75results about How to "Solve extraction difficulties" patented technology

The invention discloses a judicial case knowledge graph construction method of a dependency syntactic analysis relation extraction model. The method comprises the following steps of firstly, converting an unstructured judgment document into the structured data through an information extraction technology, then performing word segmentation, part-of-speech tagging and named entity identification processing on the structured data, and then extracting an entity relationship triple through a dependency syntax analysis relationship extraction model; and finally, importing the data in the triple forminto a Neo4j graph database in batches, and constructing a judgment document knowledge graph by utilizing the Neo4j and performing visual display on the judgment document knowledge graph. The dependency syntactic analysis relation extraction model can effectively extract the relation between the entities, is suitable for different large-scale corpora, and has the better transplantation applicability. The judgment document knowledge graph is intuitive and clear, so that the user can conveniently and efficiently master the information, and the great convenience is provided for the judicial work.

The invention provides a vibration acceleration signal frequency spectrum visual analysis method. Firstly the frequency spectrum data of vibration acceleration signals are obtained by performing Fourier transform; then the vibration acceleration signal frequency spectrum is segmented according to the high and low energy area boundaries, and statistics of the sum of amplitude of frequency components in each branch frequency band is performed; and finally a spectrum array river changing along with time is established with the sum of the amplitude of the frequency band acting as the characteristic index so as to realize visual display. Characteristic information extraction of the complex vibration acceleration signals can be realized so that the full life cycle operation state of mechanical equipment can be simply and visually reflected.

The invention relates to a novel chemical synthesis method of bromphenol PTP1B inhibitor which has a chemical structural formula as follows: wherein a group connected with 2,2' and the number 3, 3' carbon of benzene ring is bromine atom; 4,4' and the number 5, 5' carbon of the benzene ring are connected with hydroxy; and the PTP1B inhibitor has a chemical name in English: Bis-(2,3-dibromo-4,5-dimethoxy-phenyl)-methane. The compound negatively regulating an insulin signal transduction path by restricting the activity of protein tyrosine phospholipase 1B and has favorable curing effect for type 2 diabetes in insulin resistance types.

The invention relates to a citizen mediation case classification system and method based on feature migration and adaptive learning. The system comprises a data acquisition module, a feature extraction module, a feature migration module and a network training module, and is simple in system structure and wide in application range. The method comprises the steps of constructing a character vector table, carrying out the auxiliary data vectorization processing and the people's mediation data vectorization processing, inputting the vectorized auxiliary data into the neural network, extracting theauxiliary data features, migrating the extracted auxiliary data general features into the new neural network, inputting the vectorized people's mediation data into the neural network, and training aclassification model. According to the method, all texts can be effectively converted, the low-frequency words cannot be ignored, the dimension reduction is obvious, the training speed is high, and the follow-up online iterative optimization is facilitated, at the same time, the difference between the people's mediation field and the auxiliary field is solved, and the individual requirements of the specific field are met.

The invention provides a planetary gear box early fault diagnosis method based on APEWT and IMOMEDA. The method comprises the following steps: firstly, calculating a discrete scale space spectrum of aplanetary gear box fault vibration signal, and adaptively determining a frequency band division boundary on the scale space spectrum; then, APEWT is adopted to automatically decompose the fault signal into a series of modal components, and sensitive components are selected from the modal components; secondly, carrying out deconvolution processing on the sensitive component by adopting IMOMEDA, and carrying out waveform extension on the deconvolved component to obtain a lifting signal; and finally, extracting a fault characteristic frequency from the envelope spectrum of the lifting signal, and comparing the fault characteristic frequency with a theoretical fault characteristic frequency value of the gearbox so as to identify the type of a fault. According to the method, the early fault feature information of the planetary gear box can be extracted clearly and accurately, the problem that parameters in EWT cannot be determined adaptively is solved, the problem that MOMEDA has a seriousedge effect is also solved, and the fault diagnosis accuracy is improved.

