36results about How to "No prior knowledge required" patented technology

The invention discloses a method for monitoring Lamb wave engineering structural damage by utilizing signal decomposition, which comprises the following steps: collecting the Lamb wave healthy reference signals fhji of an engineering structure in a healthy state, wherein i and j can be 1, 2, 3,...; when the structure is damaged, collecting all the Lamb wave response signals f'ji of the damaged structure under the condition of an inspiring/sensing channel; utilizing the correlation operation to extract all the Lamb wave damage scattered signals of the damaged structure under the condition of the inspiring/sensing channel; obtaining the damage position and scope according to the characteristic parameters acquired in the steps above, thereby analyzing and judging the health condition of the monitored structure. According to the method, based on the condition of the traditional structural damage monitoring device, the Lamb wave damage scattered signals in a damage monitoring system are extracted by using the related treatment, thereby increasing the accuracy and stability of the monitoring system.

The invention discloses an inner and outer layer nesting ECMS (equivalent fuel consumption minimization strategy) multi-objective double-layer optimization method. The inner and outer layer nesting ECMS multi-objective double-layer optimization method includes steps of building multi-objective optimization models of plug-in hybrid electric vehicles; solving the multi-objective optimization modelsby the aid of inner and outer layer nesting multi-objective particle swarm algorithms to obtain multi-objective optimized Pareto solution set front edges; weighting equivalent fuel consumption per hundred kilometers and variation ranges of deviation of SOC (state of charge) final values and target values, building total evaluation functions related to the equivalent fuel consumption per hundred kilometers and SOC deviation and selecting the optimal charge and discharge equivalent factors and engine and motor power distribution modes corresponding to the optimal charge and discharge equivalentfactors. The inner and outer layer nesting ECMS multi-objective double-layer optimization method has the advantages that output power of engines and motors of the plug-in hybrid electric vehicles canbe reasonably distributed at CS (charge sustaining) stages, so that fuel consumption can be reduced as much as possible, battery SOC balance still can be effectively kept, and the fuel economy of theintegral vehicles can be improved.

The present invention relates to an image and video processing method, especially to a method for extracting and correlating video interested objects based on a two-stage interaction. In the method, a user performs an approximately positioning interaction by using a kind of interaction method which is not limited to a general mode and with a low requirement for priori knowledges, and then performs multiparameter extraction on the interested objects by using a quick and realizable extraction algorithm. Based on the most suitable interested object obtained by the reselection and interaction, the present invention performs searching and weighting by extracting various characteristics to obtain a final result, picks out corresponding images and additional information from a value-added information base, and shows the information at peripheral of the video. The method correlates the value-added information to the interested object of the user without influencing watching of the user by fully mining video information and ensuring preference of the user, thereby satisfying the requirment of the user that the area of interest needs to be looked into and further explored.

The invention discloses a passive target localization method for wireless sensor node random deployment. The passive target localization method includes the steps of first, deploying sensor nodes and base stations in a monitored area; second, acquiring RSS (Received Signal Strength) values on the condition without interference before monitoring; third, acquiring and classifying the RSS values at the monitoring stage; fourth, determining locations of links interfered by targets in the monitored area, namely locations of nodes constituting the interfered links, for accurately positioning the target location on the condition that the nodes in the monitored area are in random deployment; fifth, determining the target location in the monitored area. The passive target localization method can be implemented without priori knowledge. As meshing is omitted, the defect that location precision is affected by size of meshes in meshing-based localization methods is avoided.

The invention relates to a transformer winding looseness fault detection method based on an average displacement method, and belongs to the technical field of power transformer safety monitoring. The method comprises the steps that surface vibration data of a transformer with an unloosened winding are measured at preset time intervals within preset time range, the average displacement of the transformer is calculated, and a first average displacement sequence is obtained; surface vibration data of a transformer to be measured are obtained at the same time intervals within the same time range, the average displacement of the transformer is calculated, and a second average displacement sequence is obtained; difference values between the average displacements at the same time delay are calculated, the difference values are added, and evaluation factors for evaluating the winding looseness degree are obtained. The transformer winding looseness fault detection method is easy to implement, easy and convenient to operate and accurate in diagnosis, and the number of needed devices is small.

The invention provides a method for classifying pixel blocks of JPEG images and image falsification detecting and falsified area positioning methods based on the same. The method comprises the following steps of: dividing to-be-detected JPEG images into continuous and non-overlapped 8*8 pixel blocks; taking one pixel block as the center and selecting a square area combined by (2n+1)*(2n+1) pixel blocks as a sub-image; utilizing a first number distribution probability feature of AC coefficients in the front i AC coefficient regions of the sub-image to divide the central pixel block into singlecompression or double compression; and confirming a classifying result as the double compression and a connected region with an area more than M 8*8 pixel blocks as a falsified area. According to themethod provided by the invention, the falsified area can be accurately detected and positioned; the falsified areas of three different falsifying technologies can be detected according to the method;the method is wide in application scope; the method is an automatic blind detecting method for JPEG falsified images and does not require any priori knowledge; and the detection coverage rate is higher and the error detection rate is low.

The present invention belongs to the field of biological characteristic distinction, and is especially the method of extracting geometric characteristic of iris based on characteristic edge distribution. The method includes first defining geometric characteristic of iris and describing its reliability and stability; subsequent extracting geometric characteristic edge of iris in antisymmetric double orthogonal wavelet process and constituting overall edge distribution characteristic in unequal interval grading according to the inhomogeneous distribution of geometric characteristic of iris; and finally introducing characteristic distance expression and proposing legal user judging method. The method has high distinction precision and less outer factor influence.

