78results about How to "Effective selection" patented technology

The invention relates to an ultra-short-period wind speed prediction method based on a spectral clustering type and genetic optimization extreme learning machine. The method comprises the steps that S1. data are prepared; S2. the prepared data are preprocessed; S3. the preprocessed data are subjected to wavelet transformation; S4. the data obtained after wavelet transformation are subjected to normalization processing; S5. through correlation analysis, the data obtained after normalization processing are selected, so that an input variable is determined; S6. through main component analysis, the input variable generated in the S5 is subjected to dimension reduction processing; S7. through a spectral clustering type method, the data obtained after dimension reduction processing in S6 are subjected to clustering analysis, and an extreme learning machine sample space is formed with the data obtained after normalization processing in the S4; S8. through the extreme learning machine and a genetic algorithm, the data of the extreme learning machine sample space formed in the S7 are subjected to hierarchical prediction; and S9. hierarchical prediction values are added, and an ultra-short-period wind speed prediction value is obtained. Ultra-short-period and multi-step prediction on wind speed are achieved, prediction accuracy is improved, the computing amount is greatly lowered, and prediction efficiency is improved.

The invention discloses an unsupervised image division method based on a multi-target immune cluster integration technology, which mainly solves the problems of poor global optimization capability, single evaluation index, single division scheme form and difficult selection of a plurality of division schemes in the traditional technology. The method comprises the following steps of: (1) extracting gray scale information and wavelet energy information of an image to be divided; (2) sampling the image by using an area-based sampling policy to generate a test sample set; (3) selecting different characteristic vectors to form a plurality of test sample sets; (3) generating a primary division scheme by using an evolution cluster based on a multi-target immune algorithm; (5) integrating and learning an optimal division scheme in the primary division scheme; (6) marking the class of the pixel points of the image according to the selected division scheme; and (7) outputting the image division result. The invention has advantages of high average accuracy in image division and strong robustness and is applicable to the obtaining of image information and the division of image texture.

The invention belongs to the technical field of lean hematite processing, in particular to a lean hematite processing technology which comprises the steps of stage grinding, coarse-fine separation, coarse fraction gravity separation, fine-fraction magnetic separation and centrifuge beneficiation. The lean hematite processing technology disclosed by the invention adopts centrifuge beneficiation to replace reverse flotation beneficiation in the original technology, so that the whole production technological flow is optimized and becomes shorter, and the loss of fine fraction high-grade hematite particles in tailings is reduced; besides, the technological course is more adaptive to properties of the hematite, so that the influence of over-high content of iron carbonate over reverse flotation in the original technological course is effectively solved; and furthermore, due to the adoption of centrifuge beneficiation, reagents does not need to be used and hematite pulp does not need to be heated, so that expenses of reagents and coal are effectively saved and the hematite processing cost is reduced.

The invention discloses an intelligent detection and early warning method and system for an outer envelope in a prostate operation. The method comprises the following steps: 1) acquiring outer envelope image data in a prostate operation video; 2) carrying out gray processing and singular value decomposition on the outer envelope data, and extracting a principal component characteristic value of the image; 3) performing image enhancement on the outer coating image after the image preprocessing in the first step by adopting a deep bilateral learning method; 4) neural network training; 5) detection and early warning. According to the invention, the latest artificial intelligence image recognition technology is utilized to carry out target detection on the prostate envelope, so that an intelligent early warning function is provided for the operations. Different from the existing static medical image recognition technology, the method is characterized in that dynamic images of an operationsite video need to be recognized, early-warned and analyzed. By means of image preprocessing measures such as data expansion, principal component analysis and image enhancement, speed and precision are balanced, and the actual application requirements of prostate operation auxiliary early warning are met.

