1116results about How to "Precise screening" patented technology

The invention provides a vehicle license plate recognition method based on a video. According to the vehicle license plate recognition method based on the video, moving vehicles are detected and separated out with the vehicle video which is obtained through actual photographing by means of a camera serving as input, the accurate position of a vehicle license plate area is determined by conducting vertical edge extraction on a target vehicle image obtained after pre-processing, a vehicle license plate image is separated out, color correction, binaryzation and inclination correction are conducted on a vehicle license plate image, each character in the positioned vehicle license plate area is separated to serve as an independent character, feature extraction is conducted one each character, obtained feature vectors are classified through a classifier which is well trained in advance, a classification result serves as a preliminary recognition result, secondary recognition is conducted on the stained vehicle license plate characters according to a template matching algorithm imitating the visual characteristics of human eyes, and then a final vehicle license plate recognition result is obtained. The vehicle license plate recognition method based on the video has the advantages that hardware cost is reduced, the management efficiency of an intelligent transportation system is improved, the anti-jamming performance and the robustness are high, the recognition efficiency is high, and the recognition speed is high.

The present invention relates to an advertisement recommendation method, comprising: receiving an advertisement recommendation request; acquiring a user tag that is modified according to feedback information and that corresponds to a user in the advertisement recommendation request, wherein the feedback information comprises positive feedback information and negative feedback information; screening out a candidate recommendation advertisement according to the modified user tag; calculating, according to an estimated click model, an estimated click possibility of each candidate recommendation advertisement for the user on a current display page; calculating, according to an estimated non-interest model, an estimated non-interest possibility of each candidate recommendation advertisement for the user on the current display page; and screening out a target recommendation advertisement from the candidate recommendation advertisement according to the estimated click possibility and the estimated non-interest possibility. In an advertisement recommendation process, both positive feedback information and negative feedback information are considered, and advertisement recommendation accuracy is improved. Further, an advertisement recommendation apparatus is further provided.

The invention discloses a method for screening traditional Chinese medicine effect related ingredients and a model building method. The method comprises the steps as follows: taking the researched traditional Chinese medicines as objects for preparing multiple sets of sample solutions containing chemical components with different concentrations; selecting part of the chemical components for HPLC (High Performance Liquid Chromatography) analysis, and investigating linear relations between the peak areas and the concentrations; selecting the multiple sets of sample solutions in the linear range, and determining the respective HPLC fingerprint spectrums and the medicine effect indexes of the sample solutions; and obtaining relations between the peak areas and the medicine effect indexes of the ingredients by adopting a linear regression analysis method in the linear relation range, and confirming the actions of the medicine effect related ingredients on the medicine effect indexes according to the positive value and the negative value of the standard regression coefficients and / or the t check results of the ingredients so as to screen out the traditional Chinese medicine effect related ingredients. The method is simple, convenient, easy to implement, reliable and good in commonality, can embody the characteristics of comprehensive actions of the multiple ingredients on the medicine effect indexes, is suitable for screening the traditional Chinese medicine effect related ingredients, and provides a technical data support for discovering and screening traditional Chinese medicine active ingredients.

The invention aims at providing a method and device for excavating similar judgement documents based on big data. The method for excavating similar judgement documents based on big data comprises obtaining a large number of judgement documents which are disclosed, and obtaining the causes of action of each judgement document; obtaining the text topic feature information with respect to the case of the judgement document and several keyword relevant messages with respect to the party disputing content and the party appealing content in the judgement document based on the text content of each judgement document, and establishing a text feature vector of the judgement document based on the text topic feature information and the several keyword relevant messages; updating a feature word library with respect to keywords based on the several keyword relevant messages, displaying each judgement document in the large number of judgement documents in a text feature vector manner effectively and accurately, updating the feature word library with respect to keywords, thereby obtaining similar judgement documents quickly, and improving the working efficiency of excavating similar judgement documents.

A fluorescent image obtainment apparatus includes a light illumination means (130, 150) that illuminates a region to be observed with illumination-light and excitation-light at the same time, and an imaging means (110) that captures an image composed of light reflected from the region illuminated with the illumination-light and an image composed of fluorescence output from the region illuminated with the excitation-light. Further, the apparatus includes an image processing means (170) that produces an ordinary image based on the image of the reflection light and a fluorescent image based on the image of fluorescence, and a light amount control means (186) that controls the light amount of the illumination-light so that a representative luminance value of the ordinary image becomes a predetermined luminance value and the light amount of the excitation-light so that the ratio of the light amount of the excitation-light to that of the illumination-light becomes a predetermined ratio.