The invention discloses a convolutional neural network fault diagnosis method based on a multi-channel attention module. The method comprises the following steps of: 1, collecting vibration signals ofto-be-diagnosed equipment under different working conditions; 2, performing data enhancement processing and fault labeling on the vibration signals acquired in the step 1; 3, establishing a convolutional neural network fault diagnosis model based on a multi-channel attention module; 4, inputting a vibration signal under a certain working condition into the convolutional neural network fault diagnosis model based on a multi-channel attention module established in the step 3, and training the model; and 5, inputting a vibration signal under another working condition into the fault diagnosis model with the trained network parameters, carrying out fault state identification, and outputting a fault label of the to-be-diagnosed equipment to obtain a fault type. According to the method, higher fault identification capability and generalization capability can be obtained under variable working conditions, and the problem of poor generalization performance of a traditional convolutional neuralnetwork model under variable working conditions is solved.

The invention provides a shrimp meat freshness detection method based on a spatial migration Raman spectrum, and belongs to the technical field of food nondestructive testing. The method, based on a Raman scattering point light source image detection system, comprises the steps: collecting and obtaining Raman images of unhousinged shrimp samples at different positions and under 1024 wave bands; converting the Raman image intensity wavelength domain into an intensity space domain, and extracting an interested region and waveband as sub-images of the corresponding Raman image; selecting the optimal spatial offset distance to make the internal signal relatively strongest and the signal-to-noise ratio highest; removing abnormal point spectral data, and inputting the abnormal point spectral data into random forest features for feature band selection; and inputting the spectral data of the selected characteristic wave band into a support vector regression model for prediction. The method ishigh in prediction precision and high in speed for the freshness of the unhousinged shrimps, and field detection can be achieved; and meanwhile, a deep detection method is provided, and the method isexpected to be applied to deep detection of unhousinged or packaged or multi-layer samples.

The invention discloses a hyperspectral image suspicious target detection method based on low-rank sparse representation. The method comprises the following steps: reading a hyperspectral image data cube and related projection parameters; clustering spectrums of all pixels in the input hyperspectral image to obtain a plurality of clustering centers and a category to which each pixel belongs; counting an intra-class clustering result, and determining the distance between each pixel and the clustering center of the class to which the pixel belongs; selecting s pixels with the minimum distance asvarious description sample points, and adding the description sample points into a final dictionary to form a reconstruction dictionary of the scene remote sensing data; performing low-rank sparse constraint matrix decomposition on the hyperspectral image data by using the reconstructed dictionary; carrying out residual error statistics on the decomposed residual error; and outputting the T pixels with the maximum residual values and the ground coordinates corresponding to the T pixels as suspicious targets. The detection method provided by the invention has good adaptability to hyperspectralimages collected by different sensors, has a low false alarm rate, and still has a high detection capability for small targets.

The invention discloses a CT image beam hardening artifact correction system based on adjustable double factors, and particularly relates to the technical field of X-ray CT imaging. According to the system, in the correction process, a high-density region in a CT image is segmented to obtain edge information; projection data of the high-density area is calculated; the attenuation coefficient of the substance to the ray is approximately calculated to obtain artifact data caused by a high-density region; and finally, artifact data is reconstructed by using a classical filtering back projection algorithm, and the reconstructed artifact image is substracted from the original image to obtain a corrected CT image. The system can provide high-quality images for industrial CT and medical CT detection, and compared with a previous CT image beam hardening artifact correction method, the system can quickly and effectively eliminate hardening artifacts without priori knowledge.

The invention provides a reductive-rule-based anti-interference decision-making method of a wireless communication system. According to a rough set theory, a decision-making space is simplified to obtain a reduced decision-making rule; screening of an initial population generated by a genetic algorithm randomly is carried out according to the decision-making rule to keep an individual meeting the decision-making rule; and on the step of the genetic algorithm, optimization seeking is carried out on a communication parameter. In order to improve practicability of the algorithm, a maximum transmission rate that is obtained on the condition of a minimum transmitting power under a designated error rate is used as a target function of communication anti-interference decision making, thereby avoiding the dependence on an experience weighted value by the traditional algorithm. The simulation result demonstrates that the convergence speed of the genetic algorithm is substantially improved on the premise that calculation accuracy is guaranteed; and the practicability is high.

The invention relates to a survival analysis method for predicting machine damage time, which decomposes a survival analysis problem for predicting the machine damage time into sub-problems of a timeslice, and greatly reduces the difficulty of using a neural network to model a long-time sequence prediction problem after decomposing a time sequence prediction problem on the whole time length; therisk probability of each time slice is modeled by using the same neural network, and the final survival probability is obtained through a conditional probability rule. On the premise of not carrying out any assumption on the time distribution of the damage time of the machine, a prediction model can be trained by combining big data. The invention not only can be used for predicting the survival probability of discrete time slices, but also can play a role in predicting the survival probability of continuous time. Experiments prove that the prediction accuracy of the survival analysis model trained through the deep neural network is far higher than that of a traditional method. And through parallel calculation, the algorithm can perform long-distance survival probability prediction under the condition of not increasing the operation time.