A mixed bus type-based bus operation time control system and method. Bus routes are classified, and the city bus routes are divided into a part with significant difference and a part with non-significant difference according to the difference of passenger flow in peak and low-ebb periods, so that whether a mixed bus type bus operation mode needs to be adopted is determined; categories of buses of the mixed bus type bus operation mode include buses of a conventional-capacity bus type with M seats and buses of a small-capacity bus type with N seats; bus scheduling work is to determine the car type of a regular bus of each time and departure interval time; based on getting-on and getting-off data of passengers of each station and related data of bus operation cost and the like, an optimal departure interval of the buses of the conventional-capacity bus type of the shift and an optimal departure interval of the buses of the small-capacity car type are calculated; costs of the bus operation modes of the two bus types at the optimal departure intervals are compared, and the regular bus with the lowest cost is selected and executed; and the abovementioned steps are repeated till a time period is over.

The invention relates to a personnel action recognition and position estimation method based on channel state information (CSI) contained in a received signal. The method comprises the steps of in theoffline stage, at different reference position points, performing different action postures by a personnel target, and measuringCSI information of a router transmitting signal on a receiving networkcard; constructing a database and a database corresponding to each action label; preprocessing the constructed training database by using a data normalization and principal component analysis method;and performing action label-based classification learning on the database by utilizing a support vector machine to obtain an action-based classification model; in the online stage, according to the received CSI information, after data normalization and data preprocessing of principal component analysis are carried out, obtaining a personnel target action estimation result through an action classification model; and on the basis, obtaining a position estimation result of the personnel through a position regression model corresponding to the personnel action target estimation result.

The invention discloses a construction method of a morphological filter based on adaptive multi-scale AVG-Hat transformation. The construction method comprises following steps of: 1, determining the number of initial multi-scale structure elements of a signal and an initial structure element value according to a set parameter index based on a bearing fault signal; 2, constructing a set formed by the initial multi-scale structure elements through morphological expansion operation; 3, calculating results of the corresponding morphological AVG-Hat transformation of the bearing fault signal under the initial multi-scale structure elements, and constructing a set of the result; 4, selecting a ratio of the arrangement entropy of the filtered bearing fault signal to sparseness degree of an envelope spectrum as an evaluation index, and adaptively determining an optimal weight coefficient corresponding to the filtered initial multi-scale structure elements; and 5, according to the weight coefficient, constructing an optimal multi-scale morphological AGV-Hat filter. The construction method is advantageous in that the construction method is compatible to signal de-noising and fault feature extraction and accuracy of filtering results can be kept.

The invention provides a three-dimensional object characteristic view selection method, which comprises the following steps of: selecting a three-dimensional object as a target three-dimensional object, and acquiring multi-angle views of three-dimensional objects homogeneous with the target three-dimensional object; acquiring reserved multi-views for the multi-angle views of the three-dimensional objects, and setting initial weights; calculating a view distance matrix of any two groups of multi-angle views which belong to different three-dimensional objects; performing on-view reinforcement learning on the weights of the multi-views of all the three-dimensional objects to obtain final weights; and selecting a plurality of views with the maximum weights as representative views from the reserved views of the original three-dimensional objects. The method analyzes the multiple views of the analyzed object and considers the relevance of the multiple views of the homogeneous objects so as to select the representative multiple views by a novel on-view reinforcement learning method. The method has the advantages of simple design and easy realization.

The invention discloses a method for determining a container model of a single tooth, in particular to a method for determining a container model based on tooth surface characteristics and a local three-dimensional right angle coordinate system, which mainly comprises the following steps: based on a segmentation tooth digital model, reflecting the curve degree at the peak by utilizing complex collection of a triangular facet and a discrete Gaussian curvature, collecting surface feature points of a single tooth; based on the surface feature points of the single tooth, establishing corresponding local three-dimensional right angle coordinates for incisor teeth, cuspid teeth, premolar teeth, and odontoprisis; and respectively establishing the container models of the incisor teeth, the premolar teeth, the odontoprisis and the cuspid teeth by using extreme values of the single tooth model under the three-dimensional right angle coordinate system. The invention has the advantages of novel and reasonable design, convenient implementation, high operating speed, high data processing accuracy and good implementation effect, can extract characteristic points of teeth and facilitate the operation on teeth, and can be used in the tooth correcting system.

