162results about How to "Fast and accurate identification" patented technology

The invention discloses a social contact interactive method, device, terminal and system. The method comprises the steps that second social contact application users and display equipment at the periphery of a first social contact application user are identified according to a service set identifier (SSID) of a wireless fidelity (WIFI) network; if the second social contact application users at the periphery of the first social contact application user are identified, social contact data interaction connections between the first social contact application user and the second social contact application users, the display equipment and the like are established, and image-text data of the social contact interaction can be displayed on the display equipment. According to the social contact interactive method, device, terminal and system, the surrounding social contact application users are identified through the SSID of the WIFI, based on the fact that the WIFI has the characteristic that signals cover in a short distance, the surrounding users can be identified quickly and accurately, and the social contact application users within a small range at the periphery are found, identified and connected accurately and quickly so that the image, text, voice and video information interaction can be conducted, and based on small screens of intelligent terminals and large screens of public display screens, the image, text, video and voice information interaction which is commonly used by the multiple screens can be achieved.

The invention discloses a quick detecting method of the edible vegetable oil quality through using an attenuated total reflection Fourier infrared spectroscopy, and the problems that the current detecting method can not verify whether the edible vegetable oil is repeatedly heated at high temperature or whether swill-cooked dirty oil is mixed in fresh edible vegetable oil can be solved. The method comprises the steps of: indicating the change of oil ingredients by fully utilizing the changes of the peak shape and the peak height of an infrared second derivative spectrometry; measuring the content of trans-fatty acids in the edible vegetable oil through absorbance of a characteristic absorption peak at 966.50cm<-1>; judging whether the edible vegetable oil is heated through the change of the characteristic peak at 988cm<-1>, and calculating the heating time; and determining whether the edible vegetable oil is swill-cooked dirty oil through the change of the characteristic absorption peak at 988cm<-1>. The quick detecting method of the edible vegetable oil quality is simple in operation, low in cost, free of pollution, high in detecting precision, and quick and accurate and can be capable of judging an edible vegetable oil sample to be detected which is repeatedly heated at the high temperature, the swill-cooked dirty oil and the swill-cooked dirty oil mixed in the fresh edible vegetable oil.

The invention discloses a communication machine room protection monitoring management system based on big data. The system comprises an environmental noise detection module, a sound comparison and separation module, a waveform recognition and analysis module, an environmental parameter storage database, a fault frequency detection module, an equipment association construction module, an equipmentoperation parameter detection module, a data adjustment and analysis module, a big data server and a prompt early warning terminal. According to the invention, the mixed noise in a machine room is analyzed; whether each device has a fault and the type of the fault are determined; the failure frequency of the device, the repair waiting time and the temperature change condition of the device are analyzed; the operation life evaluation coefficient corresponding to each device under the influence of the device and other devices is comprehensively analyzed by combining association tightness coefficients, fault repair time and the like of other devices and a device to be detected; the characteristic of high detection accuracy is achieved, protection early warning reminding is achieved, and protection monitoring management can be conveniently carried out on the devices in the machine room.

The invention discloses a multi-language text automatic broadcast method and system. According to the method, language identification is performed on a multi-language text to be broadcasted, and text fields corresponding to multiple languages are respectively acquired; the start and end points of each text field are respectively identified so as to obtain language identifiers; and the corresponding sound broadcast type of each language is correspondingly called according to the language identifiers, and all the text fields are broadcasted in turn. The system is provided with a character identification module and a voice broadcast module. The method and the system are intelligent and flexible so that rapid and accurate automatic identification and broadcasting of the multi-language text can be realized, and the situation of omitting language information in identifying the multi-language text can be avoided.

The invention, which belongs to the technical field of machine learning and deep learning, discloses a traffic sign detection and recognition method based on a self-built neural network. According tothe method, on the basis of a shot image, a non-real area of interest of a traffic sign in the image is obtained by using color segmentation in a digital image theory; a real area of interest is obtained by using an SVM classifier; and then the real area of interest is inputted into a self-built convolutional neural network for identification and classification. Therefore, the usage state of the traffic sign can be identified and classified quickly and accurately; the real-time requirement of quick, reliable, and accurate identification is realized; and the time and effort are saved.

