156results about How to "Take into account the accuracy" patented technology

The invention discloses a method for quickly identifying two-dimension code system type in images, comprising a learning training process and a classification identifying process. The learning training process is as follows: collecting and building a sample image set of various two-dimension code images; converting each sample image into a grey image, performing Gaussian smoothing filtering and binaryzation to obtain binaryzation images; scanning prospect boundaries of the binaryzation images in the horizontal and vertical directions, obtaining an outer boundary point set of the two-dimension code; enabling the two-dimension code to be horizontal by rotating images, achieving horizontal correction of the two-dimension code; performing partitioning, combining and normalizing for the two-dimension code; performing fast wavelet transform for the normalized sample image to obtain a wavelet characteristic sample set. The classification identifying process is as follows: extracting wavelet characteristic of the to-be-identified image to build a distance measurement model; using the K nearest neighbor algorithm to identify code system type. The method is convenient and quick, has real-time performance, accuracy and high identification rate.

The invention discloses an information pushing method. The method comprises the steps as follows: obtaining scoring parameters corresponding to user historical operations in a preset duration for the same type of pushed information, and meanwhile, obtaining similarity parameters between to-be-pushed information and information corresponding to the user historical operations; according to a preset diversity regulation factor and a preset accuracy regulation factor, calculating pushing scores corresponding to the to-be-pushed information according to a preset algorithm by using the obtained scoring parameters and the similarity parameters; pushing the corresponding to-be-pushed information according to the obtained pushing scores. The invention further discloses an information pushing device. The information pushing method and device have the benefit that the information is pushed according to the pushing scores due to the fact that the pushing scores of the to-be-pushed information are calculated on the premise that the accuracy and the diversity of the to-be-pushed information are balanced, and both the accuracy and the diversity of the information pushing are considered.

The embodiments of the invention disclose a positioning identification and a positioning method thereof. The positioning identification comprises at least four identification mark points; and an annular encoding point which comprises two or more than two annular patterns used for storing number information of the identification mark points, wherein the number information is expressed through one or multiple pieces of data, each annular pattern is composed of one or more pattern units according to a predetermined splicing sequence, and each pattern unit is provided with at least one graphic feature used for distinguishing the pattern units so as to store one piece of data. The information storage capacity of original annular encoding is expanded, compared to a conventional two-dimensional code storage mode, the encoding mode is simple, the information reading speed is fast, the identification mark points and encoding points are also separated, and the positioning accuracy is improved.

The invention relates to an electromyographic signal classification method based on a multi-kernel support vector machine. For a sample with complex distribution, based on the classification performance of a single-kernel support vector machine, the classification accuracy and the quantity of support vectors are easily influenced. The method combines a multi-kernel support vector machine method with a binary tree combination strategy and comprises the following specific steps of: collecting electromyographic signals of the lower limbs of a human body through an electromyographic signal acquisition instrument; denoising the electromyographic signals containing interference noise by using a wavelet coefficient inter-scale correlation denoising method; extracting the features of the denoised electromyographic signals to obtain the features of the electromyographic signals by using denoised wavelet coefficients; and classifying on the basis of the multi-kernel support vector machine. The method can well meet the multi-classification requirement of lower extremity prosthesis control, and takes into account both accuracy and instantaneity, and has broad application prospects in the multi-movement mode recognition of intelligent prosthesis control.

The invention relates to a computer technology, and provides a news information recommendation method for a smart television, used for solving the problem that news information cannot be recommended by the existing smart television. The technical scheme provided in the invention can be summarized by comprising the following steps of: constructing a corresponding massive news information resource library in the background at first, then, constructing the latest news information library from the massive news information resource library according to the time freshness, analyzing an information browsing record under a terminal account stored by a service side, extracting a characteristic keyword set, obtaining information of the geographical position, where a smart television client side is, and a social hotspot keyword set in the current time simultaneously, then, searching an information content having relatively high association degree with keywords or nearby information contents from the latest news information library by using browsing record characteristic keywords, social hotspot keywords and the position information of the smart television client side, and finally, obtaining the final information recommendation set. The news information recommendation method for the smart television disclosed by the invention is convenient for users and applicable to smart television systems.

