261results about How to "Reduce the number of samples" patented technology

The invention discloses a sparse sampling and signal compressive sensing reconstruction method. The method comprises: establishing a signal sampling interval of each time, sampling point number, and the number of points recovering, establishing random sparse sampling lower than a Nyquist sampling theorem value; and designing a measurement matrix by random sampling timing sequence values, designing a transformation matrix of a sparse expression domain of signals, determining a compressive sensing matrix, and separated compressive sensing optimizing signal reconstruction in a nonlinear manner. The method is based on rationality of objective world rules, and makes full use of signal sparsity, uses transformation space to describe the signals, and establishes theoretical framework of new signal description and processing, so under the condition that information is not lost is ensured, signals are sampled by speed much lower than required speed of a Shannon's sampling theorem. Simultaneously, signals can be recovered completely, that is, sampling of signals is converted into sampling of information. The invention provides a whole set of complete method. The method can be used in one-dimensional and multidimensional signals, and can process audio frequency, videos, nuclear magnetic resonance, and other signals.

The invention provides a method for identifying a voltage sag reason. The method for identifying the voltage sag reason comprises the following steps of (1) using the discrete Fourier transform to calculate a spectral density function, converting the voltage in a time domain to the voltage in a frequency domain, (2) calculating the characteristic quantity related to identification of the voltage sag reason, (3) defining a sag phase number N1, judging whether swell and a sag combination N3 exist or not, and (4) identifying the voltage sag reason. The method for identifying the voltage sag reason is comprehensive and higher in adaptability to changes of voltage sags, and can be easily improved, verification is carried out on the basis of really-measured data of a power grid, and the calculation result has better engineering practicability.

The invention discloses a quick splicing method and system of infrared images. The method comprises steps of acquiring n images of a scene, wherein every adjacent two images are partially overlapped;using the SIFT algorithm to carry out characteristic point detection and characteristic point pre-matching on the images, and using a Manhattan distance as the similarity measurement during characteristic point matching; removing error matching, using the RANSAC algorithm to remove outer points, estimating model stability, and arranging the matching point pairs with the high reliability to be thefront line in the RANSAC algorithm; carrying out cylindrical projection on the images, projecting each pixel point of each picture to the same coordinate system, and using the linear weighting methodto carry out image fusion on the matched pictures so as to acquire a panoramic image of the whole scene. The system corresponds to the method. According to the invention, by improving the classic SIFTalgorithm and the RANSAC algorithm, under the condition that the splicing effects are not influenced, the splicing speed is increased and the splicing precision is quite high.

The invention relates to a block and sparse principal feature extraction-based rapid collaborative saliency detection method. The method sequentially includes the four steps of feature extraction of image blocks and original pixels, sparse principal feature extraction, clustering-based feature computation and collaborative saliency map generation. With the block and sparse principal feature extraction-based rapid collaborative saliency detection method adopted, limitations of application of existing collaborative saliency detection methods to various kinds of aspects which are brought by problems of memory occupation and time consumption which are further caused by ignoring of other saliency targets as well as increasingly high-definition images and the increase of the number of processing images in the prior art can be avoided.

The invention relates to a fisheye lens image distortion correction method based on an orthogonal projection, which comprises the following steps: 1) using a fisheye camera to take a picture, that is,a fisheye image; 2) performing distortion correction on the fisheye image: obtaining the radius and center point of the fisheye image according to the geometrical characteristics of the fisheye image; using an orthogonal projection to transfer the fisheye image into a hemispherical surface; using lateral and longitude curves to carry out mesh division on the hemispherical surface, thereby transferring into lateral and longitude coordinate values of the hemispherical surface; finally, projecting the lateral and longitude coordinate values to a target image coordinate to realize coarse correction conversion of the fisheye image; after using the orthogonal projection to carry out coarse correction on the fisheye image, using a plurality of angular points of the distorted checker image and the target image to carry out affine transformation; then, using a look-up table to save a coordinate mapping relation table from the fisheye image to the correction image; and using the look-up table to achieve accurate correction of the fisheye image in the checkerboard region. The method in the invention is stable in performance, safe and reliable.

The invention discloses a support vector machine high-voltage circuit breaker fault diagnosis method based on fuzzy clustering. The method comprises the steps of firstly, extracting a feature vector from a stroke-time curve of a circuit breaker moving contact, using the feature vector as a database of fault diagnosis, secondly, conducting fizzy clustering processing on a data sample, generating a new clustering center matrix, thirdly, using the clustering center matrix as a training sample, applying a support vector machine to conduct training, and fourthly applying the high-voltage circuit breaker fault diagnosis method to diagnose test data. According to the support vector machine high-voltage circuit breaker fault diagnosis method based on the fuzzy clustering, the efficiency of the high-voltage circuit breaker fault diagnosis can be effectively improved, the time of the fault diagnosis is reduced, the quality of the fault diagnosis is improved, and the support vector machine high-voltage circuit breaker fault diagnosis method has great significance for research on the safety and the reliability of the power grid.