The present invention discloses a bearing fault signal feature extraction method based on adaptive multiscale AVGH conversion. The method comprises the steps: 1, according to the parameter index of the bearing fault signals, determining the number of the initial multiscale structure elements and the initial structure element values of signals; 2, constructing the set formed by the initial multiscale structure elements; 3, calculating the results of the morphology AVG-Hat conversion corresponding to the bearing fault vibration signals in the initial multiscale structure elements, and constructing the set of the results; 4, selecting the specific value of the permutation entropy of the filtered bearing fault vibration signals and the spectrum envelope sparseness as an evaluation index and adaptively determining the optimal weight coefficient corresponding to the filtered initial multiscale structure elements; 5, constructing the optimal multiscale morphology AVG-Hat filter according to the weight coefficient; and 6, calculating the processing result of the bearing fault vibration signals through the filter, and extracting the fault feature components in the signals through the signal spectrum envelope analysis to perform bearing fault diagnosis.

The invention relates to an image semantic segmentation error annotation data screening method and system. The method comprises the following steps: carrying out image recognition prediction on an original image participating in training by utilizing a trained network model to obtain a confidence coefficient of each pixel point in the original image, and calculating according to the confidence coefficient to obtain an accuracy rate, a recall rate and / or a cross-to-parallel ratio of a target category; judging whether the annotation data are suspicious error annotations or not by utilizing the obtained accuracy rate, recall rate and / or cross-merging ratio; and screening out suspected mistakenly labeled data. According to the method and the system, data with wrong annotations can be screenedout and can be used for secondary annotation and training, and an original model is optimized, so that the model recognition capability is improved.

The invention relates to an accelerated determining method for heat-resisting life of a plastic material and belongs to the technical field of a determining method for service life of a resin material. The accelerated determining method comprises the following processing steps: 1) preparing a sample; 2) hanging the sample under a stress-free condition in an air exchange type high-temperature accelerated aging box, blasting drying box or drying box, heatproof aging till the material enters into powdered state, thereby obtaining aging time; 3) confirming use temperature of the sample under actual use state; and 4) confirming the practical service life of the sample. The accelerated determining method for the heat-resisting life of the plastic material has the beneficial effects that: 1) the long-time durability evaluation of a great quantity of parts is avoided, a material aging process is simulated under a high-temperature condition and the material aging is accelerated, so that the material evaluation period is reduced by above 80%, the developing period of the product is shortened and a great quantity of manpower and material resources are saved, and 2) the material sample can be effectively selected according to the accelerated determining method, the flexibility is high and the defect that large equipment is required for evaluating large parts is overcome.

The invention discloses an AP selection indoor positioning method of a path loss model, which comprises the following steps: (1) offline data collection; (2) establishment of an initial fingerprint database; (3) division of sub-areas of the fingerprint database; (4) effective selection of the AP. Although the AP selection method based on the path loss model can effectively select the APs that havesimilar contributions to positioning, the stability of the AP is not considered, and the Fisher criterion can achieve effective selection of the stable AP. Therefore, the AP selection indoor positioning method of the path loss model in the present invention introduces the Fisher criterion into the AP selection method based on the path loss model, thereby making up for the problem of insufficientconsideration for selection of the stable AP, so that the constructed wireless map is more reliable, thereby improving the positioning accuracy.

The invention discloses a question pushing method. Education learning system, the method is used for obtaining relevance characteristics among questions and effectively pushing the questions accordingto changes of knowledge point mastering degrees of different students along with time, and comprises the following steps that historical question data are obtained, the historical question data are classified according to knowledge points, and each knowledge point forms a corresponding question data set; inputting the question data set into a preset recurrent neural network model according to a time sequence to obtain relevance characteristics; obtaining a question state st of the question bank data based on the relevance characteristics and preset current answering characteristics of the question bank data; and according to the question state st and the Markov decision process, obtaining a question queuing sequence, and selecting the front question in the queuing sequence as a target question for pushing. The invention further discloses a question pushing device, electronic equipment and a computer storage medium. The questions are effectively pushed for different students.