The invention provides an identification method of rice varieties based on laser-induced breakdown spectroscopy (LIBS). The identification method comprises the following steps of S1, measuring a spectrum in a certain range of N standard rice varieties; S2, performing data processing on the spectrum of the N standard rice varieties, S3, performing spectral line identification on data subjected to average processing, so as to acquire wave crest spectral lines of chemical elements of which the contents are at the top; S4, selecting the wave crest spectral lines in S3 to synthesize a characteristic spectrum as an input value of a BP neural network and training to obtain a BP neural network structure; S5, measuring the spectrum in the certain range of one rice variety to be measured by utilizing the LIBS; S6, performing data processing on the spectrum in S5; S7, performing spectral line identification on the spectrum of the rice variety to be measured to obtain the spectral lines of elements which are same with the elements in S3; S8, selecting the wave crest spectral lines in S7 to synthesize the characteristic spectrum as the input value, utilizing the BP neural network in S4 to identify the rice variety to be measured, so as to obtain an identification result; S9, judging whether the identification result is one of the N standard rice varieties or not.

The invention relates to a method for matching and retrieving three-dimensional building models based on shape recognition. At present, relatively mature information retrieving technologies are still limited to the aspect of text retrievals. Retrievals of images, three-dimensional models and the like are usually retrievals of relevant text information, which makes retrievals of such data and models low in accuracy. In the method, both self upper direction characteristics of a three-dimension building model and a principal direction on a horizontal plane are used, recognition theories are combined with classic LightField description operators by a model normalization method, and Horizontal LightField description operators are extracted from a series of model images corresponding to a simplified field by simplifying and rendering a light field of a standardized building model, wherein the Horizontal LightField description operators only require rotary matching in horizontal direction. Matching and retrieving of the three-dimensional building model is realized by utilizing shape qualitative description operators and shape quantitative description operators. Compared with previous methods, the method of the invention can better consider human recognition of space objects, and can give consideration to both accuracy and high efficiency of coupling. The invention provides a new method for building a three-dimensional digital city system in a rapid, cheap and automatic way.

The invention is a confirming method for medical image bad dots automatic detection threshold. The invention detects the bad dot with formula alpha(x, y) = (grav(x, y)-avg(x, y)) / avg(x, y) and I median= Median (I original), I (x, y) =| I median - I original|, the automatic checking threshold of bad dot includes: collecting detection image, threshold detection, video detection, and detection threshold definition. It is a process method combining threshold detection and video detection. The method is simple, upgrades the detection speed, and guarantees the high accuracy.

The invention discloses a micro-blog emotion analysis method and system. The micro-blog emotion analysis method and system provided by the invention adopt several micro-blog text data collected by a focused network crawler in a preset period of time as target topic data, and input each target topic data into a micro-blog emotion classifier to obtain the emotion types of each target topic data. Theinvention adopts a weakly supervised learning method based on expression words and emotion words to filter emotion micro-blog, A Chinese micro-blog corpus is constructed by selecting the same amountof positive micro-blog data and negative micro-blog data, and the micro-blog emotion classifier obtained by training FastText classifier with the corpus can give attention to the accuracy and timeliness of classification, and can accurately reflect the emotional trend of topic.

Provided is a method of measuring and calculating distribution of the water temperatures of a river channel below a dam of a reservoir. Actually-measured data of the weather and the hydrology are combined, changes of the water temperatures along the river channel below the dam of the reservoir are measured and calculated, a position below the dam of the reservoir serves as the first measuring and calculating section, and the spacing distance of sections along the river channel is calculated according to an integral multiple of the one-day and one-night flow path of a river reach with the mean velocity. Heat exchange of water in the river channel is fully considered, detailed analysis does not need to be carried out on hydrodynamic processes, the mean velocity is used for replacing hydrodynamic calculation, the analysis of hydrodynamic equations required by numerical simulation calculation can be avoided, and workloads of calculation are reduced greatly. Meanwhile, according to the method, main factors leading to the changes of the water temperatures of the river channel are fully considered, and accuracy of calculation is guaranteed. As for the method, the good accuracy is obtained, the amount of calculation is relatively small, complex boundary condition data are not needed, work efficiency can be improved greatly, application fields can be widened, the accuracy and usability can be taken into account better, and the method can effectively assist in making discharging schemes of the water temperatures and the flow quantity of the reservoir.

