81 results about "Constraint algorithm" patented technology

The invention relates to an image-based plant three-dimensional shape measurement and reconstruction method and a system, and the method comprises the steps of obtaining two images of a target objectin different positions in the vertical direction, obtaining internal parameters of a shooting source, extracting two-dimensional profile information or two-dimensional framework information of the target object from the images, carrying out feature matching on a profile or a framework through the epipolar constraint algorithm, calculating data of a three-dimensional shape structure of the target object according to the internal parameters and the result of the feature matching, and finally obtaining a three-dimensional model of the target object by using spline curves and curved surfaces for fitting and generating the three-dimensional model of the profile or the framework. The method and the system design the corresponding data acquisition method according to the rapid, accurate and non-destructive data acquisition demands of main crops of agriculture by facing to three-dimensional shape acquisition and reconstruction of the main crops, as well as measurement and calculation of main agronomic parameters based thereon, thereby having an important practical significance and broad application prospects for developing practical tools.

The invention relates to a commodity distribution path planning method based on an MoCD algorithm (multi-object constraint algorithm based on Dijkstra algorithm), which includes the following steps of: firstly taking intersections or endpoints of roads as nodes and building a model of an urban road network according to actual urban road conditions; secondly determining commodity price information, and screening manufacturers which accord with commodity sending destinations; then selecting the manufacturers according to the required commodity number and the manufacturer stock quantity, and a supply and demand scheme which has multiple manufacturer combinations so as to satisfy required commodity quantity existing; and finally planning an optimal path by using the Dijkstra algorithm based on manufacture geographic positions and vehicle positions determined in the last step, and using a step by step search method to improve the search efficiency of the invention algorithm. The transport cost is saved, and the vehicle loading rate and the distribution efficiency are improved by planning the optimal path.

The invention discloses a structural parameter optimization method for a six-degree-of-freedom parallel platform. The structural parameter optimization method comprises the steps that firstly, the working space and the dexterity of the parallel platform are analyzed on the basis of a multi-target-multi-constraint algorithm; secondly, the constraint conditions for platform parameter optimization are analyzed by taking the working space size index and the dexterity index of the parallel platform as targets; and lastly, the process of the parameter optimization method is designed on the basis of a Monte Carlo method, and the optimal structural parameters, integrating the working space size with the dexterity, of the six-degree-of-freedom parallel platform are obtained. According to the method, the working space size and the dexterity of the six-degree-of-freedom parallel platform can be simply and intuitively described; the constraint conditions of the six-degree-of-freedom parallel platform in the structural parameter selection and optimization process are determined, the multi-target-multi-constraint algorithm for parameter optimization of the six-degree-of-freedom parallel platform is proposed, and structural parameter optimization can be conducted; the process of conducting structural parameter optimization through the Monte Carlo method is given, and the optimal structural parameters, integrating the working space with the dexterity, of the six-degree-of-freedom parallel platform can be obtained.

The invention discloses a two-dimensional-control reflecting mirror and a control method thereof, and a laser scanner including the same. Three piezoelectric driving structures support the reflectingmirror and the supporting points arranged in a positive triangle mode. Problems of poor stability and serious inter-shaft interference of the existing non-symmetric supporting structure of the reflecting mirror in the prior art are solved; and thus the reflecting mirror supporting structure is stabilized, thereby improving the control precision of the two-dimensional-control reflecting mirror. Besides, because of constraining algorithm controlling on the three piezoelectric driving structures, the displacement of the reflection point location in a longitudinal direction during the deflection process of the reflecting mirror becomes zero, so that the stability of the geometric model of the optical feedback path is guaranteed and thus the control precision of the two-dimensional control reflecting mirror and the scanning laser pointing control precision are effectively guaranteed.

The invention discloses a coal-miner posture detection system and method based on binocular vision and inertial navigation. The coal-miner posture detection system comprises a strap-down inertial navigation system, a laser target, a camera and a laser orientation instrument, wherein the laser orientation instrument is positioned on a coal-mining working face. The coal-miner posture detection method comprises the following steps: by utilizing the laser orientation instrument, transmitting cross laser onto the laser target; utilizing the two cameras to carry out real-time shooting on the laser target, and based on the binocular vision principle, tracking a cross laser point on the laser target; adopting an epipolar constraint algorithm based on sequential consistency to carry out three-dimensional matching on a mark point and carrying out real-time calculation on the posture of a coal miner; then by the strap-down inertial navigation system, obtaining real-time angular velocity and acceleration of a body of the coal miner, calculating a real-time posture angle of the coal miner, utilizing Kalman filtering to realize fusion of inertial navigation and vision information so as to obtainthe accurate posture angle of the coal miner. The coal-miner posture detection system and method disclosed by the invention have the beneficial effects that the measurement error of the posture angleof the coal miner can be effectively reduced and the correction effect for the positioning error of the posture of the coal miner can be improved.

