68 results about "Algorithmic complexity" patented technology

An INS assistance-based satellite navigation spoofing-type interference resisting method is disclosed. According to the invention, by employing position and speed information provided by INS, spoofing-type interferences are eliminated in two steps. Firstly, invisible satellites are eliminated based on satellite elevation angles; and then, by using satellite ephemeris information and the position and speed information provided by the INS, the pseudo range and pseudo range change rate curve of a real visible satellite and a corresponding receiver are calculated and compared with the satellite resolving pseudo range and pseudo range rate received in a channel, thereby effectively eliminating satellite signals which mismatch conditions. The method of the invention overcomes the defect that conventional satellite navigation spoofing-type interference resisting needs the foreknowledge of signal types and adopts different interference resisting methods for spoofing-type interference signals of different types; and the method of the invention expands the algorithmic applicability and reduces the algorithmic complexity, thereby being more applicable to complex spoofing-type interference environments.

A view spot triggering method based on the explication point in a GPS intelligent guide system comprises: collecting and triggering the data correlated with the explication point, acquiring the position, direction and speed information of the current user by a GPS positioning signal processing module. The triggering steps comprise: searching whether all of the explication spots in the scenic zone meeting the triggering requirement according to the preset triggering cycle and the preset triggering radius, selecting an explication spot from the explication spot assembly based on the explication spot selection regulation, searching the view spot not playing the music and relative to the explication spot, selecting a view spot based on the view spot selection regulation and starting the music playing module. The data quantity or work quantity of the collected or mapped data is small. The view spot triggering data can be vectorized which is better than the mode of triggering the zone in an graphics mode. The explication point triggering mode is influenced by the number of the explication point and the triggering radius values, wherein the data quantity is small, the algorithmic complexity is low and the performance is high.

The embodiments of the invention disclose a collaborative editing, cooperative processing method, apparatus and device, and a storage medium. The collaborative editing method includes the steps that aserver receives a first editing operation for a copy of a target editing content sent by a first client, if the server determines that the first editing operation and received at least one target editing operation form a concurrent operation, converting first editing operation according to editing attribute information of each editing operation in an editing operation sequence matched with the target editing operation and the first editing operation, to obtain a first conversion editing operation; after the server uses the first conversion editing operation to edit the locally stored copy ofthe target editing content, the server adds the first conversion editing operation to the tail of the editing operation sequence and sends the first conversion editing operation to each collaborativeclient, so as to achieve complete support for collaborative document coherence, user intent retention, and collaborative undo functionality, and to reduce algorithmic complexity and improve algorithmperformance.

The invention relates to a webpage text extraction method based on the maximum text density. The method includes the following steps of (1) preprocessing a webpage, processing character codes and standardizing the webpage, (2) analyzing the webpage into a DOM tree and extracting tag text blocks in the webpage according to specific tags, (3) calculating the maximum text density, and (4) extracting texts, carrying out sequencing according to calculated text densities after all the tag text blocks are processed, and selecting a tag with the maximum text density, wherein the tag and content of a nested sub-tag serve as a text block and the text is obtained after the tag is eliminated. The webpage text extraction method based on the maximum text density is low in algorithm complexity, has universality and has a good effect on webpages with complex structures.

The embodiment of the invention relates to a method and a device for processing video data. The method comprises the following steps: preprocessing the video data to obtain a grayscale frameset; performing normalization treatment on the grayscale frameset to obtain a normalized grayscale frameset; calculating a grayscale mean of the normalized grayscale frameset, and determining the density latitude extracted by a column diagraph according to the grayscale mean; and dividing the density latitude into more than one grayscale section, and obtaining a Hash value through comparison of the number of pixels in different grayscale sections. The device comprises a preprocessing module used for preprocessing the video data, a normalization module used for performing normalization treatment on the grayscale frameset, a column diagraph extracting module used for determining the density latitude extracted by the column diagraph, and an acquisition module used for dividing the density latitude into more than one grayscale section and acquiring the Hash value. The method and the device can effectively resist the geometric distortion, and reduces the algorithmic complexity to make algorithm have good real time performance.

The invention provides a detection method for paper money splicing signals. The method comprises the following steps: positioning a detection region of paper money, acquiring image pixel of the detection region; obtaining coverage lines of the detection region; obtaining a pixel gray distribution diagram of the detection region; obtaining and analyzing characteristics of all coverage lines, finding out a line whose black signal and white signal are most obvious, marking the line; and presetting an experience difference value threshold of a genuine note, comparing with a detection result, to obtain a conclusion. Through circumferential coverage, the method can accurately position white abnormal lines and black abnormal lines, and whether a signal is a splicing signal is determined by the characteristic difference of the white abnormal lines and black abnormal lines. The method does not need to process binaryzation and hough transformation, so time complexity is greatly reduced. Since the binaryzation and the hough detection need to calculate each point of an image, but the method just needs to calculate the lines in characteristic number, algorithmic complexity is reduced, identification time of splicing paper money is reduced, and paper money identification efficiency is improved.