The present invention relates to the field of image identification and segmentation technology, especially to a drawing object identification and extraction method in a complex scene. The method comprises the following steps of: the step 1: constructing a detection and identification model, and employing an image set with object frames and object type marks to train the model; the step 2: collecting a scene image comprising user drawing content; the step 3: employing the trained detection and identification model to locate a drawing area in the image, and identifying a drawing object; and thestep 4: selecting the drawing area, and employing an image segmentation technology to extract a drawing object contour area. The drawing object identification and extraction method in a complex scenecan be applied in the field of children drawing teaching, can increase intelligent interactivity of the drawing process and can improve the drawing experience.

The embodiment of the present invention provides a method for positioning the hidden danger of a door system of a train based on a probability petri network. Firstly, faults, which occur and possibly occur, at each key position of the system are analyzed to determine the hidden danger; a probability Petri network (PPN) model is established according to observation events of the hidden danger of the system and a logical relationship so as to describe a time sequence and a causal relationship between the events; then according to system initial information and the established probability Petri network model, an initial identification M0, a related matrix C and a transition sequence U are generated; a state equation is utilized to analyze the state description changing process of the probability Petri network; and finally, the hidden danger occurrence probability of the system is combined and the established system probability Petri network model is adopted to perform inference analysis on a discrete event system. By analyzing a probability value of a system hidden danger place in a stable state, positioning on the system hidden danger can be realized.

The invention belongs to the field of artificial intelligence, the invention discloses an abnormal behavior recognition method, device and apparatus based on voice and image features. The method comprises: performing feature extraction on the image of the user by utilizing a human body feature extraction model obtained after learning and training to obtain a to-be-recognized image feature vector;performing feature extraction on the voice of the user to obtain a to-be-recognized voice feature vector; performing cross fusion on the to-be-recognized image feature vector and the to-be-recognizedvoice feature vector to obtain a to-be-recognized fusion feature vector; and processing the to-be-recognized fusion feature vector by using an abnormal behavior recognition model obtained by learningand training the convolutional neural network to judge whether the action of the user belongs to an abnormal behavior, and if so, proving that the user belongs to a dangerous person, starting a corresponding interception function to intercept the user. Therefore, the users with abnormal behaviors can be identified more quickly and accurately, the safety coefficient of an enterprise is effectivelyimproved, and effective guarantee is provided for the safety of the users of the enterprise.

The invention belongs to the image processing field, and specifically relates to a focusing method and apparatus in the mode of holding certificate in hands. The focusing method in the mode of holding certificate in hands comprises the steps: acquiring an image to be focused; determining a face region from the image to be focused; according to the face region, utilizing a certificate identification model to determine a certificate region in the image to be focused, wherein the certificate identification model is obtained through training of a training sample set and is used for identifying the character content in the image; and determining the face region and the certificate region as the focusing regions to carry out focusing. The focusing method can determine the face region and the certificate region in the image to be focused and focus the face region and the certificate region so as to improve the definition of the information on the certificate in the taken image.

The invention provides a remote sensing tobacco field identification method based on an improved vegetation index. The remote sensing tobacco field identification method comprises the following steps: acquiring a multi-temporal synthetic aperture microwave radar remote sensing image and a visible light near-infrared multispectral remote sensing image of tobaccos in a monitoring area in the whole growth period; acquiring spectral reflection and backscattering data of remote sensing images of crops in a monitoring area in each growth period of tobacco, analyzing spectral reflection and backscattering differences between the tobacco and other crops in the monitoring area, constructing a characteristic vegetation index SPVI, and generating an index time sequence curve; determining an identification characteristic threshold, and extracting a tobacco suitable planting area according to tobacco growth characteristics, monitoring area land utilization data, DEM data and remote sensing data; and in a tobacco suitable planting area, establishing a classification feature set according to a feature vegetation index SPVI time sequence curve of the sample by using spectral reflection and backscattering differences between land classes, and identifying a tobacco field in a monitoring area. The method solves the problem of remote sensing identification of the tobacco field in the cloudy and rainy area, and the tobacco field identification is more accurate and convenient.