The invention discloses a fine-grained vehicle image retrieval method and device based on multi-scale constraints, and the method comprises the steps: carrying out multi-scale region information labeling of a first training data set, and obtaining a second training data set; processing the second training data set based on an improved bounding box constraint algorithm and a Helen constraint algorithm to obtain a third training data set; training parameter values of target parameters of a fine-grained neural network model by adopting the third training data set to obtain a trained fine-grainedneural network model; and inputting a to-be-identified vehicle image into the trained fine-grained neural network model for identification to obtain a target vehicle image of the same type as the to-be-identified vehicle image. Compared with the prior art, the workload of pre-labeling the image category or selecting the box to label the object position is reduced, the cost is saved, and the efficiency is improved.

The invention relates to a novel wireless positioning method by adopting distance geometry under the NLOS environment. On the basis of TOA measuring values, the method provides two kinds of algorithms based on the distance geometry theory under the honeycomb cell environment on two-dimensional space, and the two algorithms include the filtering algorithm and the constraint algorithm. Firstly, the distance geometry relationships between base stations and between the base stations and a mobile station are utilized for constructing a constraint function, the filtering algorithm conducts position estimation by selecting a set of measuring values with the relatively small error from a plurality of sets of measuring values, and the constraint algorithm conducts nonlinearly constrained optimization on the obtained measuring values to a certain extent and then conducts position estimation. It is guaranteed that under the situation that non-sight-distance measuring values are not identified and removed in advance, the non-sight-distance error can be effectively restrained, and the positioning accuracy can be improved.

The invention discloses an SCR denitrification control method based on variable-constraint multi-model prediction control. The method is implemented by combining a multi-model prediction control algorithm, an intelligent model weighing algorithm and variable constraint conditions. The method comprises the following steps that 1, a precise NOx prediction model is obtained through the intelligent model weighing algorithm; 2, the optimal real-time ammonia demand is calculated by beans of continuous rolling optimization of the multi-model prediction control algorithm and the constraint conditions changing by following the working conditions. According to the method, the conflict generated between the model precision and the control algorithm when an existing prediction control strategy is applied in an SCR denitrification system is solved; compared with a traditional simple multi-linear-model linear weighing method, the model precision of the algorithm is improved; traditional immobile constraint conditions are changed into the variable constraint conditions changing by following the working conditions, constraint adjustment is more accurate, therefore, the computing burden of a resolution control instruction is reduced, a project is easy to implement, the large lag characteristic of the SCR system is overcome, and the NOx concentration regulation quality is significantly improved.

The invention discloses a multi-objective optimization method of reverse osmosis deboronization seawater desalination system considering environmental influence, based on the precise model of desalination mechanism and the superstructure model of reverse osmosis system, considering the changes of pressure, concentration and flow of brine along the axial direction of pressure vessel, and the increase of salinity caused by the mixing of brine in power exchanger, the constraint of operation conditions is added to ensure the safe operation of the system, and the multi-objective optimization problem with the cost of water production, energy consumption and environmental impact (carbon dioxide emission) as the objective function is obtained. Constraint algorithm is used to solve the problem, andthe optimal solution of different objective function weights is obtained by fuzzy sorting method. A system optimization scheme obtain by the invention can effectively reduce the cost of seawater desalination, the energy consumption and the emission of greenhouse gas carbon dioxide, so that the water produce by the system meets the standard of boron content in drinking water, has important significance for energy conservation and emission reduction, and has a very good application prospect.

The embodiment of the invention discloses a method, device and equipment for carrying out three-dimensional reconstruction on a two-dimensional image and a storage medium, and belongs to the technicalfield of image processing. The method comprises steps of obtaining a two-dimensional picture set; performing feature point extraction on a single picture in the two-dimensional picture set to obtaina feature point set; obtaining the size of the picture and the position information of the feature points in the feature point set in the corresponding picture; acquiring position information of the feature points in a world coordinate system based on an OpenCV camera calibration method; based on an LSD straight line detection method and the position information of the feature points in the worldcoordinate system, obtaining a horizontal line segment in the picture; constructing a three-dimensional grid chart corresponding to the picture based on a constraint Delaunay algorithm; and combiningthe three-dimensional grid diagrams constructed by the pictures in the two-dimensional picture set to obtain a three-dimensional surface model of the object, thereby completing three-dimensional reconstruction. According to the invention, three-dimensional simulation reconstruction can be carried out on the two-dimensional picture rapidly and accurately; and the manpower and material resource consumption during three-dimensional reconstruction can be reduced.