The invention provides a depression risk screening system and method based on a virtual reality scene electroencephalogram signal. Under the stimulation of a virtual reality scene, real-time acquisition, processing and analysis of data are carried out on the electroencephalogram signal by an electroencephalogram acquisition system, and finally, a depression risk is screened by a data mining method. The system provided by the invention comprises a virtual reality induction system, a electroencephalogram signal acquisition system and a data analysis system; the virtual reality induction system is used for establishing different virtual reality scenes; the electroencephalogram signal acquisition system is used for acquiring the electroencephalogram signal of a human brain, which is generated under the stimulation of the virtual reality scene and is continuously changed, and outputting the processed electroencephalogram signal and extracted real-time feature information; the data analysis system has a classification discrimination model trained by labeled healthy populations and depression risk populations, compares feature information required in a depression screening process with feature parameters of the classification discrimination model, and distinguishes and discriminates the healthy populations and the depression risk populations.

The invention discloses a method for identifying unconventional cells in a pathological section, comprising: preprocessing an electronic scanning pathological section to obtain an effective discriminating region, inputting the effective discriminating region into a full convolution network to be pre-trained, then replacing the full convolution network head fine-tuning network with a fully connected layer so that the full convolution network has an ability to extract unconventional cell features so as to determine the locations of unconventional cells and classify the effective discriminatingregions effectively; in combination with the prediction result vote of multiple common classification networks, outputting stable classification results. The identification method of the invention canautomatically discriminate the probability of existence of unconventional cells in each 20* magnified field of view in the pathological section, and outputs the unconventional cells with a probability value over 0.5 as a recognition result, greatly reduce workload of the artificial screening pathological section, and can quickly and accurately screen out the unconventional cells.

The invention discloses a battery selecting method comprising steps of performing fully charging and discharging circles and measuring capacities of a plurality of batteries; calculating the average battery capacity value and calculating relative deviations; removing batteries that have large discharging capacity deviations; performing charging and discharging processes, and measuring voltages in the processes; calculating the average voltage value and calculating relative deviations; and selecting batteries with good consistency. According to the selecting method provided by the invention, by selecting firstly according to discharging capacities of batteries, and then by performing charging and discharging tests close to actual working conditions for the batteries and measuring voltages, batteries with good consistency are selected rapidly and accurately, thereby raising the accuracy of battery selecting, improving battery pack security, and prolonging battery service lifetime.

The invention relates to a method and device for detecting a defect of the surface of a steel rail based on data processing. The device comprises four sensors, a sensor bracket (5), a base (6), a railtransport roller (7), the steel rail (8), an encoder bracket (9) and an encoder (10); the four sensors are respectively arranged on four inner corners of the sensor bracket, the Z axes of all the sensors form a 45 degree included angle with the horizontal direction and are intersected at one point; the rail transport roller is fixed on the base and is used for supporting and transporting the steel rail. The method adopts the triangulation distance measurement principle of a linear laser, four line structured light laser sensors are used for collecting steel rail cross section profile data, and the steel rail profile data collected by all the sensors are subjected to coordinate conversion: rotation, translation, graphics splicing, and obtaining of the data splicing calibration value of allthe sensors. The method and devic can significantly improve the production efficiency, reduce the manual participation degree, and reduce the detection leakage rate and the misjudgment rate.

The invention provides an incoming call processing method and a device of mobile terminals and a mobile terminal. The method includes the following steps that the mobile terminal receives an incoming call request of a caller; the mobile terminal determines whether the mobile terminal is presently in a silent mode or not; if the mobile terminal is presently in the silent mode, an incoming call number or contact person information is searched in an incoming call times record which the silent mode corresponds to according to the incoming call number or the contact person information of the incoming call request; if the incoming call number or the contact person information is found out in the incoming call times record, the incoming call times of the incoming call request are added by one, and whether the incoming call times are larger than a preset threshold value or not is determined; and if the incoming call times are larger than the preset threshold value, the mobile terminal prompts a user, and otherwise, the user is not prompted. By means of the incoming call processing method and the device of the mobile terminals and the mobile terminal, an urgent incoming call can be automatically determined, the user can not be disturbed and can not miss important incoming calls and short messages, and the user experience is improved.

