39results about How to "Capture comprehensive" patented technology

Disclosed is a non-linear process fault detection method based on two-dimensional dynamic kernel principal component analysis, belonging to the fault detection technical field; the method comprises the following steps: the first step is to determine the sampling parameters, namely, to judge the execution process, determine the sampling parameters, select the data parameters which affect the fault, and then judge whether to carry through training or testing; the second step is training, that is, to collect the data of normal work, pick up the non-linear principal component of the training data through two-dimensional dynamic kernel principal component analysis, calculate the square prediction error of the training data and determine the control limit; the third step is testing, that is, to collect the online observation data, pick up the non-linear principal component of the online observation data through two-dimensional dynamic kernel principal component analysis, calculate the square prediction error of the real-time online observation data, compare the control limit of the square prediction error of the real-time online observation data with the control limit of the square prediction error of the training data, display and gives an alarm if the control limit of the former exceeds that of the latter. The invention can timely detect the fault in the production process and reduce losses in the industrial production process.

The invention discloses a flight path prediction method. The method comprises the steps: considering the meteorological environment of a flight, extracting meteorological features through a convolutional neural network, modeling flight path features through a recurrent neural network, and finally outputting a series of future flight paths, specifically including the steps of flight path data preprocessing, meteorological feature extraction, time sequence model definition, model compiling and training, and flight path prediction. Through the technical scheme of the invention, the long-term flight trajectory can be accurately predicted in an actual environment.

The invention discloses an electricity stealing prevention intelligent video evidence obtaining system. A first image data output end of an image collection and compression terminal of the system is connected with a first general asynchronous transmit-receive transmitter interface of a microprocessor in a metering box. A control signal communication end of the image collection and compression terminal is connected with a safe digital input and output clamping port of the microprocessor in the metering box. A second image data output end of the image collection and compression terminal is connected with an image data input end of a main station server. A total data communication end of the microprocessor is connected with a microprocessor total data input end of the main station server through a communication module. A communication end of a meter reading module is connected with a second general asynchronous transmit-receive transmitter interface of the microprocessor. A communication module of an infrared temperature detection module is connected with a digital-analog signal converter interface of the microprocessor. A communication end of an embedded safety control module is connected with a third general asynchronous transmit-receive transmitter interface of the microprocessor in the metering box. According to the invention, technical support is provided for fighting against electricity stealing according to law.

The invention relates to the technical field of the network security, and specifically relates to a system capable of quickly discovering and processing network fault, and a processing method thereof.The processing method comprises the following steps: continuously monitoring and analyzing running state and traffic data of target network equipment by using a server, and establishing a traffic database after acquiring the traffic data of the target network equipment, and then performing real-time analysis and judgment on the traffic database; shielding a fault source by adopting a corresponding processing rule once discovering the network abnormality, thereby realizing the quick recovery of the network fault. Through the processing method disclosed by the invention, the network fault caused by network illegal behavior can be quickly processed, and the technical problem of low operation and maintenance efficiency is solved.

The invention discloses an upper limb intelligent rehabilitation system based on virtual reality. The upper limb intelligent rehabilitation system comprises a human body posture estimation module, a data transmission module, a training mode selection module, a model management module, a scene management module, a motion synchronization mapping module, an interaction management module, a training management module and a motion evaluation module. Through more humanized, intelligent and low-cost interactive experience, a simpler and more comfortable exercise rehabilitation environment is providedfor the dyskinesia patient, so that the rehabilitation effect of the patient is improved, and the rehabilitation time of the patient is shortened; and meanwhile, according to the system, arm movementand two-hand fine movement can be effectively and accurately fused, so the dyskinesia patient can conduct rehabilitation training recovery more comprehensively and effectively.

The invention discloses a system for monitoring SMS spam based on the length of content and a realization method thereof. A monitoring rule definition module synchronously configures preset monitoring rules, a monitoring value monitored according to flow and a suspicious threshold value monitored according to the content length to an analysis module and a database management operation module; when the information flow value of a calling user is larger than the monitoring value, the analysis module sends all information of the calling user to the database management operation module; the insert database module inserts all information of the calling user to the database; the database judges the content length of the short message, and if the content length of the short message is longer than the suspicious threshold value, the calling user is a suspicious user. The system and the realization method of the invention greatly improve monitoring efficiency and accuracy and better protect the right of users.