The invention discloses an abnormal access behavior identification method and device, computer equipment and a storage medium. The method comprises the steps of obtaining log data; according to the log data, calculating all time intervals corresponding to the situation that the designated user accesses the same designated URL for two adjacent times, wherein the designated URL is any one of all URLs contained in the log data; according to the time interval, calling a preset calculation formula to calculate a variable coefficient related to the time interval; and determining whether the behaviorof accessing the specified URL by the specified user is an abnormal access behavior or not according to the variable coefficient and a preset variation threshold. According to the invention, the timeinterval data is calculated from the obtained log data; the time interval data is subjected to discrete analysis to obtain the variable coefficient, and finally whether the behavior of accessing thespecified URL by the specified user is the abnormal access behavior or not is intelligently and accurately determined according to the variable coefficient and the preset variation threshold, so thatthe intelligence of identifying the abnormal access behavior of regularly accessing the URL is effectively improved.

The invention discloses a crack image acquisition method, a crack image detection method, computer equipment and a readable storage medium, and belongs to the technical field of computer vision. The crack image acquisition method comprises the following steps: detecting target equipment in a sample image to generate an equipment image set, detecting target components in the equipment image set togenerate a component image set, and processing the component image set to generate a crack image set; training a detection model based on the crack image set; compared with the prior art, layer-by-layer detection of a sample image and segmentation of a detection result are carried out, and target detection and semantic segmentation technologies are combined to identify cracks, so that specific pixels of the cracks can be rapidly and accurately identified instead of only identifying an external rectangular area where the pixels are located, and thus the obtained crack image is more accurate. According to the crack detection method provided by the invention, the crack detection model is trained on the basis of the crack image acquired by the crack image acquisition method, and the crack canbe quickly and accurately identified.

The invention discloses an identification and classification method for phosphorite in different grades on the basis of Raman spectrums and PCA-HCA (Principal Component Analysis-Hierarchical Cluster Analysis) and belongs to the technical field of spectrum application. The method comprises the following steps: acquiring Raman spectrum original information of different samples of phosphorite of different grades by using a Raman spectrometer; preprocessing the acquired Raman spectrum original information, establishing an identification model according to the preprocessed Raman spectrum original information together with principal component analysis (PCA) and hierarchical cluster analysis (HCA), so as to classify the samples of phosphorite to be tested. The method is capable of simplifying pretreatment steps of samples, avoiding tedious chemical processes and shortening the detection time, has the advantages of rapidness, high accuracy, nondestructiveness and the like, has significant promotion functions on identification on phosphorite of different grades, provides a novel idea for rapid, real-time and on-line detection in wells, and has great significances in improving phosphorite exploitation efficiency, reducing exploitation cost and protecting the environment.

The invention discloses a molecular method for quickly identifying closely related species of juniperus. The molecular method comprises the following steps: firstly, extracting leaf total DNA (Deoxyribose Nucleic Acid) or seed endosperm DNA from dry leaves or seeds of juniperus species; secondly, performing PCR (Polymerase Chain Reaction) amplification and sequencing on DNA samples which are qualified by detection through a karyogene Chil primer; thirdly, purifying a PCR amplified product; fourthly, performing a sequencing reaction on the purified product; fifthly, further purifying the sequencing reaction product; sixthly, performing data collection on a purified product by using a genetic analyzer, or performing the sequencing and the data acquisition on the PCR product amplified by the karyogene Chil through the genetic analyzer; seventhly, opening an original sequence peak chart obtained by sequencing through software, reading basic group data information of individuals with high sequencing quality, comparing and correcting an original sequence and identifying the sequences of sabina convallium (JM), a sabina chinensis (JS), juniperus tibetica (JT) and karyogene (Chil) as the closely related species of the juniperus; eighthly, analyzing and comparing the difference degree among the sequences and identifying the species. The molecular method has the characteristics of quickness and accuracy for identification.