The invention relates to a method for sorting the consistency of lithium titanate single batteries. According to the method, both static performance parameters such as capacity, voltage and internal resistance and dynamic parameters such as polarization impedance, large-current charging-discharging capacity and self-discharging capacity are considered. The method comprises the following steps that1, primary sorting is carried out according to parameters such as discharge capacity, open-circuit voltage and alternating-current internal resistance; 2, secondary sorting is carried out after the single batteries undergo high-temperature standing for a period of time; and 3, third-time sorting is carried out on the single batteries subjected to secondary sorting by detecting high-rate pulse discharge conditions, and finally the lithium titanate single batteries with high consistency are obtained. According to the method, the problem that the consistency of single batteries in an existing lithium titanate battery system is difficult to sort is solved; the performance of the battery packs which are formed by the lithium titanate single batteries sorted through the method can be better than that of battery packs which are grouped through a common grouping method; and meanwhile, compared with an existing conventional sorting method, the sorting method is simple to operate, free of redundant equipment investment, short in testing time, and facilitates industrial popularization and application.

An embodiment of the invention provides a method and a device for determining a high-speed rail user and relates to the field of communication. By the aid of the method and the device, the high-speed rail user can be accurately determined, the service use quality of the high-speed rail user is analyzed through tests, the network test cost is reduced, and the mobile communication network quality is improved. The method comprises steps as follows: S1, whether the updated LAI of user equipment is identical with the LAI in a train location area table or not is judged, wherein the user equipment is an observed user; S2, if the updated LAI of the user equipment is determined to be different from the LAI in the train location area table and the LAI matching number of the user equipment is larger than or equal to a second preset threshold value, the user equipment is determined as the high-speed rail user; S3, if the updated LAI of the user equipment is determined to be identical with the LAI in the train location area table and the time difference is smaller than or equal to a first preset threshold value, the user equipment is determined as a suspected high-speed rail user, one is added to the LAI matching number of the user equipment, and S1-S3 are repeated. The method and the device are used for determining the high-speed rail user.

The invention provides a conference video shooting, recording, and propagating method, a conference video shooting, recording, and propagating device, and a server. Conference documents of a conference are acquired; the content of the conference documents is determined according to frequencies of sentences and words, or based on machine learning and accumulated business dictionaries, and the key words in the conference content can be discriminated; a time segment related to each key word is analyzed, and the video content and the audio content of the time segment are captured, and then the video content and the audio content are integrated automatically to form the theme conference video of each key word, which is stored in the conference documents; all of the conference documents are uploaded to a corporate communication platform. The conference can be recorded automatically, and the key conference content can be discovered.

The invention relates to a method for effectively screening self-discharging batteries. The method comprises the following steps: (1) determining low-voltage-state voltage knee points of batteries by using a trickle charge method; (2) carrying out charging and discharging on the batteries, allowing the voltage of the batteries to be finally located at the low-voltage-state voltage knee points, and after the voltage of the batteries tends to be stable, measuring and recording the voltage OCV1 of each battery; (3) storing the batteries at a set temperature for a period of time and measuring and recording the voltage OCV2 of each battery; and (4) calculating the voltage drop value delta V of each battery, wherein delta V satisfies the equation that delta V = OCV1 - OCV2, and if the voltage drop value delta V of a battery exceeds a set value, the battery is a self-discharging battery need to be screened out. The method provided by the invention is more scientific, accords with principles, can more accurately screen out self-discharging lithium ion batteries and reduces irreversible capacity loss of batteries in the phase of screening of self-discharge; thus, excellent electrochemical performance and safety performance of batteries in subsequent usage are guaranteed.

The invention provides a self-discharging screening method of a lithium ion battery. The method comprises the following steps: step 1: pre-charging a battery to be screened to be in a full or half state; performing high-temperature aging treatment and storing at constant temperature; marking detection open-circuit voltage as OCV1 and storing at constant temperature; marking detection open-circuit voltage as OCV2; calculating a voltage critical value U0 in the same phrase, wherein the battery with the condition that OCV2 is less than U0 is a serious self-discharging battery; and step 2: removing the self-discharging battery by using a 3o circulating way based on voltage pressure before / after the laying-aside of the battery removed in the step 1 in a full-charge state or a half-charge state. According to the self-discharging screening method of the lithium ion battery, the battery is screened after the battery is laid aside in the full-charge state or the half-charge state and the open-circuit voltage is determined, and the corresponding self-discharging batteries are removed by using the 3o circulating way, so that the self-discharging screening of the lithium ion battery is realized.