The invention discloses a communication radiation source individual identification method and system. According to the technical scheme, the method comprises the steps of: receiving communication radiation source signal, carrying out radio frequency preselection amplification, carrying out frequency mixing, carrying out intermediate frequency filtering, carrying out A/D conversion, carrying out digital orthogonal demodulation, carrying out rectangular integration double-spectrum transformation, dividing rectangular integration double-spectrum characteristic vectors into a training sample set and a testing sample set, constructing a rectangular integration double-spectrum characteristic dictionary, carrying out non-linear transformation, carrying out mapping to a cooperation projection matrix, constructing a classifier, obtaining classification residual errors, and using the type corresponding to the smallest classification residual error as type of a communication radiation source individual. According to the invention, a small sample problem in a communication radiation source individual identification process is solved, the time complexity of the algorithm is lowered, and the phase and amplitude information distortion problem in the process when an existing method based on the time domain, frequency domain or time frequency domain is used to process non-stable or non-Gaussian signals is adopted.

The invention discloses an aircraft skin opposite joint gap calculation method based on end face extraction. The method specifically comprises the steps of acquiring aircraft skin actual measurement point cloud data; carrying out rasterization processing on the actual measurement point cloud data; calculating the planeness of the point cloud data in each grid through principal component analysis,extracting a plane grid in each grid, and fitting the point cloud data in the plane grids into a plane by adopting a weighted least square algorithm; extracting point cloud data of a skin plane area by adopting an area growth algorithm based on the fitted grid plane; and fitting a plane by adopting an iterative weighted least square algorithm, judging whether the fitted plane is an end face or not, and calculating a butt joint gap of the end face of the skin. According to the invention, the aircraft skin gap alignment gap can be accurately calculated, and the problem that the aircraft skin gapalignment gap is difficult to detect is solved.

The invention provides a controlling method and system of an encrypted delivery box. The controlling method comprises the following steps: a mobile terminal acquires the encryption identification information of a current delivery box and reports the encryption identification information to a delivery-box control server; the delivery-box control server extracts the information of the delivery box of a client according to the local identification information of the terminal and sends an open command to the delivery box; a delivery-box open command is generated according to the received open command and the verification information of the client, and the delivery box is controlled to open according to the command. The controlling method solves the problems of poor timeliness and low efficiency in physical distribution in the prior art, so that deliveries to be received can be correctly confirmed, and the delivery difficulties caused by reasons that account numbers are incorrectly input or are difficult to remember are solved; in addition, through a manner of dynamic encryption, the controlling method disclosed by the invention effectively avoids the leakage of the information of users, so that the accuracy and the validity of delivery are improved.

The invention provides a multi-information fusion based automatic collection method for batched numerical control workpiece machining progress. A machine tool numerical control system is connected with a computer terminal through a network cable, and machining start-stop information is read and processed through an interface program; in addition, a power sensor is introduced in between a machine tool motor and a frequency converter, the power sensor is connected with the computer terminal through an analog/digital conversion module to acquire real-time power information during machining; the power sensor acquires machining quality information during machining by communicating with a quality inspection system; progress data in the process of machining are acquired according to the machining start-stop information, the real-time power information and the machining quality information and by means of comprehensively processing the information. The method has the advantages of timely feedback, statistical accuracy, rich statistical content and the like, support is provided for improvement of management and control ability of enterprises on production fields and optimized operation of machining workshops, convenience is brought to customers to monitor executive conditions of orders anytime, and service capability for customers is improved.

The present application provides a text information rule analysis method for a search engine, which comprises the following steps: acquiring text of a natural language original document; carrying outfeature extraction on the text of the natural language original document to generate a text feature vector; using a pre-trained vector matching model to match the text with a sample in a sample library according to the text feature vector to obtain a target sample; determining a semantic distribution rule mode of the text by using a pre-trained semantic distribution rule mode according to the consistency of the text features between the original document of the sample of the target sample and the corresponding index set of the target sample, and determining a semantic distribution rule mode ofthe text; According to the semantic distribution pattern of the text, the text of the natural language original document is converted into an index set. A system for analyzing textual information patterns is also provided. The present application realizes the discovery of the word distribution rule of the bearing core semantics of the original document of the natural language, and realizes the search engine index term extraction with high accuracy.