The invention provides a circular center detection method based on Gaussian fitting sub-pixel edge detection and an improved RANSAC algorithm, belonging to the field of aircraft ground simulation. Themethod comprises the follow concrete steps: an original image is matched by a model and RIO processing is conducted on the image; edge pixel extraction and thinning is carried out; sub-pixel edge location based on Gaussian fitting is carried out; an improved RANSAC circle fitting algorithm is provided; the coordinates of the center of a circle are calculated. The invention roughly locates the position of the circular marker through a model matching mode, and performs RIO processing to reduce the size of image processing. Gaussian fitting improves the accuracy and stability of the algorithm. The improved RACNAC algorithm reduces the sampling times, optimizes and screens the candidate circles, reduces the running time and improves the accuracy and stability.

The invention belongs to the field of event-triggered control design, and discloses an event-triggered control design method for a switching system with actuator saturation. In order to reduce the sampling times of the system and save network resources effectively, the event-triggered control design method applies an event-triggered strategy to the switching system with actuator saturation and ensures the stability of a closed-loop system adopting event-triggered control. The event-triggered control design method specifically comprises the steps of: step 1, establishing the switching system with actuator saturation; step 2, designing a trigger condition, and determining a trigger moment t<e><k>; step 3, designing conditions for the closed-loop system; step 4, designing preconditions in order to meet establishment of the conditions; step 5, and designing conditions to obtain a result that exponents of the closed-loop system are stable. The event-triggered control design method is mainlyapplied to the event-triggered control occasions.

In a system analysis apparatus, a per-SrcIP filter section captures packets in client-Web server segments on a per-SrcIP basis. An ON/OFF filter section allows captured packets to pass through only when its flag is set to “ON”. A request/response analysis section controls a flag of the ON/OFF filter section of the next segment so that the packets of the next segment are captured only when the present segment is in a response waiting statue. A range control section controls a SrcIP range of packets to be captured by the per-SrcIP filter section.

The invention belongs to the field of controller design, aims at decreasing sampling points of an extended state observer, effectively saving network bandwidth resources, reducing network communication pressure and meanwhile ensuring the performance and stability of a closed-loop system adopting active disturbance rejection control, applies an event trigger mechanism to controller design and provides a nonlinear system event trigger controller designing method based on the extended state observer.The nonlinear system event trigger controller designing method specifically comprises the following steps that step 1, a single-input single-output nonlinear system model (shown in the description) is established; step 2, for estimating a state of the system and an extended state defined in the step (2), the extended state observer is designed; step 3, a sampling moment ti is determined; step 4, the state, sampled at the sampling moment ti, of the extended state observer is transmitted to the controller.The nonlinear system event trigger controller designing method is mainly applied to controller design occasions.

The invention relates to a method for reconstructing the spectral reflectance of a single-pixel detector based on sparse priors, which performs principal component analysis on the training sample set, and obtains the first three principal components of the spectral reflectance data of the training sample set as the reconstruction basis function Vector; through the single-pixel detector to collect the spectral energy of a single multi-spectral test color block, the energy value U is obtained; in the process of solving the training sample set, the basis function vector B, the basis function vector coefficient a and the specific coefficient of the measurement matrix are obtained The reflectance of the test sample is reconstructed by the obtained spectral energy U collection for a single multi-spectral test color patch. The present invention can make full use of the spatially sparse feature of spectral reflectance and the sparse prior knowledge of the relative spectral power distribution of the lighting source based on the principal component orthogonal basis, reduce the optical complexity of the multi-spectral data acquisition system, reduce the number of samples, and improve the spectrum of the reflectance Improve reconstruction efficiency and improve reconstruction accuracy.

The invention relates to a power transmission line icing state assessment method based on information fusion of multiple sensors. The method comprises the following steps: (1), obtaining an environment temperature and environment relative humidity, determining whether an icing condition is reached, and if so, executing step (2); (2), obtaining an environment temperature difference, icing duration and an icing thickness, and according to expert experience, calculating icing membership values of the environment temperature difference, an equivalent icing thickness and the icing duration; (3),performing data fusion on each icing membership value by use of a BP nerve network, and respectively calculating an icing-free probability, an icing probability and a severe icing probability; and (4), according to the icing probability, the severe icing probability and the icing duration, outputting an assessment result. Compared to the prior art, the method provided by the invention has the advantages of accurate assessment and the like.