The invention discloses an intelligent sorting AGV. The AGV comprises an AGV body, a goods allocation support, a material box, a lifting system, a goods carrying table and goods forks. The goods allocation support is installed at two ends or one end of the AGV body, the material box is placed on the goods allocation support, the lifting system is installed on the AGV body, the goods carrying tableis installed on the lifting system, and the goods forks are installed on the goods carrying table. According to the invention, one or more material boxes can be intelligently sorted according to tasks given by a single-machine control system and each material box is a material box needing to be sorted, so the sorting efficiency is high, invalid logistics and logistics waiting do not exist; the lifting system is arranged, so the positions of the goods carrying table and the goods forks can be controlled through the lifting system so that the material boxes can be accurately clamped; a warehouse area goods shelf does not need to be jacked up, so the sorting efficiency and the automation degree are high.

The invention discloses a multi-satellite combined inversion method for the earth surface soil humidity based on GNSS-IR, and belongs to the technical field of microwave remote sensing. The method comprises the steps of extracting a satellite altitude angle, an azimuth angle and a signal-to-noise ratio by adopting monitoring data of a measurement type GNSS receiver, and selecting multiple effective satellites through a space-time overlap ratio; separating reflected signals in the signal-to-noise ratios of the satellites by wavelet analysis, resampling the reflected signals with a low elevationangle, and obtaining the multipath delay phase of each satellite through nonlinear least square sine fitting; and finally, establishing a nonlinear relationship between the multipath delay phases ofthe multiple satellites and the soil humidity through a least square support vector machine to realize high-precision inversion of the soil humidity. The method is wide in application range and high in inversion precision, the defects of a traditional satellite reflection signal separation method can be effectively overcome, the problem that soil humidity information around a measurement station is difficult to comprehensively reflect only by adopting a single satellite is solved, and the problems in the prior art are solved.

The invention discloses a method for onsite load test and evaluation of wind turbine blades and aims to provide an onsite load evaluation technique for accurately testing and evaluating load borne by the blades in practical operation so as to provide guidance for design and operation of the blades and even a unit. The method has the advantage that corresponding load onsite evaluation methods can be established aiming at wind resource conditions of specific wind fields and special blade structural forms to provide support for design, research and development of novel wind turbine blades and operation evaluation of existing wind turbine blades.

A process for extracting useful material from the rejected ore powder by high-intensity gradient magentic field includes such steps as creating a strong magnetic field (at least 6X10 to the power 5 A / m), arranging a secondary magnetic pole unit in said magnetic field, making the rejected ore powder flow through said magnetic field, and recovering said useful material from the secondary magnetic pole unit.

The invention discloses a foreign matter intrusion detection method and system for a power transmission line. The method comprises the steps of extracting a straight line segment in an image by utilizing Hough transform; and selecting two straight line segments with the length greater than a threshold value and the farthest distance, connecting the two straight line segments to form an initial segmentation frame, expanding the initial segmentation frame to obtain an expanded segmentation frame, cutting out a lead area image block from the preprocessed image according to the expanded segmentation frame, and inputting the lead area image block into a foreign matter intrusion detection model to obtain a foreign matter detection result. The region of interest is effectively selected before foreign matter intrusion detection, interference of foreign matters in the background is avoided, and the foreign matter detection precision is improved.

The invention relates to a large-scale building user behavior analysis method based on improved clustering fusion. The method is used for determining the power consumption mode of the large-scale building user. The method comprises the following steps: (1) obtaining the total load data of the large-scale building user to be analyzed and the sub-item measurement data; (2) constructing the comprehensive evaluation index of clustering effect, and selecting various high-quality clustering methods; (3) Clustering the total load data of large-scale building users to obtain different clustering results by using the selected high-quality clustering method; (4) fusing the clustering results obtained by the high-quality clustering method to obtain the final power consumption mode. Compared with theprior art, the invention can absorb the advantages of different single clustering algorithms, the effectiveness and the accuracy of the clustering algorithm are higher than those of the single clustering method, and the expansibility is improved.