The invention discloses a community discovery algorithm in an enterprise map based on relationship weights. The community discovery algorithm comprises the following steps: (1) weights of relationships in the enterprise map are determined; (2) the similarity between enterprises are determined; (3) Modularity is used to evaluate the results of community discovery in complex networks; (4) convergingis carried out based on modularity constraint algorithm ; (5) Confidence probability table is calculated based on LPA algorithm. By discovering the communities in the enterprise map, the algorithm divides the entire enterprise map into sub-graphs. The divided enterprise map can not only help to analyze the industry prospects according to its community attributes, but also can be used to analyze the financial guarantee information, cooperation information, investment information and so on. And according to community attributes, the government or relevant departments are helped to formulate policies to adjust the market in a timely manner.

Provided is a satellite navigation deception interference suppression method based on deception interference DOA estimation. The method comprises the following sequentially performed steps that firstly radio frequency signals received by array antennas are down-converted into digital intermediate-frequency signals; influence of noise on estimation of coming direction of deception interference is reduced by utilizing the periodic repetition characteristic of C / A codes, and then a CLEAN method is adopted to estimate coming direction of multiple deception interference sources and the number of the deception interference sources is judged; and finally a linearly constrained minimum variance (LCMV) algorithm is utilized to obtain antenna beams when deception interference coming direction constrains are null, and antenna null is broadened by adopting a first-order differential constraint algorithm. Multiple forwarding sources can be effectively suppressed so that the method is easy to realize and small in computation amount and can act as an independent anti-deception interference module to be embedded in a common receiver.

The invention discloses a multi-constraint wait-free hybrid running water scheduling modeling method based on an idle matrix. The multi-constraint wait-free hybrid running water scheduling modeling method comprises the following steps: firstly, carrying out constraint preprocessing: for a constraint which exhibits processing unit skipping, setting the processing time of a task on a unit as 0, adding a big penalty term when a target function value is calculated for the constraint which exhibits preorders or conversion forbidding, and adding the preparation time after previous task processing time for preparation time which exhibits sequential dependency; secondly, for a stage with a parallel machine, processing by adopting a soft compression regulation method based on a Gantt chart to construct the idle matrix; and finally, on the basis of the idle matrix, constructing a scheduling model according a scheduling target. A wait-free hybrid running water scheduling problem with the complex constraint is converted into a common wait-free running water scheduling problem without the constraint, and the scheduling model is established on the basis of the idle matrix so as to lower the complexity and the solving difficulty of the model.

A method for defining trajectories of fibers on a layup surface for producing at least one ply having a given theoretical orientation, by laying up fibers, including the steps of providing the definition of the constraint curves and / or the definition of at least one constraint grid with association of at least one constraint vector to each node of the constraint grid, the direction of a fiber at an analysis point of the la up surface, being obtained by calculation of the normalized weights of the constraint vectors of the constraint curves and / or by calculation by of the normalized weights of the constraint vectors of the constraint grid, and by weighting by the normalized weights of the constraint vectors.

The invention provides an image segmentation method based on correlation matrix self-learning and an explicit rank constraint. Low rank representation, similarity learning and a cluster structure constraint are incorporated into the same frame and are applied to medical CT image segmentation. The method comprises the steps that (1) a CT original image is preprocessed; the gray scale is used to correct and enhance the image; histogram equalization is carried out; and a median filter method is used to filter a target area to reduce interference and noise in the image; (2) an algorithm based on correlation matrix self-learning and the explicit rank constraint is used to solve the eigenvalue projection matrix of the image; and (3) a traditional clustering algorithm is used to cluster and split a projection matrix, and a region of interest is calibrated.

The invention discloses a magnetic resonance imaging simulating method for irrelevant movement in a myocardial microcirculation perfusion voxel, and belongs to the field of magnetic resonance imaging computer simulating. The method solves the problems that an in vivo experiment qualitative evaluation method cannot precisely quantize the detection effect, and a traditional simulating experiment quantitative evaluation method is large in evaluation difficulty, high in evaluation cost and long in evaluation period. The simulating method comprises the steps that firstly, a network avoidance algorithm, a boundary avoidance algorithm and a fluid branch constraint algorithm are utilized for building a virtual myocardial microcirculation network model; secondly, a blood perfusion model in a blood vessel and a water molecule diffusion movement model outside the blood vessel are built; thirdly, on the basis of the diffusion magnetic resonance imaging principle, an IVIM MRI mechanism is simulated, and a magnetic resonance fading signal is generated; fourthly, nonlinear fitting is performed on the fading signal, and the simulating detection result of the myocardial microcirculation perfusion model is obtained. The method can provide the reliable simulation quantitative evaluation conclusion.