The invention provides a control method of a high-reliability express box. The control method comprises the following steps of: a mobile terminal and an express box report respectively to an express box control server; the express box control server extracts the client express box information according to terminal local identification information, sends opening instruction and opening key to the express box and the mobile terminal, respectively; the mobile terminal matches the received opening key with the opening key of the express box, andgenerates an express box opening instruction; the express box controls the opening according to the express box opening instruction. By utilizing the control method and system, the problems of poor express accuracy and low efficiency in the prior logistics art are solved. The to-be-received expresses can be confirmed accurately, the pickup difficulties caused by wrong account or difficult-to-remember account are avoided, meanwhile, by the double information identification of the mobile terminal and the express box, the leakage of user information is avoided, and the veracity and effectiveness of express pickup are improved.

Embodiments of the present invention relate to a wood identification method and system based on depth learning of structural feature images. The method comprises the following steps: acquiring wood cross-sectional structure image data; Dividing the image data into a plurality of image blocks of uniform size; Establishing a training set and a test set corresponding to the image data according to aplurality of the image blocks; Construction of Multi-Layer Convolution Neural Network for Wood Image Identification; The training set is used for deep learning of the wood image identification multi-layer convolution neural network; The depth learning model is tested by the test set, and the model parameters are optimized according to the test results, and a wood image identification depth learning algorithm model is generated. A wood image data to be discriminated is identified according to the image recognition depth learning algorithm model. Thus, the accurate and fast identification of thewood species to be identified can be realized.

The invention discloses a wind profile radar signal and data processing method based on two channels. The wind profile radar signal and data processing method is implemented by a signal processor and a data terminal, and comprises the specific steps that: the signal processor sets two groups of processing parameters; 2, the signal processor generates two groups of power spectrum data; 3, the data terminal processes the two groups of power spectrum data respectively; and 4, the data terminal selects a processing result. Therefore, wind profile radar signal and data processing based on the two channels is implemented. The method has the advantages that: the data output rate of a wind profile radar system is not changed; the new increased processing load is relatively small; the method is easily implemented; and the extraction capacity and the estimation precision of a wind profile radar to weak signals are improved effectively.

The invention discloses a method for rapidly detecting virus of a potato tuber. The method is characterized by comprising the following steps of: extracting the RNA of the potato tuber, performing reverse transcription-polymerase chain reaction (RT-PCR), detecting PCR products through a nucleic acid colloidal gold chromatography test strip, and the like. The method suitable for rapidly detecting the virus of the potato tuber is established by combining the RT-PCR technology with the nucleic acid colloidal gold chromatography technology on the premise of optimizing a method for extracting the RNA of the virus of the potato tuber. The method has the greatest characteristic that: an RNA extracting solution is developed, the RT-PCR detection can be directly performed after a potato sample to be detected is ground, an RNA extraction step in the conventional method is eliminated, and the problem that the RNA is difficult to extract due to high starch content of the potato tuber is solved; and moreover, the developed nucleic acid colloidal gold chromatography test strip can finish the detection of the PCR products within ten minutes, electrophoresis and the dyeing of ethidium bromide are not required, and the method is rapid, practical, safe and pollution-free.

The invention relates to a device and a method for extracting steel rail impressed characters under a working condition. The device comprises a base, a guide mechanism and a carrier roller, the guidemechanism and the carrier roller are mounted on the base, a steel rail is supported and conveyed through the guide mechanism and the carrier roller, a support is further arranged on the base, and a line structure sensor is arranged on the support and perpendicularly right faces the surface of the rail web of the steel rail. The method comprises the following steps: acquiring steel rail surface point cloud data by using a linear structured light sensor; and fitting the point cloud data of the arc section where the steel rail impressing character is located through a least square circle fittingmethod based on radius constraint, screening out point cloud data corresponding to the steel rail impressing character, and converting the point cloud data into a two-dimensional image so as to extract a steel rail impressing character image. The method has a series of advantages of being simple in algorithm, good in extraction effect, high in reliability and the like, and can effectively solve the problem that steel rail impressed characters are difficult to extract under the working condition.

