51results about How to "The description is valid" patented technology

The invention belongs to the technical field of the pattern identification and image processing, and relates to a license plate detection and identification method based on the maximum stable extremal region, which comprises the following steps: extracting the maximum stable extremal region (MSER) to obtain text regions of candidate license plates; adopting one effective feature for description of each extremal region, classifying the extremal regions into 'text' regions and 'non-text' regions with a classifier obtained in prior training, and extracting a license plate from original images based on the characteristics of the structure of the license plate; and describing with the characteristics of the shape context, and completing the character identification through template matching. Since the maximum stable extremal region has the affine invariance, high stability and multi-scale features, and the regions are determined only according to the gray value and are not light-sensitive,the license plate detection and identification method with the maximum stable extremal region as the base is applicable to the complex background, and has the advantages of good stability and high identification rate.

The invention discloses a moving target detection method, which comprises the steps of background modeling and foreground detection, wherein the background modeling is performed by using a gaussian mixture model, and an operation of model updating is performed so as to obtain B gaussian distributions, for a pixel in a frame image to be detected, if in the B gaussian distributions subjected to sequencing, at least a gaussian distribution is matched with the current pixel value, the pixel is a background pixel, otherwise, and the pixel is determined as a foreground pixel. According to the invention, foreground noises can be effectively filtered so as to obtain an extremely clean background, thereby, an efficient denoising effect is kept, and the accuracy and completeness of foreground object detection are enhanced.

The invention discloses a video target tracking method which is based on multi-cue integration and particle filtration and a device thereof. The method carries out the self-adaptive integration according to three cues of color, edge and feature points to obtain target observation information, then the particle filtration technology is used for updating the target state, and the specific steps comprise: A. a target template is extracted and target parameters are initialized; B. dynamic prediction is carried out on a particle set according to a motion model; C. the particle weight is updated according to multi-cue integration information; D. the motion state of a target is updated. The video target tracking method and the device can improve the tracking effect and the tracking stability of the moving target under the complex environment and the tracked target can comprise human heads, pedestrians, cars, etc.

The invention discloses a video human body tumble detection method and system based on a track weighted depth convolution sequence pooling descriptor. The method comprises the steps of obtaining a convolution feature map of each frame; providing a new track attention map and can be used for positioning a character area in a video; the convolutional feature map of the video frame and the track attention map are weighted to obtain track weighted convolutional features, and the features can effectively describe visual features of a character area in the video; and a clustering pooling method is provided to eliminate redundancy in the sequence; and finally, the track weighted convolution feature sequence is encoded by using a sequential pooling method, and the obtained result is the track weighted depth convolution sequential pooling descriptor. By using the descriptor, the current highest accuracy is obtained on the SDUFall data set, and a good effect is also obtained on the UR data set and the multi-view data set.

The invention discloses a particle filter and perceptual hash-based video target tracking method which comprises the following steps: 1) selecting a target to be tracked from an original image as a target area, extracting a perceptual hash characteristic of the target area as a template feature, and initialization positioning a weight particle set; 2) performing a state coordinate prediction for the weight particle set and labeling the prediction area; 3) extracting the perceptual hash characteristic for each predicting particle area, and calculating a similarity degree of the predicting particle area and the target area; 4) according a position contribution of a distance from the target and a similarity degree refreshing particle weight of the perceptual hash characteristic, adopting a minimum variance to estimate and obtain a new target area; and and 5) according the size of the weight and a resample estimation value, obtaining the weight particle set of a next frame, and repeating from the step 2 to the step 4 till an end of all video sequences. The particle filter and perceptual hash-based video target tracking method disclosed by the invention is high in efficiency, and can realize fast and effective tracking on the premise of great illumination variance and change of the target scale.

