40results about How to "Optimize calculation method" patented technology

The invention discloses a method for autonomously localizing robots on the basis of laser radar. The method includes randomly generating N particles to form particle swarms around initial locations ofthe robots, and updating the particle swarms according to robot real-time movement distances and real-time rotation angles measured by sensors of the robots at current operation moments of the robots; computing the superposition quantity of point cloud of the laser radar and obstacles of maps for each particle to use the superposition quantity as a score of the particle, computing weighted position and posture average values of the particle swarms by the aid of the score, which is used as a weight, of each particle and utilizing the weighted position and posture average values as AMCL (adaptive Monte Carlo localization) estimation positions and posture; utilizing the AMCL estimation positions and posture as initial values, acquiring scanned and matched positions and posture by the aid ofscanning and matching algorithms on the basis of Gauss-Newton iterative processes and utilizing the scanned and matched positions and posture as the optimal positions and posture of the robots at thecurrent operation moments; re-sampling the particle swarms by the aid of AMCL algorithms to ultimately obtain the global optimal positions and posture of the robots during operation. The global optimal positions and posture of the robots are used as localization results. The method has the advantage that the localization convergence rate can be greatly increased, and the localization precision andthe localization stability can be greatly enhanced.

The invention relates to an image semantic segmentation method based on a local area conditional random field model. A full convolution neural network structure of the invention extracts input picturefeatures and obtains a rough segmentation result. The region selection structure filters the edge of the segmentation result map, and selects the segmentation result as the largest circumscribed rectangle of the pedestrian, bicycle and motor vehicle parts. The local region conditional random field model establishes the conditional random field model in the rectangular region and refines the segmentation result of the rectangular region. The invention effectively combines the advantages of the precision of the conditional random field model with the advantages of the speed of the full convolution neural network. The computational method of the conditional random field model is optimized so that the time complexity of the model is greatly reduced. The segmentation accuracy of the traditional full convolution neural network is improved. The application of probability graph model and full convolution neural network is designed as an end-to-end system.

The invention discloses a real-time data calculation method. The real-time data calculation method comprises the following steps: when a configuration instruction sent by a configuration end is received, analyzing the configuration instruction to obtain configuration information; generating calculation task is generated according to the configuration information, calculating a data source according to the calculation task, and storing a calculation result in a preset Hbase database; When a data real-time calculation request sent by a query end is received, obtaining corresponding target indexinformation and a target time range according to the data real-time calculation request; According to the target index information and the target time range, obtaining corresponding index data from the preset Hbase database according to a preset rule, and obtaining a real-time value through calculation according to the index data. The invention further discloses a data real-time calculation deviceand equipment and a computer readable storage medium. According to the invention, the development cost can be reduced, and the code reuse rate is improved.

The invention relates to an optical character recognition method based on an LVQ (Learning Vector Quantization) neural network. The method includes the following steps: performing preprocessing including image denoising, character segmentation, binaryzation and characteristic extraction on a to-be-recognized character image; sending a single character characteristic into an input layer of the LVQ neural network and acquiring a winning competition-layer nerve cell according to a competition algorithm and performing a recognition rejection test; and inputting a character characteristic which complies with a recognition condition to the competition layer of the LVQ neural network for recognition and outputting the recognition result. The optical character recognition method based on the LVQ neural network is applicable to the field of automatic recognition of a small character set which has strict demands on false rate.

The invention relates to an object name editing distance calculating method and an object name editing distance matching method based on information entropy. The object name editing distance calculating method comprises the following steps of (10) counting the using frequency of each character and the total number of object names and defining that the character which is frequently used in an object name only appears in the object name once; (20) calculating the information entropy of characters according to the ratio of the total number of the object names to the using frequency of the characters to obtain editing cost of the characters; and (30) enabling the editing cost of inserting or deleting of a character to be equal to the editing cost of the character when editing distances of the object names are calculated and performing substitution operation under the condition that the editing cost of substitution is zero if two characters are the same and the editing cost of the substitution is the sum of the two characters if the characters are different. The invention also provides the corresponding matching method. By the object name editing distance calculating method and the object name editing distance matching method based on the information entropy, the absolute difference between two object name character strings can be reflected accurately, the similarity between two object names can be recognized effectively, and an effect of handling the problem of matching of data such as names is high.