The invention discloses a feature vector-based fast and high-precision robustness matching method. Due to the changes of the environment, the influences of the target motion and the defects of sensors, the shot images are not only influenced by the noise but also have severe grayscale distortion and geometric distortion, therefore, how to achieve and realize high precision, high matching accuracy rate, fast speed, strong robustness and strong anti-interference performance becomes a goal pursued by the matching method. The invention discloses a sub-pixel-level fast and high-precision robustness matching, comprising the steps of: respectively extracting the SIFT (Scale Invariant Feature Transform) feature vectors of two pictures to be processed; carrying out PCA (Principal Components Analysis) dimension reduction treatment on the two pictures; matching the two feature vectors by utilizing Kd-tree; then screening the obtained matching points by utilizing an RANSAC algorithm; and causing the matching to achieve the sub-pixel level through a surface fitting technology so as to obtain the feature point pair of the high-precision robustness. Furthermore, the matching speed is increased by using an even point getting method. Through the invention, the obtained image matching result has extremely high matching precision and faster running speed and the experiment effect is excellent.

The invention discloses a cancer driving gene screening method based on gene mutation frequency and belongs to the field of cancer medicine. The method comprises the steps that (1) tumor gene mutation data is acquired; (2) the mutation data is preprocessed; (3) adjacent genes of each gene are screened; (4) background mutation data of each kind of mutation is calculated; and (5) hypothesis testing is used to screen mutation genes according to mutation values. According to the method, a modern high-throughput sequencing technology, DNA data processing software, a classic clustering algorithm and a classic statistical method are utilized and combined with biotic factors influencing gene mutation, so that screening of cancer driving genes is more precise, and the method has an important meaning for both research and development of novel anti-cancer drugs and clinical diagnosis and treatment of cancers.

An embodiment of the invention discloses a filtering method and a filtering device for point cloud data. The filtering method comprises the steps of: converting the point cloud data into raster data, and acquiring the lowest point in each raster as the initial point; performing morphological opening operation on the raster data to obtain an initial terrain surface; for each initial point, determining the current initial point as a potential ground seed point if a difference value between an elevation value of the current initial point and a value of the target raster which the current initial point falls into; constructing an initial triangulation network model according to the potential ground seed point; and determining a ground point according to the initial triangulation network model based on an irregular triangulation network filtering algorithm. By adopting the filtering method and the filtering device, the ground seed point can be reasonably determined, and the filtering efficiency is improved while guaranteeing the accuracy and precision of the initial triangulation network model, thereby the ground point in the point cloud data can be screened quickly and accurately.

The invention discloses a gene TaSnRK2.10 relevant to the wheat thousand seed weight, a molecular marker TaSnTK2.10-4A-caps relevant to the gene and application of the marker. The DNA of a wheat variety to be detected is subjected to PCR (polymerase chain reaction) multiplication through a marker primer TaSnTK2.10-4A-caps; a multiplication product is subjected to cleavage through SalI incision enzyme and is also subjected to electrophoretic separation; for example, PCR products are two bands with the sizes of 793bp and 316bp; the wheat variety is a variety with haplotype with high thousand grain weight of the gene; the PCR product is only a band with the size of 106bp; the wheat variety is a variety which is not provided with the haplotype with high thousand grain weight of the gene. According to the gene TaSnRK2.10 relevant to the wheat thousand seed weight and the molecular marker TaSnTK2.10-4A-caps thereof, the wheat variety or strain with high thousand grain weight can be conveniently detected or screened, and the selection process of the high-yield variety of wheat can be greatly accelerated.

The invention relates to a test case screening method and device, computer equipment and a storage medium. The method comprises the steps of when a to-be-detected code file is detected each time, recognizing a code modification label of the to-be-detected code file, and storing the code modification label in a history table; obtaining a plurality of test cases, and executing the plurality of testcases based on the to-be-tested code file each time; associating the test cases of which the test results are test failure with the corresponding code modification tags, and recording corresponding association relationships in a historical table; calculating the failure times of each test case associated with each code modification tag based on the historical table; and determining a target test case corresponding to each code modification label according to each code modification label and the failure times of each associated test case. By adopting the method, the pertinence of the test casescan be improved when a service system after code modification is tested.