The invention discloses an information recommending method based on trust in a social network. According to the method, the trust degree between individuals is utilized to realize the information recommendation, for different individuals in the social network. The method comprises the following steps: 1) giving a representation model for information recommendation of the social network; 2) considering from the two aspects of individual trust relationship and individual confidence coefficient, dividing the trust assessment into a recommender individual, an information source and a recommending route, and establishing a comprehensive trust assessment calculating method; and 3) giving the information recommending method based on trust assessment. According to the method, the information is effectively transferred between the mutual trust individuals in the social network and the failure rate of the information recommendation is reduced.

The invention discloses an information physical fusion system modeling method based on SysML / MARTE. A SysML subset and a MARTE subset are extracted and used for modeling continuous behaviors, random behaviors and nonfunctional properties of a system, and the purpose is to construct a model of the information physical fusion system in a multi-view modeling mode. The modeling method includes the following concrete implement steps that system requirements are analyzed; according to the system requirements, needed modeling elements are selected from the extracted SysML subset and the extracted MARTE subset; an expanded requirement diagram is used for defining property constraint needing to be met to achieve requirement modeling; the extracted SysML modeling elements are used for modeling an architecture of the system to achieve architecture modeling; the extracted MARTE modeling elements are used for modeling the continuous behaviors, the random behaviors and the nonfunctional properties of the system. The SysML / MARTE modeling element expanded subset is constructed, the information physical fusion system is modeled on the well-defined semantic basis, and an effective modeling mode is provided for designing and developing the information physical fusion system.

The invention relates to a fault diagnosis method of sucker-rod pump well; the method comprises the following steps: acquiring ground dynamometer cards of the known and to-be-diagnosed sucker-rod pump wells; converting the ground dynamometer card of the suck-rod pump well into a suck-rod pump well dynamometer card; redrawing a boundary chain code of a normalization processed pump dynamometer card by 16-direction chain codes; zoning the pump dynamometer card which is redrawn with the boundary chain code, and extracting the characteristic vector of each zone; clustering the characteristic vectors of the quantified pump dynamometer card by spectral clustering algorithm based on particle swarm optimization, so as to finish the fault diagnosis of the suck-rod pump well; the fault diagnosis method of sucker-rod pump well uses the number of the 16-direction chain codes to quantify the characteristic vectors of each pump dynamometer card, so as to effectively and precisely describe the pattern characteristics; the fault diagnosis is achieved by clustering the quantified pump dynamometer card characteristic vectors with spectral clustering algorithm based on particle swarm optimization, without depending one training samples, thus increasing the performance of the diagnosis.

The invention discloses a non-uniform geometrical variational nodal method aiming at a reactor neutron diffusion equation. The method comprises the following steps: 1, establishing a functional containing a neutron conservation relationship inside the nodal block and flow continuity conditions on the nodal surface; 2, performing nodal discrete disposal by utilizing a finite element shape function in the x-y direction and a chip constant polynomial, and realizing non-uniform nodal geometrical and material description functions; 3, deducing to obtain a response matrix equation; and 4, solving the response matrix equation by utilizing a red-black scanning manner, and finally obtaining the neutron-flux density distribution in the whole non-uniform solving area. According to the method disclosed by the invention, the non-uniform lattice-cell structure of a nuclear reactor can be finely described; and according to the method, the nodal internal fine geometrical structure is processed based on the variational nodal method by adopting an isoparametric finite element, and then the non-uniform geometrical neutron diffusion equation is solved.

The invention discloses a construction method of a shape descriptor and an image retrieval method based on the descriptor. The construction method of the shape descriptor comprises the following steps: (a) describing a triangular space relationship between two adjacent sampling contour points and a shape mass center by using an only K value, and combining a K value histogram obtained according to K value distribution and a mass center distance histogram obtained according to the distance between the contour points and the shape mass center to construct a new shape descriptor based on contour; (b) clustering the shape contour points, taking the obtained clustering center as new feature points, taking the feature of a clustering region as the feature value of the corresponding feature points, describing a triangular space relationship of the feature of the combined objects between any two new feature point objects and a reference object by using an only L value, and constructing a new shape descriptor based on a region by using one group of L values obtained by each shape and 4-step Zernike matrix; and (c) composing the shape descriptor of the invention by the descriptors based on contour and region. Based on the shape descriptor of the invention, the invention further provides the image retrieval method.