The invention provides an infrared-based 3D four-dimensional data acquisition method and device; the method comprises the following steps: preprocessing a 2D color image gathered by one or more zooming color cameras, thus obtaining a 2D high definition color image of a target object; preprocessing an infrared image gathered by an infrared depth camera, thus obtaining depth infrared data of the target object; obtaining 3D point cloud information of each composition point of the target object according to the depth infrared data of the target object; registering and fusing the 2D high definitioncolor image with the 3D point cloud information; obtaining 3D four-dimensional data of the target object according to the 2D high definition color image with the 3D point cloud information fusion result data and the calibration information of the zooming color camera and the infrared depth camera. The method reduces characteristic point extraction, reduces the algorithmic complexity, and improves3D reconstruction precision and efficiency.

The invention belongs to the technical field of information processing, and particularly relates to a method for compensating uneven brightness of a bottom view image under nighttime infrared. Compared with the prior art, the method comprises the following steps of extending an original image by fully utilizing a mirror image method of an image, and eliminating edge effect influence of the image; using an average filter to acquire a background image of an extracted image, and removing texture details of the background image as far as possible; utilizing a difference method to eliminate the phenomenon of uneven brightness of four corners of the image, so that the algorithmic complexity is reduced, and the timeliness of the algorithm is improved; and adopting image contrast enhancement process on the differenced image, so that a gray value distribution range of the image is expanded.

The invention discloses a power allocation method in a cognition orthogonal frequency division multiple access (OFDMA) system based on proportional fairness and interference constraints, comprising the following steps: firstly carrying out power allocation among cognitive users in accordance with proportional fairness factors, total power constraints and permanent speed values; then using a quick water injection method to carry out carrier power allocation by each cognitive user; if the carrier power of each cognitive user does not exceed a power constraint value, ending an algorithm, and otherwise, modifying allocated power of a carrier exceeding the power constraint to be the power constraint value; calculating the speed of the carrier and adding the speed to the permanent speed value of the user; eliminating the carrier from the carrier set of the user; and if the carrier set of each cognitive user is a null set, ending the algorithm, otherwise, subtracting the allocated power constraint value from a total power value, and repeating the above steps until the algorithm ends. The method is better in performances of the system capacity and the proportional fairness among the cognitive users, low in the algorithmic complexity and applied to engineering.

The invention discloses a SoE online estimation method for a variable-working-condition vehicle power battery, and the method comprises the following steps: S1, building a compensation coefficient function of actual available energy, and building an OCV-SoE curve; S2, obtaining a state equation and an observation equation; S3, constructing a difference equation suitable for a recursive least square algorithm with a forgetting factor, identifying system parameters in real time, and obtaining the system parameters; S4, estimating the SoE state value of the vehicle power battery by using the system parameters obtained in the step S3 and combining an H infinity algorithm. According to the method, the recursive least square method with the forgetting factor is combined with the H infinity algorithm to identify and estimate the SoE, the forgetting factor can effectively reduce the influence of error accumulation, and the H infinity algorithm has no fixed hypothesis requirement on noise, so that the estimation is more in line with the requirements of actual working conditions, the estimation precision is higher, and the algorithm complexity is not high.

The invention is a multi-user scheduling method precoded by the Tomlinson-Hasheem, which is characterized in that in downlink multi-antenna multi-user system adopting Tomlinson-Hasheem precoding, in the process of subscriber selection, selecting the user which projects the maximal second order norm in the orthogonal space at the selected user channel space at every time; applying the heuristic Tomlinson-Hasheem method, adequately applying the projecting result of the user space orthogonal space in the selected set at the last user space so as to effectively reduce algorithmic complexity. Adopting the scheduling algorithm provided by the invention, the system throughput performance approaches to the capability of the downlink multi-user multi-antenna system, has progressive optimum performance. The inventive computation complexity is far less than traversal selective complexity and better than complexity of the classical greedy scheduling algorithm and suitable used in the practical system.