The invention belongs to the technical field of electron emission imaging, and discloses a PET (Positron Emission Tomography) annular detection imaging system based on scintillation fibers. The systemcomprises a plurality of detector modules which are sequentially connected along a central axis, wherein each of the detector modules consists of a scintillation fiber unit and a photosensor unit; the photosensor units are symmetrically arranged at two ends of the integral scintillation fiber unit; a plurality of photosensors are arranged on the photosensor unit at each end; and each photosensoris respectively correspondingly coupled with a plurality of scintillation fibers in the scintillation fiber unit. According to the system, the error influence caused by energy attenuation on positioning is effectively reduced, the reaction positions of the scintillation optical fibers are easily judged, missed detection cannot be caused due to splicing of a plurality of the detector modules, and detection information is more accurate and comprehensive, so that the PET annular detection imaging system has the characteristics of high efficiency, high sensitivity, high imaging resolution and morecompact devices.

The invention discloses a preparing system of PMMA light-guide-stage transparent membranes. The preparing system comprises a washing device, a drying device, an extruding device, a polishing device, aconveying device, a relay type membrane surface dust removal device, a film pasting device, a first cutting device, a traction device, an edge cutting device, a second cutting device and a piece loading device which are sequentially arranged and further comprises negative-pressure chip sucking devices arranged on the discharging sides of the first cutting device, the edge cutting device and the second cutting device correspondingly. Squeezing rollers below a mouth die of the extruding device are a pair of flexible rollers. The relay type membrane surface dust removal device comprises an upperdust removal roller set and a lower dust removal roller set which are arranged on the two faces of the membranes. The negative-pressure chip sucking devices each comprise a plurality of negative-pressure chip sucking parts arranged on one face or two faces of a to-be-cut product. Ultra-thin and ultra-wide PMMA membranes can be produced, inner stress of prepared products is evenly distributed, andthus the products are not prone to being cracked. Meanwhile, the very small thickness tolerance (smaller than 1.5%) and extremely perfect optical quality can be obtained, and the defect that powderedscraps press the appearances of the products to be damaged is effectively overcome.

The invention discloses a fusion method and device for multi-source satellite aerosol optical thickness data. The method comprises the following steps: acquiring a data set of a stationary satellite in a first preset time period, and acquiring a data set of a polar orbit satellite in a second preset time period; preprocessing the aerosol optical thickness of the stationary satellite and the aerosol optical thickness of the polar orbit satellite in the acquired data set to obtain the aerosol optical thickness of the stationary satellite and the aerosol optical thickness of the polar orbit satellite in each geographic block in the same time period; and establishing a data fusion model, and performing data fusion on the preprocessed aerosol optical thickness of the stationary satellite and the aerosol optical thickness of the polar orbit satellite in each geographic block in the same time period by using the data fusion model. According to the method and the device, the polar orbit satellite and the stationary satellite are selected for data fusion, and the two satellites with different scanning orbits and different detector wavebands are used, so that the environmental information can be captured more comprehensively, and the problem of insufficient time and space resolution of single data is effectively solved.

The invention discloses a shaver with novel rotatable cutter nets. The shaver comprises a main body structure, a machine body assembly and a cutter head assembly, wherein the mounting foundation of the machine body assembly is a machine body housing and a gear box; the cutter head assembly is composed of a cutter head cover, a cutter body assembly and a cutter body supporting assembly; the cutterhead cover, the cutter body assembly and the cutter body supporting assembly are combined into a whole; and the cutter body assembly is attached to the cutter head cover through a fulcrum. According to the shaver, the inner ring cutter net and the outer ring cutter net can be arranged to rotate forwards or reversely, so that the beard can be more comprehensively captured, and then a better shavingeffect is achieved; and meanwhile, a plurality of beard inlet grooves are formed in the plurality of rings of cutter nets, so that the beard and the hair can be quickly removed.

The invention discloses a social network feature dynamic extraction method based on vector compression and reconstruction. The method comprises the steps of (1) generating a training set; (2) constructing a deep semi-supervised auto-encoder network; (3) constructing a generative adversarial network; (4) training a network; and (5) completing dynamic feature extraction of the social network in thegenerative adversarial network. According to the method, the deep semi-supervised auto-encoder network is constructed and trained, high-order social network structure information can be better captured, and based on a dynamic feature extraction method of the generative adversarial network, the method has relatively short processing time and relatively high space utilization rate when being used for processing a large-scale social network.