The invention provides a method for determining bearing capacity characteristic values of tamped/compacted fill strata by using the shear wave velocity. The method for determining the bearing capacitycharacteristic values of the tamped/compacted fill strata specifically includes the steps that the fill strata of a dynamic compaction (compacted) foundation are classified by using first stone backfill materials with different particle sizes, then the bearing capacity characteristic values of the fill strata calculated by adopting shear strength indexes (c and phi) of a direct shear test on sitecorrespondingly correspond to the shear wave velocity in a borehole to establish the relationship between the bearing capacity characteristic values and the shear wave velocity, and through the established correspondence, the accurate and fast method for determining the bearing capacity characteristic values of tamped/compacted foundation soil layers is provided in addition to a load test method.An accurate basis for valuing of the bearing capacity characteristic values is provided for valuing and evaluation of a tamped foundation and a compacted foundation in geotechnical engineering investigation.

The invention discloses a downlink signal sending method and downlink signal sending equipment in a multiaerial system. The downlink signal sending method mainly comprises the following contents: obtaining a downlink channel submatrix required by a terminal by virtue of receiving pilot signal information sent by the terminal; determining a downlink channel matrix of an aerial system and obtaining a noise power matrix according to noise power fed back by each terminal; obtaining a precoding submatrix by use of the downlink channel matrix and the noise power matrix, and determining the precoding matrix of the downlink channel, so that the influence of interference and noises among the terminals on the system performance is taken into consideration in a precoding matrix acquisition process; obtaining a precoding submatrix C according to the obtained downlink channel matrix H and noise power matrix X. According to the method and the equipment, the computation complexity is simplified, the precoding matrix computing mode in the prior art is improved, meanwhile, the precoding matrix of antennae is determined by virtue of considering the factor that different terminals are influenced by noises, and the downlink signal power of the antennae is optimized.

The invention discloses a soil heavy metal detection value correction method, device and a computer storage medium. The soil heavy metal detection value correction method includes the steps that factors affecting the detection value of a soil heavy metal rapid detection instrument are determined; for soil samples used for calibration modeling, by taking the soil heavy metal detection value detected by a laboratory analysis test method as an accurate value and taking the soil heavy metal rapid detection instrument detection value as a correction value, a mixed linear correction model is built;and model parameters of the mixed linear calibration model are estimated and verified, and the verified mixed linear calibration model is used as a soil heavy metal detection value correction model detected by the soil heavy metal rapid detection instrument in the future. The soil heavy metal detection value correction method, device and the computer storage medium are used for improving the detection accuracy of the soil heavy metal content of the soil heavy metal rapid detection instrument.

The invention discloses a condition slicing method based on golden section search and software execution trace. Based on a given first-order predicate logic set, in an interval matching conditions, a variable value is selected as input of dynamic program operation based on a golden section search method, a corresponding statement coverage set is obtained by calculation through an execution trace set of program multi-time operation, and on the basis, combined with static slicing of a program, condition slicing of the program is obtained by calculation. According to the method, global analysis of the program is not needed, thus the efficiency is improved, and the space complexity is low; besides, it is convenient to acquire the execution trace of the program, and acquisition of the execution trace of the program is not influenced by program types, so that the method is applicable to interprocess and concurrent programs and is object-oriented.

The invention discloses an automatic proportioning flavoring machine which comprises a setting device for presetting seasoning proportion, a control device for communicative connection with the setting device, a weighing device connected with the control device and for weighing a utensil below a discharging port and sending weight data to the control device, and a discharging device provided witha plurality of driving parts for driving discharging. The control device sends the obtained weight data to the setting device, the setting device obtains the weight data and then calculates driving parameters of the driving parts according to preset seasoning proportion, and the control device obtains the driving parameters and controls the driving parts to work. The automatic proportioning flavoring machine can dynamically adjust seasonings according to the weights of food materials, thereby improving the accuracy of seasoning proportioning.