The invention discloses a simple and convenient exosome extraction method. The method comprises the following steps of (1) collecting exosome cell supernatant; (2) removing cell debris in the supernatant through centrifugation; (3) transferring the supernatant without the cell debris into an ultrafiltration tube, and enriching an exosome by using the ultrafiltration tube; (4) washing the exosome enriched in the ultrafiltration tube by using PBS; (5) salting out the washed exosome to obtain a relatively purified exosome; and (6) measuring the concentration of the purified exosome. According tothe method, the exosome is intercepted by using a filter membrane, and then high-molecular-weight protein is separated out through salting-out, so that the problem that the exosome is difficult to extract is solved; and the method is high in operability, and the problems that an exosome extraction reagent is expensive, operation is time-consuming, requirements for extraction equipment are high, and the product purity is low are truly solved. Therefore, by using the extraction method provided by the invention, the extraction method of the exosome is greatly simplified, and the extraction quality and extraction efficiency of the exosome are improved.

The invention relates to a coal gangue recognition and separation method based on machine vision and capacitance fusion and a device thereof. The device comprises a retaining rack arranged on a belt conveyor rack for mounting an accessory, wherein a detection unit, a conveying unit, an electric unit and a pneumatic unit are connected to the rack through the retaining rack and are arranged around the belt conveyor reasonably. The method comprises the steps of acquiring an effective thickness of a medium and center position information of a contour thereof through a CCD camera, and combining partial voltage numerical values at two ends of a resistor to recognize coal gangue; and operating a nozzle accurately to separate the coal gangue according to the position information of the center of the contour in combination with the operating speed of a conveyor belt. The method solves the problems that it is hard to extract features of the coal gangue, the recognition precision is low, the mistake separating probability is large and the like. The method is high in processing speed and high in timeliness, and can achieve very high recognition precision with few recognition features.

The embodiment of the invention provides a malicious application program detection method and equipment based on an AI model. The method comprises the steps of analyzing a target file in a to-be-detected application program installation package, extracting static information in the target file, where the static information comprises at least one kind of dimension information in a behavior dimension, an authority dimension and a content dimension; processing the static information into a digital feature vector in a feature transformation mode, wherein the digital feature vector is composed of 0 and 1; inputting the digital feature vector into a trained AI model to obtain a maliciousness detection result of the to-be-detected application program; training the AI model according to the input digital feature vector, and outputting the probability that the application program corresponding to the digital feature vector is a malicious application and/or the probability that the application program corresponding to the digital feature vector is a non-malicious application. The problems of difficulty in rule extraction, low coverage, poor expansibility, easiness in bypassing and the like during traditional malicious application detection are solved, and higher accuracy and timeliness are achieved.

The invention relates to a Matlab-based RTS noise detection analysis system for an image sensor. The system is characterized by using Matlab to read collected images and converting the images as pixel signals of a pixel gray value; performing white noise filtering to each pixel signal to obtain the pixel signals; using a gradient detecting method to obtain gradient change signals of the pixel signals; obtaining a pixel gradient signaling noise threshold via a threshold algorithm, and then denoising the gradient signals; counting number of gradient pulse of each pixel and outputting as an RTS noise detection result; averaging the signals between each two pulses of a pixel gradient signal to obtain an average dark current base of the pixel RTS noise; obtaining a reconstructed RTS noise signal by combining the average dark current base of the RTS noise and a gradient pulse signal. The system of the invention can be applied to any visible image sensors and can automatically realize RTS noise detection and whole parametric analysis. The system of the invention solves the problem that extraction and analysis of current RTS noise detection are hard, and more deeply studies all kinds of technologies of the image sensor and defect generated by radiation.