The invention discloses a power consumer power consumption evaluation method based on an intelligent power distribution network. The method comprises the following steps: S1, collecting data information of a power grid side and a user side; s2, cleaning the collected data information to obtain effective data information; and S3, according to the effective data information, establishing a power consumer power consumption evaluation model, and evaluating the intelligent power distribution network. According to the method, the data information of the power grid side and the user side is accessed,and the data information is cleaned by adopting an algorithm, so that effective selection of massive information is realized, and a foundation is laid for evaluation distribution; by setting a space-time algorithm for power consumer electricity consumption behaviors, statistics and classification of mass data are realized, the electricity consumption safety of the country and the society is improved, and the normal benefits of power grid enterprises and electricity consumers are guaranteed.

The invention discloses a defectives picking device in candy production. The defectives picking device comprises an industrial robot, a first conveying belt, a control cabinet and a vacuum sucker, wherein a defectives recycling box is arranged in front of the first conveying belt; an electronic weighing platform is arranged below the defectives recycling box; an indicating lamp is arranged on theelectronic weighing platform; a power switch is arranged behind the first conveying belt; the industrial robot is arranged on one side of the first conveying belt; a base is arranged on the industrialrobot; a speed reducer is arranged on the base; a first servo motor is arranged on one side of the speed reducer; a big arm is arranged above the first servo motor; a second servo motor is arranged on the big arm; a first rotary arm is arranged above the second servo motor. The defectives picking device disclosed by the invention has the following beneficial effects: unqualified candies can be effectively picked out so as to guarantee the product quality; moreover, with a high degree of automation, the device is convenient to operate, has a fast picking speed and is easy to repair in the caseof failure.

The invention discloses a cress seed selecting method, which comprises the following steps of: first, sifting harvested cress seeds to remove tiny and thick impurities, mixing the seeds with river sand for one month, and separating the seeds from the sand through water; floating the separated seeds for one or two days, and airing the seeds taken out of water; grading the aired seeds by winnowing, and simultaneously, removing part of impurities and seeds with poor quality; and finally, manually pressing the seeds which are classified by winnowing to further remove impurities. The method adopts the procedures of storing with san, floating on water and pressing and combines traditional methods of sifting and winnowing. Therefore, not only are the cress seeds selected effectively, but also the way of storing seeds in sand is added to the selection process of the cress seeds. In this way, two effects are realized because the seed selection and the relieve of seed dormancy can be finished in a combined way, and a large quantity of cress seeds can be selected within a short time. The method overcomes the difficulty of heavy workload in a traditional way of selecting seeds manually, and can be used for realizing the goal of breeding cress seeds in an industrialized way.

The invention discloses a power supply capability evaluation method based on an intelligent power distribution network, and the method comprises the following steps: S1, collecting the data information of a power grid side and a user side, cleaning the data information, obtaining effective data information, and building a data platform; and S2, establishing a power utilization quality evaluation platform of the intelligent power distribution network, and providing a power utilization governance auxiliary decision. The data information of the power grid side and the user side is accessed, and the data information is cleaned by adopting an algorithm, so that effective selection of massive information is realized, and a foundation is laid for evaluation distribution. By analyzing the effective data information from multiple dimensions, an evaluation system for the power distribution network is established, and scientific evaluation for power grid evaluation is realized. Distribution network weak points and power supply quality abnormal points are found in time through an evaluation system. A basis is provided for evaluation, and safe and high-quality operation of a power grid is guaranteed.

The invention relates to a feature selection method and device for multi-label data, and belongs to the technical field of data classification. According to the method, firstly, the prior probabilityof a mark is used as the weight of the mark, the correlation between features and the mark is calculated, the features are pre-screened according to the correlation, and the correlation between the features and the mark is as large as possible while the subsequent calculation amount is reduced; the method comprises the following steps: firstly, marking samples, then dividing the same type and different types of samples by utilizing correlation link values among marking sets of the samples, finally, calculating feature weight values according to a weight updating formula, sorting the feature weight values, and selecting an optimal feature subset. Through the process, the optimal feature subset can be effectively selected, and the classification precision of the multi-label feature selectionalgorithm is improved.