The invention relates to the technical field of communication, in particular to an energy efficiency power distribution method based on non-orthogonal multiple access in an MIMO system, which comprises the following steps: taking a user as a buyer and a base station as a seller; constructing the utility optimization model of the buyer and the utility optimization model of the seller; introducing an auxiliary variable to solve the optimal power purchasing strategy of the buyer; solving the optimal price formulating strategy of the seller by utilizing a convex optimization method; obtaining a user power value when the base station and the user play a game to achieve balance; and updating the auxiliary variable to calculate the optimal energy efficiency value of the user so as to obtain the energy efficiency value of the system. According to the method, only the benefits of a single user are considered, each user maximizes the energy efficiency of the user, and the algorithm complexity isreduced; and in the game process of the user and the base station, the power purchase value and price are adjusted according to the serial interference elimination residual factor, and the result iscloser to the optimal value.

The invention discloses a passive cooperative positioning method for predicting fast covariance interactive fusion on the basis of an information theory. The method includes the steps that 1, a mathematical model of a passive cooperative positioning system for two receiving stations and one external radiation source is constructed; 2, the two receiving stations separately perform prediction and updating on the state of a target according to an extended Kalman filtering algorithm; 3, according to an information theory optimization criterion and by taking original fusion state estimation into account, a K-L divergence (Kullback-Leibler Divergence) between error covariances can be obtained, and a calculation method for predicting a fast covariance interactive fusion algorithm weight is proposed; 4, an algorithm for predicting the fast covariance interactive fusion is proposed to obtain state estimation of the target. Under a passive positioning environment of the receiving stations and the external radiation source, the fusion effect can be improved, the target tracking precision is improved, and the complexity of the algorithm is low.

The invention provides a method for identifying boundary nodes in a distributed wireless sensor network. The method is used for identifying the boundary nodes in the sensor network by utilising a DSCS (Distributed Sector Cover Scanning) technology and is applicable to homogeneous or different non-uniform wireless sensor networks same in function, sensing and communication ranges. The method comprises the following steps that: (1), the nodes independently collect neighbour information in the network; (2), each node establishes a cartesian coordinate system by taking self as the circle centre and calculates an absolute angle of a neighbour node thereof; (3), the neighbour nodes are sorted in an ascending manner according to the absolute angles; and (4), the boundary nodes are identified by utilizing DSCS. According to the invention, the technological problem that the identification precision of the boundary nodes in the traditional wireless sensor network is low can be solved; the identification problem of the boundary nodes in the wireless sensor network is decomposed into the sensing sector edge coverage problem of one node; therefore, the algorithm complexity is effectively reduced; the operation time and the inter-node communication and energy consumption in the boundary node identification process are reduced; and the identification precision is improved.

The invention relates to a multi-hypothesis tracking method based on infrared target gray-scale cross-correlation and angle information. Infrared target gray-scale cross-correlation characteristic information and angle information are used, a track confidence evaluation model in the multi-hypothesis tracking method is reconstructed, the track confidence evaluation model based on the gray-scale cross-correlation and angle information is established, the model is used to complete confidence calculation, the rapid confirmation and rapid deletion of multiple target tracks are achieved, the numberof follow-up track hypotheses is reduced, a track creation time is effectively improved, and the algorithm complexity is reduced. Therefore, the track confidence evaluation model based on the infraredtarget gray-scale cross-correlation and angle information can achieve a better effect, and the method of the invention is better than a multi-hypothesis tracking method using an angle information track evaluation model.

The invention belongs to the technical field of Internet-of-Vehicles communication, and discloses an Internet-of-Vehicles information transmission method based on unmanned aerial vehicle assistance. Aiming at the problems in the prior art, an information transmission problem is formalized into a multi-objective optimization problem related to state transition probability and transmission consumption, transmission is carried out by utilizing an opportunity transmission mode of 'storage, carrying and forwarding', the state transition probability is set to simplify the original problem, the algorithm complexity is reduced, and finally, the proposed multi-objective optimization problem is solved on the basis of a Q-learning technology. Compared with other information transmission methods, themethod has the advantages that the network transmission success rate can be effectively improved, the routing overhead, the transmission delay and the transmission hop count are reduced, the algorithmcomplexity is low, the method is more suitable for the urban environment, and the intelligent traffic is realized.

The invention discloses a coordination control method and system for a main-branch path mixing area. The method comprises the steps that the main-branch path mixing area is divided into a plurality of sub-areas; for each sub-area, moving amount between actual intersections and set ideal intersections in the sub-area is obtained based on each ideal intersection separation distance in a plurality of set ideal intersection separation distances, and the best period of the sub-area and the best phase difference of all the intersections in the sub-area are determined based on the moving amount; for each intersection in each sub-area, the intersection predicted saturability of each scheme of a plurality of preset split schemes used in the next period is calculated, and the split scheme corresponding to intersection predicted saturability with the minimum value is determined to be the best split scheme of the intersection. The coordination control method and system for the main-branch path mixing area reduce algorithmic complexity, reduce systemic input quantity, and can conduct coordination control on the area dynamically in real time.