The invention discloses a method for testing the balance degree of the electrolyte of a portable all-vanadium redox flow battery. The method comprises the following steps that: step 1, an absorbance-concentration standard working curve is drawn; 2, a positive pole electrolyte and a negative pole electrolyte are sampled separately and simultaneously, and the sampled positive pole electrolyte and negative pole electrolyte are subjected to isopyknic mixing so as to undergo an oxidation-reduction reaction, and a mixed solution A can be obtained; step 3, a portable ultraviolet and visible spectrophotometer is adopted to obtain the concentration CV<3+> of trivalent vanadium ions in the mixed solution A; step 4, a reducing agent is added into the mixed solution A, so that a mixed solution B can be obtained, and the portable ultraviolet and visible spectrophotometer is adopted to obtain the maximum concentration Cm of the trivalent vanadium ions in the mixed solution B within testing time, andthe concentration of tetravalent vanadium ions in the mixed solution A is CVO<2+>, and CVO<2+>=Cm-CV3<+>; and step 5, the unbalance degree K of the whole electrolyte is calculated according to a formula that K=(CVO2<+>-CV3<+>) / (CVO2<+>+CV3<+>). The method has the advantages of convenience in operation, simplicity, rapidity, portable convenience, accurate detection and suitability for the on-sitedetection of the actual operation place of flow batteries.

The present invention relates to tumor neoantigen polypeptide and an application thereof. Specifically, the present invention provides a newly discovered tumor neoantigen polypeptide derived from human papilloma virus HPV16 and KRAS oncogenes, a screening method of the tumor neoantigen polypeptide, a compound formed by the polypeptide and an MHC molecule, and uses of the polypeptide and the compound. and meanwhile, the present invention also provides molecules and cells combined with the polypeptide or the compound, and uses of the molecules and the cells.

The invention relates to a seeding monitoring method based on a wheat seeder. The method includes the following detailed steps that (1), an analog voltage signal and land wheel speed data are collected; (2) the analog voltage signal is analyzed according to feature information of the analog voltage signal; (3) the seeding state is determined according to the analyzed feature information of the analog voltage signal and the land wheel speed data. According to the method, a high-power infrared photoelectric seeding sensor is used, the seeding sensor directly outputs the analog voltage signal, a single chip microcomputer is used for collecting the voltage signal, a state judgment algorithm is designed according to features of voltage, voltage change, and continuity and discontinuity of wheat sowing in line, the defect of poor anti-dusting performance of a photoelectric monitoring device is overcome, three seeding states such as normal state, blocking state and blank pipe state can be well monitored, real-time performance and accuracy can be considered simultaneously, and miss-seeding can be timely and accurately discovered. The invention further discloses a seeding monitoring device and system based on the wheat seeder.

The invention relates to a positioning system of an overhauling trolley for urban rail transit. The positioning system is based on the overhauling trolley and rails, and comprises a photoelectric encoder, a beacon, two photoelectric proximity sensors and a processor. The photoelectric encoder is arranged on an axle of a rotating wheel of the overhauling trolley and is connected with the processor, the beacon is fixed between sleepers of the rails, and the two photoelectric proximity sensors are respectively mounted at a front position and a rear position of a chassis of the overhauling trolley and are respectively connected with the processor. Compared with the prior art, the positioning system has the advantages of simple structure, effective functions, adaptability, flexibility, wide application range and the like.

The invention relates to a power battery package system high voltage interlocking connector detection system. The detection system comprises a plurality of high voltage interlocking connectors, each of the high voltage interlocking connectors comprises a plug and a socket, which are connected together. A signal generator, a signal detector, and a processor are provided. Two ends of the contacts ofeach socket are parallel connected with a detection resistor, and the signal generator is connected with all of the detection resistors, and the signal detector is connected with two ends of each detection resistors. Compared with the prior art, the invention is advantageous in that the connection state of all of the high voltage interlocking connectors can be detected, and the current state of each concrete high voltage interlocking connector can be discriminated concretely, and in addition, the battery package system can be detected timely during the use or after the assembly of the connectors, and therefore high efficiency and accuracy are realized at the same time, and the use safety of the battery package high voltage circuit can be guaranteed.