The invention discloses a hyperspectral image classification and wave band selection method based on multi-target immune cloning. The hyperspectral image classification and wave band selection method based on multi-target immune cloning comprises the following steps that a sparse representation classifier is used for classifying hyperspectral remote sensing images so as to obtain classified images based on pixels; the hyperspectral remote sensing images are preprocessed by means of mean shift, the processed images are segmented into a plurality of clustering results through a multi-target immune cloning clustering and wave band selection algorithm, and the optimal clustering results are selected from the clustering results so as to constitute a segmentation image; the obtained classified images and the obtained segmentation image are fused by means of the maximum voting rule so as to obtain a final result image. The hyperspectral image classification and wave band selection method based on multi-target immune cloning has the advantages that the very high accuracy rate and Kappa coefficient values can be obtained under the circumstance that few training samples exist, region consistency is well kept, the hyperspectral image classification and wave band selection method is suitable for multiple hyperspectral data, and parameters are adjusted easily and conveniently.

The invention discloses an adaptive mixed indoor positioning method based on a Wi-Fi, and the method comprises the steps: collecting the position coordinates of a few of sample points in an indoor central region and the corresponding signal intensity information, and the position coordinates of a large number of sample points in an indoor boundary region and the corresponding signal intensity information; calculating the three-side positioning coordinates of a test point through employing a three-side positioning algorithm, and calculating the fingerprint positioning coordinates of the test point through employing a position fingerprint positioning algorithm; carrying out the ordering of the Euclidean distances between the test point and all sample points, and selecting the former M groups of data with the minimum distance; determining the weight coefficients of the three-side positioning coordinates according to the number of sample points in the indoor central region of the M groups of data, determining the weight coefficients of the fingerprint positioning coordinates according to the number of sample points in the indoor boundary region of the M groups of data, and finally obtaining the final positioning coordinates of the test point. The positioning method can maintain higher positioning precision while reducing the sampling number of the indoor central region.

The invention relates to an on-line water quality monitoring system, comprising a movable water quality monitoring node and a relay monitoring node which are floating on a water surface, and an ashore base station and a data center, wherein the movable water quality monitoring node and the relay monitoring node are used for monitoring water quality parameters and uploading the obtained water quality parameters to a cluster head node; the cluster head node fuses the collected data and reports to the ashore base station according to a specified format; the ashore base station uploads the data according to a monitoring query command of the data center; the data center is used for reconfiguring the monitoring data; and the movable water quality monitoring node and the relay monitoring node employ a random sampling strategy to sample and monitor the water quality. The life cycle of a conventional water quality monitoring method can be prolonged by the on-line water quality monitoring system.

The invention provides a correlated imaging system used for carrying out correlated imaging on an object to be imaged through a thermal light source. The correlated imaging system comprises an object arm light path and a first reference arm light path, wherein the object arm light path is internally provided with a first barrel detector and the object to be imaged; the first barrel detector samples a total light field intensity signal Sm, passing through the object to be imaged, in the object arm light path; the first reference arm light path is provided with a reference detector device for sampling the distribution information of the light field intensity of the first reference arm light path; the reference detector device comprises at least one reference detector unit; each reference detector unit comprises a timing sequence controller and a plurality of reference detectors with the spatial resolving power; the reference detectors are controlled by the timing sequence controller; the reference detectors are exposed in sequence for sampling under the control of the timing sequence controller. As the reference detectors are alternately exposed in time sequence, the exposure frame rates can be overlapped, limitation of existing reference detectors on the sampling speed is broken through, the sampling speed is greatly increased, and the imaging time is shortened.

The invention discloses a method for extracting a state discrimination characteristic parameter of a metal drawing piece, comprising the following steps of: setting acoustic emission sensors at the edge of a metal workpiece and the edge of a terrace die for acquiring acoustic emission signals at different parts and transmitting the acoustic emission signals to a computer and carrying out the normalization processing on the signal in the computer; carrying out the wavelet-packet decomposition on the signal subjected to the normalization processing and reconstructing the signal in a frequency range to obtain a signal containing frequency components of a metal crack acoustic emission signal; and finally, establishing an autoregression model of a metal crack acoustic emission signal frequency range, calculating an autoregression spectrum according to autoregression model parameters, extracting accurate energy value of the acoustic emission signal frequency range according to a formula and using the energy value as the characteristic parameter for reflecting the metal crack characteristic. The invention improves the accuracy for parameter identification and state prediction of the metal drawing state discrimination, greatly reduces the sampling quantity and is suitable for high-requirement occasions for monitoring short-time pulse signals on line.