The invention provides an information propagation prediction method suitable for a file sharing oriented mobile social network, and provides a method for effectively determining a seed node, so that the selection of the seed node is more reasonable and effective. At the same time, a method to simplify the network structure is given. On the basis of not affecting the overall dissemination, the nodes which have no influence on the information dissemination are removed to improve the efficiency of dissemination. The model builds a bridge between the node influence calculation and the classical node authoritative value calculation, and then studies the information dissemination maximization strategy in large-scale social networks, which provides an efficient solution for the information dissemination prediction and calculation in large-scale social networks.

The invention discloses an aging method of a lithium ion battery. The aging method includes the steps of charging the lithium ion battery in the constant current and the constant voltage until the voltage of the lithium ion battery reaches 3.000V to 3.700V, then, placing the lithium ion battery in an environment with the temperature from 45.0 DEG C to 85.0 DEG C to be aged for 1h to 24h. According to aging method, on the premise that safety production is ensured in the aging process, the aging time is greatly shortened, batteries with internal micro short circuits are more effectively selected, and batteries with potential safety hazards are prevented from entering the market.

The invention discloses a crack propagation signal calibration device and method based on a DCPD method. The crack propagation signal calibration device comprises a cavity body, a cavity body base anda plastic slice, the cavity body is arranged at the top of the cavity body base, the shape of the cavity body is of a cuboid structure, a cavity enclosed by cavity walls of the cavity body is a cuboid inner cavity, gaps are vertically formed in the middle of the front wall and the middle of the rear wall correspondingly, scales are arranged on the two sides of the gaps in the length direction correspondingly, the cuboid inner cavity is filled with liquid metal, and the plastic slice can be inserted into the gaps to cut the liquid metal. According to the crack propagation signal calibration device and method based on the DCPD method, the condition that the liquid metal is cut to be capable of simulating crack propagation intuitively and vividly through the liquid state of the liquid metalat the normal temperature and the performance of certain fluidity of the liquid metal is utilized, an intuitive representational method for simulating crack propagation signals is provided for researching and developing a crack monitoring instrument based on the DCPD method, wherein the method can be used repeatedly, the advantages of stable performance and repeatable operation are achieved, and the manufacturing cost of the crack monitoring instrument based on the DCPD method is greatly lowered.

The invention provides a high-precision structural reliability analysis method based on a three-weighted response surface. The method comprises the following steps: 1, determining a random variable and a response surface function form; 2, a classical response surface method is adopted for the first iteration; 3, in subsequent iteration, selecting an initial sample point by adopting a Bucher method, and calculating a true value of a limit state function; 4, selecting a newly added final sample point in the iteration by adopting a continuous interpolation method, and calculating a true value ofa limit state function of the final sample point; 5, constructing a regression matrix; 6, constructing a weight matrix; 7, solving an undetermined coefficient in the undetermined response surface function by using a weighted least square method; 8, calculating reliability indexes, checking calculation points and limit state function values of the checking calculation points by adopting an important sampling method; 9, calculating a new sampling center by adopting an interpolation method; tenthly, repeating the third step to the ninth step until the difference between the reliability indexes obtained through two times of iterative computation meets the precision requirement; the method is scientific and good in manufacturability and has wide application and popularization value.

Two opposite position marks are designed for procedure of fabricating crystal grain of light emitting diode. The two opposite position marks are located at same metal plane. The two opposite position marks can be setup in modes of corner-to-corner, or corner-to-edge, or edge-to-edge. Since two marks are located at same metal plane, their surface roughness is similar. Thus, device for fabricating crystal grain can recognize them easily. Selecting crystal grain effectively, the invention reduces unusual condition from happening greatly so as to raise integrated capacity and ration of fine products.