The invention discloses an automatic reciprocating spraying device. The automatic reciprocating spraying device mainly comprises a workpiece spraying assembly, a clamping rotating movement unit, a spraying adjusting unit and a reciprocating movement unit. The workpiece spraying assembly comprises a protection sleeve and a workpiece. The clamping rotating movement unit comprises a first stepping motor, a fixed lower tip, a self-rotation upper tip, a lock nut and a clamping air cylinder. The spraying adjusting unit comprises a spraying platform, a T-shaped guiding shaft support, a connecting shaft and a spray gun. The reciprocating movement unit comprises a second stepping motor, a coupler, a ball screw, a fixing side support, a supporting side support, guiding shafts, linear bearings and a guiding shaft support. Upright fixing of the workpiece is guaranteed through the fixed lower tip and the self-rotation upper tip, pressing and detaching of the workpiece are achieved through the clamping air cylinder, the spraying platform is connected with the linear bearings on the guiding shafts and a nut on the ball screw, and the linear movement of the spraying platform is achieved; and the stepping motor drives the workpiece to do the rotation movement, and the automatic reciprocating spraying device can guarantee the spraying thickness and uniformity for the sprayed workpiece.

The invention relates to a risk oriented hazard analysis method for a chemical process device. The risk oriented hazard analysis method is mainly used for solving the problems that in the prior art, when risk oriented hazard analysis is carried out on a chemical process device, time consumption is high, and investment is large. By the adoption of the risk oriented hazard analysis method for the chemical process device, the safety integrity level is determined when a safety instrument system is taken as a protective measure; when the safety evaluation on the functions of the safety instrument system is carried out, a risk graph is used for analyzing the risks related to danger sources, a set of risk parameters are used for evaluating the risk levels, a risk level is assigned to each danger source, and an improved and extended ROHA process risk graph is used for determining the safety integrity level. The risk oriented hazard analysis method is characterized in that according to the technical scheme shown in the ROHA process risk graph (shown in figure 1), the problems are effectively solved, and the risk oriented hazard analysis method can be used for risk oriented hazard analysis of the chemical process device.

The invention discloses a high-order CFD implicit time propulsion method based on a linearization matrix hybrid solution mode, and belongs to the technical field of computational fluid mechanics. An existing method is very poor in universality and low in efficiency, computing codes need to be recompiled for different items in a control equation, programming implementation is not facilitated, and therefore the workload of fluid simulation is large, the efficiency is low, and the computing result and flow field information of fluid simulation cannot be obtained rapidly and effectively. Accordingto the invention, a total differential and chain type rule is adopted; the derivative of the dependency on the degree of freedom is directly converted into a product of the derivative of the dependency on the variable and the derivative of the variable on the degree of freedom; Compared with a traditional method, the method has the advantages that the calculation amount can be greatly reduced, the calculation efficiency is improved, the universality is realized, the programming realization is facilitated, the accuracy of the linearization matrix is considered, and the fluid simulation resultand the flow field information can be quickly obtained.

The embodiment of the invention provides a mold state monitoring method and device, an industrial personal computer, a storage medium and a system, and the method comprises the steps of obtaining a template image of a target mold and a to-be-detected image, the template image being an image of the target mold in a normal state, and the template image comprising a region of interest of the image of the target mold; according to the region of interest of the template image, adopting a pyramid template matching algorithm to identify and position the to-be-detected image to obtain a target region of the to-be-detected image; determining sensitive sub-regions contained in the target region according to a pixel difference between the target region and the region of interest; and according to the area of the sensitive sub-region, determining whether the target mold is abnormal, so that automatic monitoring of the mold quality is realized, the real-time performance and the accuracy of mold abnormality detection are considered, and the mold monitoring quality and the production efficiency are improved.