The embodiment of the invention provides a background color generation method and device and electronic equipment, and belongs to the technical field of data processing. The background color generation method comprises the steps: obtaining a color feature matrix of a target image part region on a preset region; performing clustering processing on the color feature matrix to obtain a clustering matrix; determining RGB color values representing the target image color based on a plurality of clustering centers in the clustering matrix; and converting the RGB color value into an HSV channel valuethrough a constraint algorithm, wherein the HSV channel value is used as a background color of the target image correlation region. Through the processing scheme of the invention, the background colorcan be matched with the display content.

The invention relates to an automatic carotid artery internal and external wall segmentation method based on shape prior and similarity constraint. The segmentation method comprises the following steps: 1) detecting a carotid artery lumen through an SVM method to obtain a position where the lumen is; 2) segmenting an internal wall by using an anisotropic segmentation method; 3) learning a regression model in combination of the features of an external wall based on the segmented internal wall, estimating the approximate position of the external wall and constructing a probability model; 4) segmenting the external wall by using a similarity constraint algorithm. The invention provides the automatic carotid artery internal and external wall segmentation method based on the shape prior and similarity constraint, which adopts an automatic mode, effectively realizes the internal and external wall segmentation and is higher in accuracy.

The invention discloses a network planning and optimizing method based on bandwidth reservation and a device. The network planning and optimizing method based on bandwidth reservation comprises the steps of obtaining the bandwidth required by a service of each pixel area within the covering area of each base station of a network to be planned, wherein each pixel area is a small area planned for each base station of the network to be planned; calculating the reserved bandwidth of each base station, wherein the reserved bandwidth of each base station is obtained by equivalent calculation of the bandwidth required by the service of each pixel area within the covering area of a neighbor base station of the base station in proportion; using the calculated reserved bandwidth as a disturbing influence index of a corresponding base station, and carrying out network planning calculation according to the disturbing influence index, a preset planning algorithm and a constraint algorithm to accomplish the planning of base stations in the network to be planned. According to the network planning and optimizing method based on bandwidth reservation and the device, due to the fact that the reserved bandwidth is calculated to simulate the disturbing influence, network planning and optimization can be achieved accurately.

This invention provides the correct Viterbi decode traceback starting index is obtained for all constraint lengths and frame sizes. Reverse transpose operations that depend on the last active add-compare-select unit a cascade block of the state metric update process. This last active add-compare-select unit controls selection of T counter signals used in the decode.

A method for generating noncontact measurement feature points solves measurement errors caused by pasting feature points in the noncontact measurement. The method comprises steps of: 1, projecting a first laser spot array onto a first measurement area on the surface of an object to be measured, wherein two laser spot arrays mutually contain three or more non-collinear laser spot feature points; 2, using the position, the size and the direction information of each laser spot feature point as a unique identification mark in plurality of imaging view angle; 3, deleting, by using a circular constraint algorithm, all other illegal feature points of non-laser spot feature points in the result of the laser spot feature point algorithm extracted in the step 2, and completing identical point matching of a laser spot center by means of the feature matching of an SIFT algorithm; and 4, repeating the steps 1 to 3 to complete the generation of measurement feature points in the measurement area each time. The method is simple in feature point generation and easy to operate. The feature points and the actual points are small low in errors and have density matching the laser light spot array.

The invention discloses a vehicle fine-grained recognition method and device based on combination of a multiple attention mechanism and regional feature constraints, and the method comprises the steps: constructing a multiple attention convolutional neural network model, and carrying out pre-training of the multiple attention convolutional neural network model; performing multi-scale region information labeling on the training data set to obtain a first training data set; processing the first training data set based on an improved bounding box constraint algorithm and a Helen constraint algorithm to obtain a second training data set; and training parameter values of target parameters of a multi-attention convolutional neural network model by adopting the second training data set to obtainthe trained multi-attention convolutional neural network model. Compared with the prior art, the workload of pre-labeling the image category or selecting the box to label the object position is reduced, the cost is saved, and the efficiency is improved.

