number into the path model. The method comprises the following steps: S1, based on stored express bill number data, establishing a path model of an express bill number, the path model comprising a plurality of path data; S2, obtaining atarget express bill number inquired this time, and importing the target express bill number into the path model; and S3, comparing the target express bill number with path data in the path model, judging whether the path model comprises path data matched with the target express bill number or not, if so, outputting an express company name corresponding to the path data, and if not, updating the stored express bill number data. The system comprises the method provided by the technical scheme.

The invention discloses a salmonella bacteriophage nano magnetic bead conjugate and its enrichment and separation kit. The conjugate is a complex of a salmonella bacteriophage coupled with a carboxylmagnetic bead, wherein the salmonella bacteriophage is Salmonella Typhimurium bacteriophage LPST10, and a preservation number is CCTCC NO: M 2016473. The bacteriophage of the Salmonella bacteriophagenano magnetic bead conjugate of the invention can specifically recognize and infect bacteria, and has the characteristics of good specificity and small molecular weight. The kit is capable of rapidlyand sensitively separating and enriching salmonella.

The invention relates to a technical scheme of two-dimensional code risk identification. In the scheme, a set of fraudulent two-dimensional code recognition algorithm is established from multiple dimensions, so that the fraudulent or malicious two-dimensional code can be accurately found in the two-dimensional code provided by the user. According to the scheme, a URL corresponding to a two-dimensional code is identified to identify a malicious URL. The scheme also comprises a derivation risk identification step which comprises deriving and discovering more fraudulent persons and fraudulent persons based on the discovered fraudulent two-dimensional code so as to realize self-circulation two-dimensional code risk identification.

The invention relates to the technical field of artificial intelligence, in particular to an entity recognition method and device based on semantic analysis, equipment and a storage medium. The methodcomprises the steps of obtaining a to-be-identified corpus, matching the to-be-identified corpus in a suffix dictionary, and recording all suffix word indexes to obtain a suffix word index set; segmenting and extracting the corpus to be identified between two adjacent suffix word indexes in the suffix word index set to obtain a sub-character string text, matching the sub-character string text ina candidate dictionary, recording all candidate word indexes, and combining the target candidate words with the adjacent suffix words one by one to obtain a plurality of entity words. According to themethod, the initial boundary of the entity word is defined based on the prefix dictionary, the suffix dictionary and the candidate dictionary, the matching process is high in efficiency, low in searching difficulty, wide in recognition range and high in transportability, and the new word and the entity word containing the special symbol can be recognized to the maximum extent.