The invention discloses loop-mediated isothermal amplification (LAMP) primers, a kit and a detection method for detecting common carbapenemase genes of gram negative bacilli. In the LAMP primers disclosed by the invention, klebsiella pneumoniae carbapenemase (KPC) and new delhi metallo-b-lactamase (NDM) primer groups can detect all subtypes except for NDM-10; hypoxanthine nucleotide (IMP) and vimentin (VIM) primer groups can detect common subtypes at home and abroad. The LAMP kit built by the invention is applied to joint detection of KPC, NDM, IMP and VIM genes, can cover the common carbapenemase genes of non-fermentative bacteria and enterobacteriaceae, can accurately and quickly screen the common carbapenemase genes, and has great clinical significance for timely detecting and further controlling fulminant epidemic caused by propagation of the carbapenemase genes in enterobacteriaceae. The kit disclosed by the invention is high in detection sensitivity and the minimum detection limits of the KPC, NDM, IMP and VIM genes can reach 100 CFU / reaction.

The invention relates to a photo processing method and a photo processing device. The method comprises the steps of grouping to-be-processed photos according to preset rules, detecting the quality of each group of to-be-processed photos to get the quality level of the group of to-be-processed photos, and storing the to-be-processed photos according to the quality level. By implementing the technical scheme, the storage space of a terminal or server is saved, and the storage pressure is reduced. Meanwhile, photos of high quality are screened out accurately for users, and the photos of high quality can be stored accurately and completely.

The invention belongs to the field of plasma diagnosis computer processing, in particular to a processing method and device for multi-channel energy spectrum measurement. The processing method includes the following steps: collecting a pulse signal sent by a probe and converting it into a digital pulse signal; counting and classifying the digitized pulse signal to obtain a count spectrum; converting the count spectrum into an energy spectrum; The energy attenuation produced by the beryllium foil is reduced, and the energy spectrum obtained in the third step is restored; the energy spectrum after reduction is smoothed and peaked and fitted to obtain the electron temperature. The processing device includes a pulse signal generation module, a high-speed acquisition card, an analog / digital signal conversion module, a high-speed cache module, and a processor unit with an energy counting module and a temperature calculation module. The problems of common multi-channel energy spectrum measurement and processing methods, such as high hardware cost, strong equipment specificity, limited number of energy channels, low spatial and temporal resolution, and inability to save original signals, are solved.

The embodiment of the invention provides a visual positioning method and system and a computer readable storage medium. The method comprises the steps: acquiring a to-be-positioned image, retrieving and obtaining at least one image in a pre-constructed three-dimensional visual map database to serve as a candidate image, wherein the global feature vector similarity of the candidate image and the to-be-positioned image is larger than a threshold value; respectively calculating respective relative pose information of the to-be-positioned image and the at least one candidate image, and obtaining at least one pose result of the to-be-positioned image relative to a three-dimensional visual map by utilizing the calculated relative pose information; and performing a verification test on at least one pose result, and outputting the pose result with the highest score in the verification test. According to the method, the system and the computer readable storage medium provided by the embodimentof the invention, visual accurate positioning can be realized in a complex dynamic scene and a shielded place.

The present invention discloses a wind turbine generator operation state identification method based on a DBSCAN. The method comprises the following steps: the step 1: collecting operation data of a wind turbine generator and basic parameters of the wind turbine generator; the step 2: performing data normalization processing of the wind turbine generator; the step 3: screening the shutdown data from the operation data according to the wind speed and the power data; the step 4: taking the wind speed, the power and the pitch angle data obtained through the normalization processing in the step 2 as the input data of the DBSCAN, and performing optimization of the model parameters of the DBSCAN clustering; and the step 5: obtaining the operation state identification result of the wind turbine generator according to the data classification results of the step 3 and the step 4.