The embodiment of the invention discloses a vehicle optimal efficiency obtaining method and system, and the method comprises the steps: building a whole vehicle stress model; sampling the vehicle speed of the whole vehicle to obtain feedback data, and obtaining driving resistance equivalent coefficients A, B and C according to the stress model of the whole vehicle; determining the required torqueof the whole vehicle according to the driving resistance equivalent coefficients A, B and C and the mass of the whole vehicle; screening the optimal torque output by the motor high-efficiency point bysetting a torque interval, and achieving the optimal efficiency of the whole vehicle. According to the scheme, the whole vehicle mathematical model is established, the whole vehicle stress equivalentcoefficient is used for closed-loop adjustment, and various working conditions in whole vehicle operation are met. And the calculation efficiency is improved by optimizing a calculation mode, lookingup a table, obtaining a difference value and a weight, calibrating in advance and the like.

An earth and rockfill dam complex borrow pit excavation volume calculation method based on a grid subdivision algorithm comprises the following steps that three-dimensional point cloud data of a borrow pit excavation face are scanned, registration of point clouds of different scanning batches is achieved through a point cloud registration algorithm, so that point clouds which are originally and irregularly distributed on the excavation face are obtained; the obtained original irregularly distributed point cloud is processed through a Doo-Sabin subdivision algorithm, and the point cloud forms a relatively regular excavation surface; then, Kriging interpolation calculation is performed to obtain a spatial point cloud participating in excavation volume calculation; a grid subdivision algorithm is compiled through MATLAB, and calculation of the excavation volume is achieved. According to the earth and rockfill dam complex borrow pit excavation volume calculation method based on the grid subdivision algorithm, the borrow pit excavation volume can be rapidly estimated, and the method has the advantages of being small in workload, relatively safe in measurement and high in efficiency.

The invention discloses a book similarity calculation method based on random walk, electronic equipment and a storage medium, and the method comprises the steps: obtaining user interaction behavior data for a book; determining an interactive book sequence corresponding to each user according to the user interaction behavior data; constructing a book association graph according to the interactive book sequence corresponding to each user; performing random walk calculation according to the book association graph to obtain a book similarity matrix of each book relative to other books. According to user interaction behavior data for books, the book association graph can be conveniently and rapidly constructed. Random walk calculation is made according to the book association graph, the book similarity matrix of each book relative to other books is rapidly obtained, the similarity between the books is accurately and effectively reflected from the perspective of a user, the calculation accuracy of the book similarity is effectively improved, and the calculation mode of the book similarity is optimized.

The embodiment of the invention discloses a risk rating method and device for security vulnerabilities of an information system, equipment and a storage medium. The risk rating method for the security vulnerabilities of the information system comprises the following steps: determining vulnerability types according to positions of to-be-rated vulnerabilities; determining a corresponding risk rating index and content according to the vulnerability type, wherein the risk rating indexes comprise a vulnerability occurrence probability index, a vulnerability hazard level index and a vulnerability utilization difficulty index; quantifying the content of each risk rating index to obtain a quantized value of each risk rating index; and according to the quantized value of each risk rating index and a preset weight ratio among the vulnerability occurrence probability index, the vulnerability hazard level index and the vulnerability utilization difficulty index, determining the risk level of the to-be-rated vulnerability. Through the method, the security vulnerabilities of the information system can be graded quickly.

The invention relates to the technical field of operation and maintenance and monitoring on a substation DC system, specifically, relates to a method for determining battery performance by multi-parameter weighting, comprising the following steps of : 1 collecting the cell voltage of each cell in real time and recording it once every hour; 2 automatically performing an internal resistance test every month, and recording the internal resistance and discharge voltage of each battery in the internal resistance test. The method for determining battery performance by multi-parameter weighting assigns weights according to categories, and then calculates the integrated value of each battery. The battery with the highest combined value is the worst performing battery. The calculation method is simple and accurate.

The invention discloses a map extension layout method, and relates to the field of data visualization analysis. The method comprises the following steps: acquiring an existing map boundary, forming an ordered coordinate set of the boundary, performing node positive sequence arrangement according to the distance between an extension node and a node coordinate in the ordered coordinate set to obtain an ordering result, calculating a coverage circle and an original position of the coverage circle for a newly-added node which is not in an existing map and is associated with the extension node, obtaining a moving route of the coverage circle in the existing map according to the sorting result, determining a layout positioning circle by moving the coverage circle according to the moving route, and adjusting the layout positioning circle to complete layout extension of the map.