The invention discloses a method for achieving the similar contrast of image colors by using a taper coordinate, which comprises the following steps: 1) a sample image is converted from RGB to HSL; 2) the sample image expressed by HSL color space carries out discretization; 3) an areal coordinate of each subspace is counted according to the HSL color space after the discretization.; the appearance probability of each subspace is counted; 4) the areal coordinate of each subspace is transformed to the taper coordinate; 5) the color characteristics of the sample image extracted according to step 1) to step 4) are deposited to a characteristics database; 6) the corresponding color characteristics of the images to be contrasted are exacted according to step 1) to step 4); 7) according to the distance of two images and the definition of the similarity, the similarity is set; the images to be contrasted are compared with the sample images in the characteristics database of the sample images to find out all similar images which are in line with the requirements of the similarity and then the similar images are displayed. The method of the invention can be applied to image retrieval, which greatly improves the accuracy of the image retrieval.

The invention discloses an anti-shielding infrared target tracking method, solves the problem of long-time target tracking in complex environments such as attitude change and shielding, and belongs to the field of target tracking technologies and computer vision. The method comprises: reading an infrared image sequence; selecting a target in the initial frame of image in a framing manner to obtain the central position and size of the target, taking the target in the initial frame of image as a template, obtaining a second frame of image as a current frame of image, taking the template of the initial frame of image as a template of the current frame of image, and obtaining a two-dimensional cosine window according to the size of the template and the size of a cell unit; and extracting features of the template based on the direction gradient histogram and the Haar, carrying out linear fusion, initializing or updating a target model and a target regression coefficient, and obtaining a multi-layer kernel correlation filtering response graph set based on a search box in each frame of image to carry out subsequent target tracking. The method is used for infrared image target tracking.

The invention provides a polygon-based main road selection method of an urban road network. The polygon-based main road selection method comprises the following steps: firstly, taking OSM map data as a data source and processing to generate area elements; forming a four-dimensional vector by utilizing four shape descriptors including compactness, width, complexity and circularity, and describing polygons in road network data of the generated area element; marking particle polygons and extracting; classifying the obtained four-dimensional vectors by utilizing a non-linear SVM (Support Vector Machine) classification method, and extracting the corresponding polygons after classification; and combining the extracted particle polygons with the polygons obtained by SVM classification to form seed polygons. The invention provides the shape descriptors capable of relatively effectively describing the polygons of a main road, and the polygons with different shape characteristics of the main rod are processed respectively; and finally, based on topological characteristics among the polygons of the road network, the seed polygons extracted by a support vector machine are further processed to obtain the final main road polygons.

The invention discloses a dynamic financial network monitoring analytical method, comprising the following steps especially: (1) building a dynamic financial network model, the flowing conversion process from a financial database to a directional weighted graph is included; (2) computing weight of attribute: by analyzing case database of doubtful transaction, doubtful degree weight of the attribute is obtained by statistical leaning, the weight of attribute is for short; (3) dynamically monitoring analysis of the doubtful degree of account and transaction is carried out; by comprehensively considering the common effect of multiple attributes of account and transaction, and the effect of the last time slice on the current time slice, the doubtful degree of account and transaction along with time is dynamically monitored; (4) updating the weight of attribute: when a new doubtful transaction appears, the case database of doubtful transaction and the weight of attribute are updated; (5) incremental updating the dynamic financial network model: when a new data of financial transaction appears, the financial database and the dynamic financial network model are updated.