The invention provides a detection method for illegal additives in weight-losing type Chinese patent medicines and health foods. A Q-Orbitrap high-resolution mass spectrometer is adopted and a Full MS Scan / dd-MS2 (Top N) scanning mode is carried out so that positive and negative ions are detected at the same time; the separation, high-precision first-grade mass spectrometric scanning and high-precision second-grade mass spectrometric scanning of analytes in a sample can be finished in an analyzing period to obtain accurate mass quantity and accurate fragment ion information; chromatography retention time, a first-grade mass spectrum and a second-grade mass spectrum of the sample are compared with those of a weight-losing type chemical component comparison product to determine that whether weight-losing type chemical medicines are doped into the sample or not; and rapid screening, confirmation and quantification can be carried out on a potential positive sample. The detection method has the advantages of high sensitivity, good selectivity, strong confirmation, simplicity and rapidness of operation and the like, can be used for accurately screening, has the sensitivity beyond comparison of traditional HPLC and low-resolution LC-MS, and can be used for identifying, confirming and quantitatively analyzing compounds in a single analysis period (12 minutes).

The invention discloses a three dimensional tissue and organ culture model, a high throughput automatic stereo image analyzing platform thereof, and a method utilizing the provided platform to screen antitumor drugs. The platform can be applied to the culture system of three dimensional tissue and organ / organ-like substance, three dimensional scanning and sampling system, and high-volume data storage and computer analysis system. The provided platform can simultaneously process a large amount of clinical samples and screen a plurality of antitumor drugs, is capable of greatly reducing the cost and maximally reducing the detection time, thus is widely applied to the drug selection schemes and dosage optimization in clinic, novel drug development, and basic researches on interactions among tissue, organ, biological macromolecules, and other micromolecules.

The invention discloses an epidemic prevention and control AI robot and a prevention and control method, and belongs to the field of epidemic prevention and control. The epidemic prevention and control AI robot comprises a base; an automatic or hand-push type walking mechanism is arranged at the bottom of the base; a cabinet is installed on the base; a temperature measurement and disinfection unit, a mask recovery unit, an information display unit and a main control system are installed in the cabinet, wherein the temperature measurement and disinfection unit is positioned at the upper part; the mask recovery unit is positioned at the lower part; and the information display unit is installed on the front of the cabinet. The fact that body temperature data is automatically stored and recorded in real time, and uploaded to a data management background is realized; all the body temperature data records support querying and tracing; epidemic prevention worker can see epidemic data in someways; therefore, screening and management can be conveniently carried out; the epidemic prevention and control method integrates a series of applications, such as intelligent temperature measurement,sterilization, face recognition, flexible deployment, multi-scene use and data summarization; the workload of reporting the data summarization of the epidemic prevention workers is greatly reduced; the health safety of the epidemic prevention workers can be protected; and the contact infection rate can be reduced.

The invention discloses a method for efficiently screening a leakage region of a water supply pipe network. The method comprises the steps of: dividing the water supply pipe network into virtual DMA districts; determining a leakage index of an urban pipe network; establishing a subjectively decentralized first hierarchy judgment matrix; establishing an objectively decentralized second hierarchy judgment matrix; constructing a visual AHP model by utilizing thermal map cluster; and performing overall hierarchical ranking on the virtual DMA districts from the No.1 to the No.n, verifying consistency of the overall hierarchical ranking, and screening the high leakage region of the water supply pipe network and acquiring specific problems existing in the virtual DMA districts by adopting the visual AHP model obtained in the last step if the consistency verification passes. The method solves the problems that a flow meter or a value needs to be installed on a pipeline of the DMA districts in the prior art and the screening result is inaccurate since the subjective influence in the screening of the leakage region of the water supply pipe network by utilizing the classical analytic hierarchy process is great.

The invention discloses a rolling bearing early composite fault feature extraction method based on progressive VMD. The method comprises the steps of 1, installing an acceleration sensor near a rolling bearing to collect vibration signals during bearing operation; 2, decomposing the acquired vibration signal by using a progressive VMD to obtain a series of decomposed modes; 3, screening the modesobtained after decomposition by using a dual screening criterion based on kurtosis and an energy fluctuation factor to determine a fault mode, and realizing reconstruction of the fault mode; 4, initializing the range and the search step length of the balance parameter alpha, repeating the step 2 and the step 3 for different alpha values, calculating the EFF value of the reconstruction mode, the maximum EFF value corresponding to the optimal alpha value, and then determining the reconstruction mode under the optimal alpha value; and 5, demodulating the reconstructed mode by using TEO to obtaina TEO spectrum, and obtaining a diagnosis result in combination with the bearing related fault characteristic frequency. On the basis of an EEMD recursion idea, the invention provides a progressive decomposition VMD method.