An embodiment of the present invention is directed to a code audit tool that intelligently analyzes and profiles code, such as Python code, based on a variety of previously unmeasured factors and metrics including a set of software dimensions, such as Algorithmic Complexities; Software Sizing Metrics; Anti-Pattern Implementations; Maintainability Metrics; Dependency Mappings; Runtime Metrics; Testing Metrics; and Security Metrics. Once this analysis is complete, a standardized report card or other scoring interface may be generated. This may include analytical findings as well as suggestions and recommend steps so that developers can make informed decisions, enhance their code bases and improve the score assigned to their code.

The invention discloses a long-distance laser radar point cloud data processing method, and the method comprises the steps: scanning a target environment of data to be collected through an optical-mechanical scanning system, collecting the photon flight time of each pixel point, and exporting and extracting the collected data; calculating filtering parameters by using a point cloud data segmentation method to complete parameter self-adaption; performing dimension reduction on the data by applying a principal component analysis technology; and carrying out point cloud clustering by applying a density-based spatial clustering denoising (DBSCAN) algorithm, deleting outliers, namely noise points, and finally raising the dimension of the obtained two-dimensional point cloud to three-dimensional point cloud data, thereby completing denoising and filtering of the original point cloud data. According to the method, the algorithm complexity is effectively reduced while the environment characteristics are reserved.

The invention relates to an in-orbit video image stabilizing method for a video satellite, which solves the problem that the existing video satellite image stabilizing method having complex hardware structure and high algorithmic complexity is unbeneficial to satellite embedded implementation. A video satellite image is stabilized by adopting the methods of phase correlation, image interception and light stream image stabilization. The method comprises the steps of firstly, according to the existing video satellite imaging characteristic, translating two frames of images based on the difference, and judging the image translation quantity by adopting the phase correlation method; then calculating corners of a small intercepted image, and generating global corners corresponding to the original images; and finally, implementing high-precision image stabilization by adopting light stream and an RANSAC (Random Sample Consensus) method. The method avoids the influence of inherent accuracy error of satellite positioning and orientating equipment on image stabilization, and the practicability of the method is verified by adopting a video image transmitted by the video satellite of Changguang Satellite Technology Co., Ltd. in experiments.

The invention relates to a sound source positioning method and device, a medium and equipment. According to the scheme provided by the embodiment, when sound source positioning is carried out, the microphone array can be used for collecting audio time domain signals, after the audio time domain signals are converted into audio frequency domain signals, the trained neural network model is used forscreening frequency points, the frequency points with the signal-to-noise ratio larger than a set value are determined, and then the frequency points with the signal-to-noise ratio larger than the setvalue can be used for carrying out sound source positioning. Compared with the scheme that the frequency points are not screened and all the frequency points are used for sound source positioning inthe prior art, according to the scheme that all frequency points are used for sound source positioning, the frequency points with high signal-to-noise ratio is used for sound source positioning, so that the sound source positioning precision can be effectively improved. Meanwhile, a small number of frequency points are used for sound source positioning, the algorithm complexity of sound source positioning can be effectively reduced, and the real-time performance of sound source positioning is ensured.

The invention discloses a skeleton action recognition method based on learnable PL-GCN and ECLSTM, and relates to the field of action recognition. The problems that in the skeleton action recognition process, the feature capture capacity of key frames and significant motion joints is limited, and the similar action classification capacity is weak can be solved. The method comprises the steps that a learnable graph convolutional network (PL-GCN) is provided for the problem that similar action recognition is prone to confusion, and the learnable graph convolutional network (PL-GCN) is used for improving the physical structure of a model; for the problem of weak key frame capture capability, a feature enhanced long and short time memory network (ECLSTM) is provided for enhancing time sequence features; building a skeleton graph by utilizing a graph topological structure of the skeleton sequence data; fusing the spatial features from the image after convolution and the time sequence features extracted by the ECLSTM network; and carrying out average pooling and convolution on the fused features, and then carrying out final feature classification. The method provided by the invention is superior to some current methods in action recognition progress, algorithm complexity and feature extraction capability.