The invention relates to the technical field of rock mass mechanics, and discloses a shale brittleness index evaluation method, device and system. The method comprises the following steps: obtaining stress-strain curves of a rock sample under different confining pressures, and calculating a pre-peak energy evolution characteristic value and a post-peak energy evolution characteristic value according to the stress-strain curves; calculating a pre-peak brittleness index according to the pre-peak energy evolution characteristic value, and calculating a post-peak brittleness index according to thepost-peak energy evolution characteristic value; and obtaining a comprehensive brittleness index by combining the pre-peak brittleness index and the post-peak brittleness index, and evaluating the shale brittleness according to the comprehensive brittleness index. According to the comprehensive brittleness index provided by the invention, the whole process of shale deformation and failure under different confining pressures is comprehensively considered, and brittleness characteristics of different rocks under different confining pressure conditions can be well reflected.

The invention provides a scene image generation method and device based on a generative adversarial network. The method comprises the steps of obtaining current text information; extracting text features in the current text information; extracting scene graph features in the current text information; performing feature fusion on the scene graph features and the text features to generate fusion features; and inputting the fusion feature into a pre-trained scene image generation model to generate a corresponding scene image. In this way, the scene image can be generated according to the description of the text information, so that the generated scene image is closer to the real scene image, and the position relationship between the objects in the scene image can be effectively described.

The invention relates to a method and a device for recognizing human behavior based on multi-mode channel feature fusion. The method comprises the following steps: 1) constructing an adaptive hierarchical structure for an input depth image sequence; 2) extracting DMM features at each level of the adaptive hierarchical structure; 3) serially connecting DMM features of each level to construct adaptive depth motion map features as depth image channel features; 4) selecting relatively stable skeleton joints as reference joints according to the input skeleton image sequence; 5) calculating the displacement difference between the other joint nodes and the reference joint nodes in each frame as the feature expression in each frame; 6) combining the whole skeleton image sequence to obtain a feature expression of an action sequence as a skeleton image channel feature; 7) through feature fusion and classifying the fused feature, the human body behavior recognition result is obtained. The invention can describe the human body time sequence motion information and the spatial structure information at the same time, and has good recognition effect and robustness.

The invention discloses a method and device for recognizing the behavior of a human body. The specific execution mode of the method comprises the steps: extracting the movement track features of a video image; carrying out the coding of the movement track features in a plurality of coding modes; recognizing the coded movement track features, and determining the type of the behavior of the human body. According to the execution mode, the movement track features coded in the plurality of coding modes are used for recognizing the behavior of the human body, thereby improving the recognition precision of the behavior of the human body.

The invention provides an automatic testing method of distributed big data service. The method comprises the following steps: describing a testing interface by adopting XML Schema, proposing corresponding standard so as to effectively describe a web application or API of a cloud platform. by describing the XML of the API, a testing sequence, the testing script and the testing data can be automatically generated. Through the testing method provided by the invention, the user can automatically construct the testing sequence and generate the testing data by only describing the web application orthe API of the cloud platform, thereby achieving an automatic testing aim, the workload of the user is greatly lightened, the testing efficiency and the accuracy rate are improved, and the testing cost is reduced. And meanwhile, the pressure testing and like artificial testing and impossible testing work can be performed on the web application or the cloud platform.

An embodiment of the invention discloses a method for obtaining graph structural description information. The method includes that a structure of a graph is extracted in a given graph; the structure of the graph is converted into an information flow graph, the information flow graph is used for representation of an information flow space defined by the structure of the graph, and each node in the structure of the graph carries out information flow according to information flow rules in the information flow graph; the initial information amount of each node in the information flow graph is determined in the structure of the graph; according to the initial information amount of each node, according to the information flow rules, the information amount of each node at a preset time point is calculated; and according to a time sequence of information distribution formed by information amounts of all the preset time points, the structural description information of the graph is determined. The method for obtaining the graph structural description information achieves dynamic analysis for the structure of the graph, dynamic description information of the structure of the graph is obtained, the information amount of the structural description information of the graph is increased, and the structure of the graph can be effectively described.