The invention discloses a compressed sensing based color imaging device and a compressed sensing based color imaging method. The compressed sensing based color imaging device comprises a PC (personal computer), a DLP (digital light projector), a target image, a single-pixel photon detector and a data collection and control module. The PC, the DLP and the single-pixel photon detector are connected with the data collection and control module. The PC generates a two-dimensional random binary projected intensity image. The data collection and control module controls the DLP to perform binary intensity image projection on the target image under red, green, blue and white structured light sequentially to acquire digital signals of the red, green, blue and white structured light respectively and send the digital signals to the PC, reconstructs the signals of the red, green, blue and white structured light through compressed sensing to acquire four monochrome gray level images, and performing color fusion on the four monochrome gray level images to acquire a final color image. The compressed sensing based color imaging device and the compressed sensing based color imaging method have the advantages that light filters are not needed, and simple structure and low cost are achieved; sampling frequency can be reduced, and color fidelity can be guaranteed without reduction of image resolution.

The invention discloses a bidirectional iteration parallel probability load flow calculation method combining Latin hypercube sampling. The method includes: using a Latin hypercube sampling method to sample new energy power generating power variables according to the accumulation distribution function of the new energy power generating power variables, and building the sample matrix of the new energy power generating power variables; aiming at the partitionable feature of a power grid, using a branch cutting and node tearing method to concentrate the power grid into meshes, and building a bidirectional iteration parallel load flow calculation model; using the sample matrix of the new energy power generating power variables as the input variable of the bidirectional iteration parallel load flow calculation model to perform probability load flow calculation so as to obtain the discrete result of output variables, and using kernel density estimation to fit the discrete result of the output variables to obtain the probability density function of the output variables. The method has the advantages that the Latin hypercube sampling is combined with the bidirectional iteration parallel load flow algorithm based on the concentrated meshes, sampling number is reduced, and parallel probability load flow calculation is achieved.

The invention discloses a method for identifying a fine crack impact signal of a metal deep drawing part. The method comprises the following steps: acquiring a fine crack acoustic emission impact signal of a metal deep drawing part by using an acoustic emission sensor, carrying out preamplification, filtering and A/D conversion pretreatment on the acquired signal, inputting the pretreated signal into a computer, analyzing the wavelet packets, reconstructing data at different wave bands after the wavelet packets are decomposed, carrying out time series analysis on the denoised acoustic emission signal by using a time series method, establishing a time series model, and finally, identifying the state of the metal deep drawing part in the computer by combining a fuzzy comprehensive judgment method with maximum membership grade principle. The invention can enhance the accuracy that characteristic parameters reflect the actual working conditions can greatly reduce the sampling number on the premise of ensuring to acquire sufficient information, has the advantages of accurate and clear frequency positioning, and is suitable for the occasions with high requirements for on-line monitoring.

The invention provides a rapid random optical reconstruction imaging system based on sparse constraint and a rapid random optical reconstruction imaging method based on sparse constraint. The system at least comprises an excitation light source module for generating polychromatic excitation light, a fiber coupling and full inner reflection module which receives the polychromatic excitation light and then forms collimating excitation light, an integration module which carries out imaging, scanning, space modulation and detection on the collimating excitation light, a time sequence motor control module which controls the excitation light source module, the fiber coupling and full inner reflection module and the integration module, and an imaging collection and analysis software package which detects collected data and carries out random reconstructed image processing and compression perception processing. According to the system and the method, the incoherence of a measurement domain and an image domain is perfectly satisfied, and the sparse constraint technology can play a maximum effect. While a nanometer level imaging resolution is achieved, the magnitude order of an existing super-resolution imaging speed is raised. In the condition of maintaining an original imaging speed, the spatial resolution is greatly improved.

The invention provides a compressed video capture and reconstruction system based on structured sparse dictionary learning. The system comprises a structured sparse dictionary learning module, a video signal sensing module and a reconstruction processing module. The structured sparse dictionary learning module firstly acquires a training set through a sub-space clustering method, then, a dictionary is acquired through a linear sub-space learning method and minimized-block-relevant block sparse dictionary learning method, the sensing module projects video signals in an image block mode, and acquired data are finally decoded and reconstructed in the reconstruction processing module. Compressed sampling is provided, the distributed progressive structure of the video sampling process is combined, the reconstruction accuracy and efficiency are improved for the special structure of a structured sparse dictionary matrix, the sampling efficiency of the video signals is greatly improved, reconstruction gains are acquired compared with other methods under different sampling compression ratios, and meanwhile the good expandability is achieved.

The invention relates to a soil type merging and multiple regression-based soil manganese content prediction method. The divide-and-conquer control of different spatial variation characteristics showed by soil trace elements in different soil types is involved; the spatial heterogeneity of the soil trace elements can be detected by analyzing a discrete degree of spatial distribution of effective manganese content of soil; the multicollinearity problem in local regression analysis can be diagnosed through the local regression analysis; and especially in a spatial prediction process, a comprehensive prediction model is built by measuring spatial distribution characteristics of the effective manganese content of the soil under the sampling densities of different soil samples.