The invention discloses a multi-target active power optimization scheduling method and system based on linear relaxation and proportional weight. The method comprises the following steps of: establishing a mathematical model of multi-target active power optimization scheduling of a power system; simplifying the mathematical model of multi-target active power optimization scheduling by using a model variable and constraint deletion method to obtain a quadratic convex programming model; adopting linear relaxation line loss constraints for the quadratic convex programming model, performing solving and determining, and if all the line loss constraints meet equation constraints or threshold values, solving the quadratic convex programming model subjected to linear relaxation line loss constraints through adoption of a proportional weight multi-objective optimization method; otherwise, performing mixed integer programming processing on the line loss constraints which do not meet the equalityconstraints or threshold values until all the line loss constraints meet the equality constraints or threshold values. The invention aims to provide the multi-target active power optimization scheduling method and system based on linear relaxation and proportional weight, the complexity of multi-target optimization calculation can be effectively reduced, and both the calculation efficiency and the result accuracy are considered.

The invention discloses a business processing method, device and system based on fuzzy logic. The method comprises the steps of obtaining business data, carrying out the structural processing, setting the fuzzy logic, building a corresponding evaluation system according to a fuzzification concept inputted by a business expert, building at least one mapping relation table between the fuzzification concept and a fuzzy logic judgment result, constructing a neural network model through a deep learning algorithm, performing driving processing on fuzzy logic, establishing a knowledge graph in combination with experience knowledge of business experts, performing classification, performing reasoning and description on a fuzzification concept, iterating reasoning and description results into the model to obtain a final judgment basis, and processing business requirements to obtain and output results. On the premise that a large number of existing samples are not needed, the decision-making efficiency of business experts can be effectively improved, and the decision-making accuracy is considered.

The invention provides a method and system for quickly detecting fall in showering. The method comprises the following steps: continuously performing observation for multiple times and respectively calculating power spectral density sum in a body movement frequency observation range of each time of observation; and performing pre-alarm of falling by comparing the density sum with a power density sum threshold value in falling and after falling, and performing falling confirmation by comparing power spectral density in the respiratory rate observation range with obtained echo interference levelunder a corresponding condition. The method is suitable for non-contact non-visual fall detection, does not influence normal activities, does not invade personal privacy, greatly improves the detection practicability, and has extremely high popularization value.

The invention discloses a predictive control method for the optimized output track of a DC-DC converter discrete model. The predictive control method comprises the following steps: step 1, establishing a discrete predictive model of a DC-DC converter; step 2, determining the steady-state voltage and the current characteristic of the discrete predictive model; step 3, determining the objective function of the discrete predictive model; step 4, determining a predictive control strategy. The method has the simple mathematical model and small computational amount of optimization problem solving and can realize online computation of control decisions. The method can work alone in a CCM, also can work alone in a DCM, and also can work in the mixed mode of the CCM and DCM. The method can avoid the voltage overshoot phenomenon of the DC-DC converter in the starting process and also has the accuracy and fastness of the DC-DC converter for regulation of output voltage under various running conditions.

The invention relates to a method for predicting turning machining deformation of a thin-wall complex curved surface rotating member, and belongs to the technical field of machining. Firstly, a tool cutting edge angle and cutting line speed are taken as experimental factors, a two-factor multi-level full factor test is performed to obtain residual stress distribution under each kind of combination; then, according to the obtained multiple sets of residual stress distribution, local residual stress values corresponding to different places in a surface stress layer are obtained by a multidimensional linear interpolation method, and the reconstruction of the machining of a non-even residual stress field is completed; then, according to a position relationship of a local coordinate system anda global coordinate system, the conversion from a local stress field to a global stress field is achieved; lastly, loads and boundary conditions are applied, and the rotating member turning machiningdeformation on the thin-wall complex curved surface is calculated. According to the method for predicting the turning machining deformation of the thin-wall complex curved surface rotating member, thenon-uniformity of processing residual stress distribution caused by related parameter changes in the actual processing of the thin-wall complex curved surface rotating member is considered, thereby the calculation amount of the method is moderate, the quality requirement for grids is general, and the efficiency and the accuracy are simultaneously achieved.