The invention discloses a facial expression capturing method, system and equipment, relates to the technical field of computer vision, and particularly relates to a method, system and equipment for acquiring facial action information in real time by utilizing a face detection technology, a key point detection technology and a facial action fitting technology and then displaying the facial action information through a deformation model. According to the facial expression capturing method, system and equipment, a rich deformation model is utilized, an advanced facial action fitting technology isadopted, and each subtle change of the face is captured to the maximum extent, and the operation efficiency is improved to the maximum extent.

The invention provides a multi-round dialogue generation method and system based on information enhancement, and relates to the technical field of multi-round dialogue generation. The reply is generated through a pre-constructed multi-round dialogue model, the multi-round dialogue model comprises a multi-dimensional word level encoder, a speech level encoder and a decoder, and the reply generation process comprises the steps that multi-round dialogue historical information is acquired and preprocessed; the preprocessed multi-round dialogue historical information passes through a multi-dimensional word level encoder based on a multi-information embedding mechanism, and sentence vectors are obtained; inputting the sentence vector into an utterance level encoder to obtain an utterance level encoding vector; and decoding the utterance-level coding vector through a decoder to generate a reply. According to the method, context information is comprehensively captured through a multi-information embedding mechanism, and more relevant context information is generated to generate replies.

The invention discloses a vibration signal and neural network-based TBM disc cutter wear identification system, which comprises a vibration signal acquisition and processing subsystem and a neural network identification model, and is characterized in that the vibration signal acquisition and processing subsystem comprises a sensor module, a data acquisition module and a data processing output module which are connected in sequence; the sensor module is installed on the hob and used for obtaining hob rock breaking vibration signals and converting the vibration signals into piezoelectric signals. The data acquisition module acquires a piezoelectric signal at a set sampling frequency to obtain a time domain signal, and the data processing output module calculates the time domain signal by adopting a fast Fourier method to obtain a frequency domain signal and outputs frequency domain data; and the neural network identification model identifies the wear state of the single-blade disc cutter according to the frequency domain data. The recognition system has good stability and reliability, has good feature extraction and recognition capability, can comprehensively capture the features of the vibration frequency domain data, and achieves a good accuracy level.

The invention discloses an unsupervised pedestrian re-identification method based on online hierarchical clustering. Comprising the steps of collecting pedestrian sample pictures and forming a training set; inputting the training set into a pedestrian re-identification model and outputting features of all pedestrian sample pictures, performing online hierarchical clustering on the features of all the pedestrian sample pictures by using an online hierarchical clustering method, then generating a pseudo tag corresponding to the training set, performing training and iterative training on the pedestrian re-identification model according to the pseudo tag, and obtaining a trained pedestrian re-identification model; selecting a to-be-detected pedestrian picture and extracting features of the to-be-detected pedestrian picture; and calculating the cosine similarity between the features of the to-be-detected pedestrian images and the features of the remaining to-be-detected pedestrian images in the to-be-detected pedestrian data set, and according to a preset similarity threshold, retaining all to-be-detected pedestrian images higher than the preset similarity threshold to realize pedestrian re-identification. According to the invention, an unsupervised pedestrian re-identification task can be completed, and the accuracy is high.

The invention discloses a multifunctional endoscope for stomach detection, and particularly relates to the field of stomach detection. The multifunctional endoscope comprises a connecting pipe, an outer ring is arranged at one end of the connecting pipe, a connecting ring is fixedly mounted at the other end of the connecting pipe, one end of the connecting ring is provided with a knob, and the outer surface of the knob is provided with a thread. According to the invention, the outer ring, a micro camera, an inner ring and a first connecting shaft are arranged, during use, the micro camera on the outer ring is turned on, the outer ring is in an unfolded state, the visual field of the micro camera is increased, therefore, the micro camera can capture the situation of the stomach more comprehensively, when the micro camera is in a working state, illumination at one end is turned on to work, after detection is completed, a first torsional spring on the connecting shaft can drive the outerring to be folded inwards, the micro camera on one side of the outer ring is buckled on the inner ring, the space is reduced, the endoscope can be conveniently taken out of the stomach, and meanwhilethe micro camera is protected to a certain degree.