The invention discloses a zero value insulator automatic detection method based on an infrared image. The method comprises the following steps that an infrared image acquisition apparatus carries outimage acquisition on an insulator on a power transmission line so as to form a test image and sends the image to the outside world; a central processing unit receives the test image and carries out preprocessing so as to acquire a preprocessed infrared image; a characteristic area and a characteristic point of the preprocessed infrared image are acquired; the insulator is positioned and an insulator area is acquired; the insulator is segmented; and a zero value insulator is determined, if there is the zero value insulator, an industrial control computer emits an alarm sound, otherwise, there is no response. The invention also provides a zero value insulator automatic detection apparatus based on the infrared image. In the invention, the zero value insulator is determined; the insulator canbe accurately and automatically positioned; and accuracy of zero value insulator detection and determination can be effectively increased.

The invention relates to a pedestrian detection, belongs to the technical field of detection, and solves problems in the prior art that the real-time performance is poor, the shielding is difficult toprocess and a method is not suitable for occasions with the apparent illumination changes and strong noise. The method comprises the following steps: inputting an image, and obtaining gray scale values of the pixels; obtaining three CRLBP operators according to the gray scale values of the pixels, and obtaining a CRLBP texture feature spectrum; calculating HOG features of the input image, CRLBP histogram features, and HOG features of a CRLBP texture feature spectrum; carrying out the fusion of the HOG features of the input image, the CRLBP histogram features, and the HOG features of the CRLBPtexture feature spectrum to obtain an image descriptor; reducing the dimensions of the image descriptor through a principal component analysis method, and employing a classifier for a dimension reduction result to achieve pedestrian detection and recognition. The method is high in detection efficiency, is good in real-time performances, and is good in robustness for illumination and noise.

The invention discloses a similar case calculation method based on a criminal behavior chain, and the method comprises the following steps: 1, extracting related information from a case description part of a judgment document, employing Bert + Crf as a sequence labeling model, carrying out the extraction of key information, and finally constructing the criminal behavior chain; 2, based on the structural features of the criminal behavior chain, adopting a graph neural network model as a basic calculation model to realize similarity calculation of the criminal behavior chain; and 3, combining aresult obtained by the graph neural network model with the text content information to realize finding of similar cases by a crime behavior chain-based similarity calculation method. According to theinvention, the characteristics of a criminal behavior chain are fully utilized, a deeper calculation method is provided for traditional similar case searching, the similarity between similar cases isimproved, technical support is provided for pushing of class cases, and a very good effect is achieved.

The invention provides a background modeling and motion object detection method with image gradient and gray scale integration. The method comprises: 1, collecting N frames of video monitoring data as a training sample; 2, establishing an image-gray-scale-based single Gaussian model; 3, selecting pixel points to establish a gradient model; and 4, carrying out moving object detection. In addition, the invention also provides a background modeling and motion object detection apparatus with image gradient and gray scale integration. The apparatus consists of an image acquisition unit, a gray scale background modeling unit, a gradient background modeling unit, a model judging unit, a moving object detection unit and a background model updating unit. According to the invention, backgrounds are distinguished into a single back ground and a complicated background based on a gray scale standard difference and different background models are established for description; and with the single background model, the background can be described rapidly and effectively, and with the complicated background model, the complicated background change can be described well. Therefore, accuracy of the moving object detection can be improved well.

The invention provides a thin mutual reservoir description method taking a reconstructed sand-to-ground ratio as a seismic description target. The method comprises steps of S1, explaining fine seismicstructures of top and bottom surfaces of a sand layer group; S2, generating a plurality of seismic attribute bodies by applying seismic means; S3, analyzing the correlation between the various seismic attributes and a sand-to-ground ratio and a net-to-gross ratio, determining whether the correlation has zoning or not, and selecting the seismic attributes with high correlation in zones for subsequent research; S4, respectively reconstructing the sand-to-ground ratio according to the zones in the S3, calculating the correlation between the reconstructed sand-to-ground ratio and the net-to-grossratio and the optimal seismic attributes, and giving a linear formula; S5, making an optimized seismic attribute planar graph, and converting the seismic attribute planar graph into a reconstructed sand-to-ground ratio or net-to-gross ratio planar distribution graph by utilizing the formula in the S4; and S6, deriving a sand-to-ground ratio and net-to-gross ratio planar graph by utilizing the reconstructed sand-to-ground ratio and net-to-gross ratio planar graph. The method is advantaged in that the reconstructed sand-to-ground ratio is utilized as the seismic description target, so the reconstructed sand-to-ground ratio and the seismic attributes have high similarity, the distribution rule of a thin mutual reservoir is effectively described, and the rule can effectively guide explorationand development practices.