The invention discloses an optical proximity correction method for improving a process window. The method comprises the following steps: 1, inputting a target graph and cutting the target graph; 2, simulating the target graph and calculating initial values of first to third parameters; 3, setting the parameters; 4, correcting according to the current value of the first parameter to obtain a corrected graph, recalculating to obtain the updated values of the first to third parameters, and calculating a corrected comprehensive error; 5, judging whether the corrected comprehensive error is not more than a maximum permissible error, and judging whether the frequency of performing the step 4 is equal to or more than a maximum correction frequency; when a judgment result in the step 5 is that the corrected comprehensive error is more than the maximum permissible error and the frequency of performing the step 4 is less than the maximum correction frequency, re-performing the steps 4 and 5.

The embodiment of the invention provides a three-dimensional reconstruction method, device and system, a medium and computer equipment. The method comprises the following steps: carrying out the three-dimensional reconstruction of an image of a target object through a three-dimensional reconstruction network, obtaining an initial parameter of the target object, and enabling the initial parameter to be used for building a three-dimensional model of the target object; optimizing the initial parameters based on pre-acquired supervision information used for representing features of a target object to obtain optimized parameters; and performing skeleton skin processing based on the optimization parameters, and establishing a three-dimensional model of the target object.

The invention, which relates to the field of mechanical fault diagnosis, discloses a mechanical fault diagnosis method based on a dual-tree complex wavelet packet sub-band average kurtosis graph, thereby improving the robustness and accuracy of mechanical fault feature extraction and fault type identification. According to the invention, with utilization of non-stationary and instantaneous impactcharacteristics of a mechanical fault vibration signal, a sub-band average kurtosis graph method based on a dual-tree complex wavelet packet is provided, wherein the efficiency and the accuracy are both considered in dual-tree complex wavelet packet transformation; the sub-band average spectral kurtosis can eliminate the influence of random impact interference on the kurtosis graph, can find the frequency band where the fault characteristic signal is located more accurately; the filtered signal is analyzed through a square envelope spectrum and thus the accurate extraction of the bearing faultcharacteristics and the precise recognition of the fault type are realized. The method is suitable for mechanical fault diagnosis.

The invention discloses a beam measurement method of user equipment, the user equipment and a computer readable medium. The beam measurement method of the user equipment comprises the following steps:computing handover frequency of the beam in the preset duration; computing the measurement cycle of the beam based on the handover frequency of the beam; and performing intra-cell beam measurement and inter-cell beam measurement based on the measurement cycle of the beam and a measurement result of the beam. By applying the above scheme, the measurement cycle and measurement content can be dynamically adjusted according to the handover frequency of the beam and the measurement result of the beam, and the measurement accuracy and the user equipment power consumption are both taken into consideration.

The invention discloses an Internet of Vehicles intrusion detection method and system based on deep reinforcement learning. The method comprises the following steps: calculating statistical characteristics of traffic data at historical moments; establishing a flow prediction model based on a deep reinforcement learning algorithm-a deep deterministic strategy gradient algorithm; wherein the input of the flow prediction model is the statistical characteristics, and the output of the flow prediction model is predicted flow; establishing an intrusion detection model based on a depth deterministic strategy gradient algorithm; wherein the input of the intrusion detection model is the statistical characteristics and the predicted traffic, and the output of the intrusion detection model is a traffic threshold; and performing Internet of Vehicles intrusion detection by comparing the predicted traffic with the traffic threshold. According to the method, the complexity variability of the Internet of Vehicles, the finiteness of computing resources of the infrastructure server side and the accuracy of network intrusion detection can be considered, and the method is more practical. For a complex system of the Internet of Vehicles, the intrusion detection method based on deep reinforcement learning provided by the invention has better performance compared with other existing multipurpose methods.

The present invention relates to a method and a system for day-ahead generation schedule establishment of an energy base system. The method comprises the steps of: 1) establishing a day-ahead generation schedule optimization model of an energy base system comprising a direct current regulating measure according to wind power prediction information obtained in advance; 2) establishing a plurality of application examples by employing different types of wind power prediction information as input variables of the day-ahead generation schedule optimization model of the energy base system, and determining an optimal day-ahead generation schedule of the energy base system according to the obtained operation index comparison result of the energy base system and focused operation indexes. The method can integrate the wind power prediction information into the optimal model to perform comparison, analysis and verification, and considers the direct current regulation measure to obtain a better economic performance on the premise of reasonable operation of the system. Therefore, the method provided by the invention can be widely applied to the day-ahead generation schedule establishment of theenergy base system.