The invention relates to an automatic firefighting explosion-prevention system and method for a bus. The automatic firefighting explosion-prevention system is characterized by at least comprising nozzles, a firefighting bottle, a control box and an audible and visual alarm, wherein the firefighting bottle is communicated with a plurality of nozzles through pipelines; the firefighting bottle and the audible and visual alarm are electrically connected with the control box respectively; the control box is electrically connected with a plurality of detectors through internal CAN buses. Due to the adoption of the whole system, monitoring on fire hazard or explosion in a whole carriage can be achieved, a fire extinguishing agent can be sprayed into the whole carriage in all directions, automatic alarm and automatic firefighting bottle driving can be achieved, and the rapid fire fighting and explosion prevention properties of the bus can be effectively improved.

The invention discloses an information processing method, a system and an apparatus. The information processing method comprises the following steps of acquiring a video including lip movements of a subject; identifying the video so as to generate identification information corresponding to an expression meaning of the subject, wherein the meaning the expressed through the lip movements; displaying the identification information.

The invention discloses an aero-engine system identification modeling method based on a mixed sequence, and the method comprises the steps: firstly establishing a part-level model of an aero-engine, and converting a linear time-invariant modeling problem of a single working state into a system identification problem for solving least squares according to the input and output of each working statepoint and state quantity data; secondly, selecting an appropriate model class, an estimation method and a signal source to solve a least square problem, and establishing a linear time-invariant modelof a single working state; and finally, establishing a polynomial nonlinear system under the ground condition in a linear regression mode. According to the method, the problem of state space modelingof the aero-engine in the multivariable control field can be effectively solved.

The invention provides a vehicle user identity identification method and a vehicle management system, and belongs to the data processing field. The vehicle user identity identification method comprises steps that terminal equipment is used to acquire license plate information of a vehicle; vehicle user identity identification information is acquired; the identity of the vehicle user is determined according to the vehicle user identity identification information and the license plate information, and the vehicle user identity identification information and the license plate information are corresponding to the identity of the vehicle user, and therefore the identity of the vehicle user is identified accurately and quickly, and vehicle management convenience is improved.

The invention relates to a method and a device for identifying semantic difference between source code versions. In one embodiment of the invention, a method for identifying the semantic difference between the source code versions is provided, comprising: respectively obtaining first debugging information of a first source code version and second debugging information of a second source code version; comparing the first debugging information with the second debugging information, and determining whether a second function matched with a first function in the first source code version exists in the second source code version; and identifying the semantic difference based on the determined result. In one embodiment of the invention, a device for identifying the semantic difference between the source code versions is provided. By adopting the method and the device provided by the invention, the semantic difference between the source code versions can be identified quickly and accurately.

The invention provides a method for identifying mature rape honey. The method comprises the following steps that: to-be-detected honey is detected by adopting liquid chromatography-quadrupole rod-time-of-flight mass spectrometry; and excimer ion peaks with the mass number of 207.07 +/-0.01 are adopted as characteristic peaks, when it is judged that the to-be-detected honey is mature rape honey when the time-of-flight mass spectrometry contains the characteristic peaks. According to the method, it is determined that the excimer ion peaks with the mass number of 207.07 +/-0.01 can serve as the characteristic peaks for identifying the mature rape honey, only the mature rape honey contains the characteristic peaks, and non-mature rape honey does not contain the characteristic peaks. The methodfor identifying the mature rape honey is quick and accurate, has important significance for identifying the mature rape honey, and provides a guarantee for controlling the quality of the honey and maintaining the market order.