The invention discloses a video target tracking method which is based on multi-cue integration and particle filtration and a device thereof. The method carries out the self-adaptive integration according to three cues of color, edge and feature points to obtain target observation information, then the particle filtration technology is used for updating the target state, and the specific steps comprise: A. a target template is extracted and target parameters are initialized; B. dynamic prediction is carried out on a particle set according to a motion model; C. the particle weight is updated according to multi-cue integration information; D. the motion state of a target is updated. The video target tracking method and the device can improve the tracking effect and the tracking stability of the moving target under the complex environment and the tracked target can comprise human heads, pedestrians, cars, etc.

The invention discloses an information physical fusion system modeling method based on SysML / MARTE. A SysML subset and a MARTE subset are extracted and used for modeling continuous behaviors, random behaviors and nonfunctional properties of a system, and the purpose is to construct a model of the information physical fusion system in a multi-view modeling mode. The modeling method includes the following concrete implement steps that system requirements are analyzed; according to the system requirements, needed modeling elements are selected from the extracted SysML subset and the extracted MARTE subset; an expanded requirement diagram is used for defining property constraint needing to be met to achieve requirement modeling; the extracted SysML modeling elements are used for modeling an architecture of the system to achieve architecture modeling; the extracted MARTE modeling elements are used for modeling the continuous behaviors, the random behaviors and the nonfunctional properties of the system. The SysML / MARTE modeling element expanded subset is constructed, the information physical fusion system is modeled on the well-defined semantic basis, and an effective modeling mode is provided for designing and developing the information physical fusion system.

The invention discloses a substation main and auxiliary equipment panoramic model and a modeling method. The method comprises the following steps: carrying out spatial association constraint modelingon various main and auxiliary equipment of the transformer substation; establishing full-life-cycle time historical attributes of main and auxiliary equipment objects of the transformer substation andtime dynamic attributes of all sensing monitoring points of the equipment, describing space state attributes of the equipment with the full-life-cycle time historical attributes of the equipment andcomponents of the equipment based on the space association model and the time association model, and establishing a space-time association model of the main and auxiliary equipment of the transformersubstation. According to the substation main and auxiliary equipment panoramic model, spatial characteristics and association constraint relationships of equipment objects and monitoring points of theequipment objects are considered, description of time characteristic dynamic change elements of the equipment is also considered, dynamic change of spatial characteristic attributes along with time is considered, time is brought into a space system, and penetration perception of the state of the substation equipment is realized.

The invention discloses a non-uniform geometrical variational nodal method aiming at a reactor neutron diffusion equation. The method comprises the following steps: 1, establishing a functional containing a neutron conservation relationship inside the nodal block and flow continuity conditions on the nodal surface; 2, performing nodal discrete disposal by utilizing a finite element shape function in the x-y direction and a chip constant polynomial, and realizing non-uniform nodal geometrical and material description functions; 3, deducing to obtain a response matrix equation; and 4, solving the response matrix equation by utilizing a red-black scanning manner, and finally obtaining the neutron-flux density distribution in the whole non-uniform solving area. According to the method disclosed by the invention, the non-uniform lattice-cell structure of a nuclear reactor can be finely described; and according to the method, the nodal internal fine geometrical structure is processed based on the variational nodal method by adopting an isoparametric finite element, and then the non-uniform geometrical neutron diffusion equation is solved.