The invention relates to a 3D beamforming modeling method based on an NR-U / Wi-Fi coexistence system, and belongs to the technical field of communication. According to the modeling method, a channel is decomposed into a horizontal dimension and a vertical dimension for a 5G multi-cell multi-user scene of an unlicensed frequency band. In the horizontal dimension beam forming, interference between adjacent cells is eliminated through prepositive precoding, and interference between users in the cells is eliminated through postpositive precoding. The vertical dimension beam forming enables the base station to more accurately position the user by expanding the dimension of the antenna array, thereby improving the total capacity and the anti-interference capability of the system, effectively reducing the interference between NR-U / Wi-Fi equipment and the algorithm complexity, and further improving the system performance.

The embodiment of the invention discloses a face attention judgment method and device, equipment and a storage medium. The method comprises the following steps: acquiring a face image; processing theface image by using a pre-trained multi-task convolutional neural network, and determining a face frame position, a face orientation and a face posture; determining a coarse face orientation directionand a fine face orientation angle according to the face orientation and the face attitude; and determining a human face attention result according to the human face frame position, the human face orientation and the human face posture. According to the embodiment of the invention, the process of determining the human face orientation and the attention result is combined, and the human face attention concentration result is obtained by determining the human face orientation information, so that the attention judgment efficiency is improved, the judgment time is saved, and the algorithm complexity is reduced.

The invention discloses a method for recovering a monocular SLAM scale through detection and calibration, and the method comprises the steps: calibrating a camera, obtaining a first depth from a pointto the camera on a visual plane by perspective transformation calculation, detecting and setting a Marker, and detecting and setting a second depth from the Marker to the camera through a deep learning module; adding the first depth and the second depth into the SLAM back-end optimization process to obtain the monocular SLAM scale, so that the monocular SLAM scale is recovered, the calculation process is effectively simplified, and the algorithm complexity and related calculation amount are effectively reduced.

The invention discloses a denoising and interpolation united method and system for Bayer images. The method comprises steps as follows: pixels in an original Bayer image are classified into G and BR, and interpolation is performed on G channel and BR channel respectively; data after interpolation are transformed from G domain and BR domain to delta domain and summation domain; horizontal filtering processing is performed on the data after domain transformation; interpolation is performed on the delta layer, a delta_B layer and a delta_R layer are recovered, and meanwhile, the delta layer obtained in the last step is stored as a delta_G layer; and longitudinal filtering processing is performed on the data after interpolation and color recovery is performed to obtain processed data. By means of the denoising and interpolation united method and system for the Bayer images, algorithmic complexity can be reduced greatly, and a large amount of hardware resources can be saved.

The invention relates to a navigation multipath effect inhibition method based on space-time big data; the method comprises the following steps: creating a space-time big data multipath information table, using time and coordinates solved by an original input signal as a two dimension index, and using the corresponding navigation position coordinate offset amount as contents; 2, sharing the space-time big data multipath information table to other navigation receiver users in the table creation area in the step1; 3, loading the space-time big data multipath information table to the local database; 4, searching the space-time big data multipath information table in the local database when the navigation coordinates received by other navigation receiver users have hopping, retrieving the corresponding navigation position coordinate offset amount and correcting same, thus obtaining a positioning result. The method can inhibit the multipath effect according to the sharing ideas, and can reduce the multipath inhibition algorithmic complexity in the navigation digit intermediate frequency signals.

The invention discloses an automobile anticollision intelligent early warning system based on millimeter-wave radars. The automobile anticollision intelligent early warning system based on the millimeter-wave radars comprises an automobile body, the millimeter-wave radars, a Beidou satellite positioning system, an inertial navigation system, automobile door trigger switches, automobile lock control switches and a main control module, wherein the millimeter-wave radars, the Beidou satellite positioning system, the inertial navigation system, the automobile door trigger switches, the automobilelock control switches and the main control module are arranged on the automobile body; the millimeter-wave radars are uniformly installed on the periphery of the automobile body; the Beidou satellitepositioning system, the inertial navigation system and the millimeter-wave radars are correspondingly and electrically connected with a signal amplifier, a demodulator and the main control module in sequence; the automobile door trigger switches are installed on opening handles on the inner parts of four automobile doors of the automobile body and are connected with the main control module; and the automobile lock control switches are installed at hinge joint positions of the four automobile doors of the automobile body and the automobile body to restrict the opening and closing of the automobile doors. According to the automobile anticollision intelligent early warning system based on the millimeter-wave radars, the conception is original, the detection reliability is high, the algorithmic complexity and the equipment cost of the complete set of the system are reduced greatly, the detection accuracy is improved greatly on year-on-year basis, and the safety of driving is improved.