Belonging to the technical field of Chinese medicinal material identification, the invention relates to a PCR method for rapid identification of fritillaria cirrhosa authenticity. Specifically, the invention relates to primers for identification of fritillaria cirrhosa authenticity and application thereof in identification of fritillaria cirrhosa authenticity. The invention further relates to a PCR method for rapid and accurate identification of fritillaria cirrhosa authenticity with the primers.

The invention discloses a remote sensing biological aerosol pollution alarm method based on cloud fluorescence data analysis, and relates to the technical field of aerosol detection. The remote sensing biological aerosol pollution alarm method comprises the steps that a biological aerosol fluorescence spectrum database is established through multi-channel signals, feature information of various biological aerosols in the database is enriched, and the accuracy of biological aerosol recognition is improved; in the process of identifying and early warning the bioaerosol to be detected, a multi-wavelength fluorescence detection signal and a mie scattering signal are utilized to eliminate the influence of obstacle interference and the like of a corresponding wavelength channel on the fluorescence detection signal; the corrected fluorescence detection signal is dispersed into fluorescence spectrum signals with corresponding wavelengths by using a spectrograph, and the fluorescence spectrum signals are integrated and recorded to form spectrum data of the to-be-detected biological aerosol under laser with different wavelengths; and the to-be-detected biological aerosol is analyzed and identified by utilizing a cloud database technology through a spectrum identification algorithm in combination with multi-wavelength spectrum data, so as to realize more accurate and rapid biological aerosol identification.

The invention discloses a tea variety classification method based on fuzzy inter-cluster separation and clustering. The method comprises the following steps: S1, carrying out Fourier near infrared spectrum collection on a tea sample; S2, carrying out pretreatment on the near infrared spectrum of the tea leaf sample by using multivariate scattering correction; S3, realizing near infrared spectrum dimension reduction treatment by using principal component analysis; S4, realizing identification information extraction of the near infrared spectrum data by using linear discriminant analysis; And S5, carrying out tea variety classification by using fuzzy inter-cluster separation clustering. According to the method, the problem that the clustering effect is not ideal when a complex data structureis processed by using traditional fuzzy inter-cluster separation and clustering is solved. The method has the advantages of high detection speed, nondestructive detection, capability of processing complex spectral data, high tea variety classification accuracy and the like.

The invention discloses a system for identifying a vehicle driver through track association degree. The system comprises a time sequence track acquisition module, a data cleaning module and a track association degree calculation module. The time sequence track acquisition module is used for acquiring time sequence track data comprising the vehicle driving track and the behavior track of the driverand forming a track database. The data cleaning module is used for processing and integrating data including vehicle information and driver information and after classifying and matching the collected track data, performing track restoration by combining time and space dimensions, determining and dividing dimension intervals according to business requirements, and respectively forming a three-dimensional resonance image taking the vehicle track scene and the behavior track, time and space of the driver as three elements. And the track correlation degree calculation module is used for comparing the vehicle track with the behavior track of the driver according to the determined dimension interval, setting a resonance threshold value according to business requirements, setting a group exceeding the threshold value as a group with high track matching degree, and determining the current driver of the vehicle from the group.

The invention belongs to the technical field of anti-counterfeiting products, and particularly relates to a special anti-counterfeiting code hot stamping foil and a preparation method thereof. The anti-counterfeiting code hot stamping foil is composed of a gum layer, an aluminum layer, a printing layer, a holographic information layer, an imaging layer, a release layer and a PET layer which are sequentially arranged from top to bottom, and anti-counterfeiting information printed by the printing layer is composed of printed images and texts and special anti-counterfeiting codes converted into picture forms. Based on the special code anti-counterfeiting hot stamping foil, original data is converted into a picture form, and then typesetting design is carried out on the converted picture and an identification printed image, so that the printing quality of an invisible code picture can be guaranteed, the requirement of high-speed printing can be met, variable information tagging can be carried out one by one, and the problem of high-speed tagging of a large number of commodity identifications is solved; when the anti-counterfeiting code is damaged by 75%, the anti-counterfeiting code can still be read, the error correction rate is high, and the anti-counterfeiting code is not easy to copy.