The invention discloses a park integrated energy system dynamic planning method and device considering multiple uncertainties and a medium. The method comprises the steps of determining an energy hub framework of a park integrated energy system, and determining power balance constraints of all energy networks and device condition constraints of all coupling devices; establishing a whole-life-cycle cost model for the park integrated energy system, wherein the whole-life-cycle cost model comprises a staged dynamic planning equipment investment cost part and a whole-life-cycle operation cost part; taking the minimum cost of the whole life cycle as an objective function, and constructing an optimization model based on the determined constraint condition; decomposing the optimization model into a main problem and a sub-problem by adopting a column sum constraint algorithm, and converting the sub-problem by adopting a large M method and a dual theory; and solving through a main-sub problem iterative algorithm to obtain equipment capacity configuration of each planning stage and an output configuration scheme of each piece of equipment at each moment in the operation stage. The invention is high in economic benefit, and theoretical support is provided for development of a comprehensive energy system.

The invention discloses a rail transit vehicle axle counter magnetic field EMC semi-physical simulation method, which is characterized in that an electromagnetic field constraint algorithm of an axle counter is adopted, a simulation model of a magnetic induction line of the axle counter is calculated by using a magnetic field line of the axle counter according to related conditions of a steady-current electromagnetic field, and connection simulation is synchronously implemented through a hardware I / O module and the axle counter; according to the semi-physical simulation method, the simulation data and the detection data are synchronously compared in the single semi-physical simulation process, so that the semi-physical simulation data are obtained, and the data are uploaded to the SQL database of the cloud through the network to be processed and exported. According to the method, the deviation in the simulation operation process of the axle counter can be reduced, the simulation data of a part of real objects are combined, so that the data are more accurate in the test process, simulink simulation belongs to off-line simulation, the simulation process is kept to have good and stable timeliness, and the reliability of the data is maintained.

The invention provides a control method for uplink resources in an LTE (Long Term Evolution) network. The method comprises the following steps: determining the weight parameter of the transmission rate of each user according to a preset user permission and a transmission demand; grouping each user according to positions through a clustering algorithm to obtain a plurality of user groups; performing mathematic modeling according to the weight parameter of the transmission rate of each user and the user groups by using a single-target multi-constraint algorithm to obtain a mathematic model; solving the mathematic model by using a preset maximum flow algorithm, and performing uplink resource control on each user. The method provided by the invention is low in calculation complexity, and is more accurate to uplink resource control of the users.

The invention discloses a community division method, which comprises the following steps that: receiving the information of multiple uses to be divided, and establishing a topological structure diagram according to the information of multiple uses to be divided; calculating the degree of each node in the topological structure diagram, taking the node with a highest degree in the topological structure diagram as a seed node, and obtaining the neighbor node of the seed node; according to the neighborhood attribute of each neighbor node, calculating the input vigor proportion of the seed node for each neighbor node; according to the input vigor proportion of the seed node for each neighbor node, through a grid constraint algorithm of a structural hole, obtaining the constraint factor of the seed node for each neighbor node; and dividing the neighbor node of which the constraint factor is greater than a preset relationship threshold value and the seed node into the same association, and deleting the neighbor node of which the constraint factor is greater than the preset relationship threshold value and the seed node from the topological structure diagram to improve community division quality.

The invention relates to a low energy consumption underwater acoustic data compression and reconstruction method. An M*N-dimensional matrix randomly formed by elements 0 and 1 is used as a sampling matrix Phi to compress underwater acoustic data, and then an N*N dictionary matrix Psi is built by combining (Daubechies, Db) Db-20 wavelet filter. In the method, a wavelet dictionary matrix is used forsparsity expression of the underwater acoustic signal for the first time, and a sparse solution in a frame with sparsity expression built is resolved by using an AL0 norm constraint algorithm; and aresult of the sparse solution and the dictionary matrix Psi are used, so as to recover the compressed underwater acoustic signal finally. The method has obvious implementation effect in compression and recovery simulation of a sweep signal, and converts signals with non-sparse time frequency effectively, and based on the method, a corresponding optimizing algorithm is designed and is applied to compression and recovery of non-sparse signals for the first time. The defect that the compressed sensing algorithm cannot be used directly at present is overcome.

The invention provides a cell tracking method based on correction by cellular partition combined with multi-frames and original image feedback. The method comprises the following steps of: segmenting cell sequence images, tracking inert cells by a regional overlap method, carrying out cellular partition for each frame of image, and building a cellular information storage center; applying a topological constraint algorithm introducing a Delaunay triangulation network combined with a cellular region to match the cells, storing a cell matching result into a temporary-storage track, then identifying the types of the cells in the temporary-storage track, and correcting the classified cells by a multi-frames and original images feedback method; and finally, updating the cellular information storage center and the storage information of the inner parts of the cells, and outputting a result. The method uses the multi-frames feedback and original images feedback method to correct the matching result and segmentation faults, and therefore the accuracy rate of a cell tracking system is further increased and meanwhile the higher efficiency can be realized.