84 results about "Minimization algorithm" patented technology

A 3-dimensional scanner capable of acquiring the shape, color, and reflectance of an object as a complete 3-dimensional object. The scanner utilizes a fixed camera, telecentric lens, and a light source rotatable around an object to acquire images of the object under varying controlled illumination conditions. Image data are processed using photometric stereo and structured light analysis methods to determine the object shape and the data combined using a minimization algorithm. Scans of adjacent object sides are registered together to construct a 3-dimensional surface model.

The invention relates to a camera calibration system and method thereof, which is capable of easily performing camera calibration using a concentric circle pattern. According to the invention, a method of calibrating a camera calibrates the intrinsic parameters of the camera required to measure geometric information of an object using projection invariable characteristics of concentric circles. The method includes the steps of: taking images of the calibration pattern consisting of two or more concentric circles located in the same plane and having different radius at different angles to obtain projected images; calculating the central point of the projected images using a given algorithm, and calculating the principal point and focal point of camera using a nonlinear minimization algorithm based on the central point thus obtained.

The method electromagnetically measures a pipe inner diameter ID and a pipe ratio of magnetic permeability to electrical conductivity μ₂/σ₂ by means of a measuring arrangement 1 comprising a transmitter coil 2 and a receiver coil 3, both coils being coaxial to and longitudinally spaced from each other, the measuring arrangement 1 being adapted to be positioned into the pipe CS and displaced through the pipe. The method comprises the steps of:

• • a1) exciting the transmitter coil 2 by means of a transmitter current I_i, the transmitter current having a first excitation frequency f₁,
  • a2) measuring a receiver voltage V_i at the receiver coil 3,
  • a3) determining a transimpedance V_i/I_i between the transmitter coil 2 and the receiver coil 3 based on the transmitter current I_i and the receiver voltage V_i, and determining a measurement ratio M_i based on said transimpedance,
  • b) repeating the excitation step a1), the measuring step a2), the transimpedance and the measurement ratio determination step a3) for at least a second excitation frequency f₂ so as to define a measurement ratio vector [M₁, M₂, . . . M_n],
  • c) calculating a prediction function vector [G₁, G₂, . . . G_n] based on the first and at least the second excitation frequency, a plurality of potential pipe ratio of magnetic permeability to electrical conductivity and a plurality of potential pipe inner diameter ID, and
  • d) applying a minimizing algorithm onto the measurement ratio vector [M₁, M₂, . . . M_n] and the prediction function vector [G₁, G₂, . . . G_n] and determining the pipe inner diameter and the pipe ratio of magnetic permeability to electrical conductivity corresponding to a maximum solution of the algorithm.

The invention discloses a dynamic PET (positron emission tomography) image reconstruction method based on a Poisson TV. According to the method, PET images are reconstructed through conversion of a reconstruction problem mathematical model and an equivalence problem on the basis of a Low Rank method, and an alternate minimization algorithm is adopted during solution of the equivalence model problem when the PET images are reconstructed in combination with Poisson and Low Rank models. Therefore, the Low-Rank algorithm is utilized effectively, and problems of low resolution and noise interference of results produced during dynamic PET image reconstruction by a computer are solved; compared with existing reconstruction method through tests, the method can realize the better reconstruction effect.

A quiet zone generation technique is proposed for interference mitigation for a receive antenna by injecting the very interference signals via iterative processing, generating quiet zones dynamically for receive (RCV) antennas. The receive antenna may feature multiple receiving apertures distributed over a finite area. Optimization loops consist of four cascaded functional blocks; (1) a pick-up array to obtain the interference signals, (2) element weighting and/or repositioning processors, (3) an auxiliary transmit (XMIT) array with optimized element positions, (4) a diagnostic network with strategically located probes, and (5) an optimization processor with cost minimization algorithms.

To minimize interferences between transmit (Tx) and receiving (Rx) apertures in limited space of an antenna farm for communications and/or radar applications are very tough problems. Among the tools for solving the problems are many conventional techniques listed in the references [1,2,3]. However, solutions for co-site interference mitigation may not be generic ones but more specific to geometries of antenna farms, Tx apertures and Rx antenna locations, and beam positions of the Tx beams.

The invention discloses a compressed-sensing-based de-artifact fusion method of a high-dynamic-range image. Compressed sampling is carried out on an inputted multi-exposure image sequence; reconstruction is carried out by using a reconstruction method to obtain multi-exposure image sequence after compressed sensing; and then an image set after compressed sensing is normalized, multi-exposure imagede-artifact fusion based on PatchMatch and a rank minimization algorithm is carried out on the image set to obtain a high-dynamic-range (HDR) image of a target. According to the invention, on the basis of latest research results of K-SVD dictionary learning, compressed sensing and de-artifact fusion, the sampling rate, the storage space and the computational complexity are reduced effectively; and a de-artifact de-blurring HDR image is obtained.

The invention provides a design method of a classifier for high-precision face recognition. The design method comprises the following steps of: (1) inputting a vector type face image data set for standardized processing to obtain a face image sample set, wherein the human face image data set must contain the known classes of face image samples and can contain unknown classes of face image samples; (2) utilizing the L1-minimization algorithm to calculate the sparse representation or sparse coding of each face image sample reconstructed by face image samples except the sample; and (3) utilizing the sparse representation or sparse coding of face image samples and the classification information of the classified face image samples to build the optimal model of the classifier, and solving the regularization optimization problem to obtain a classification function. The design method provided by the invention has an explicit way of expression, so that the real-time performance in face recognition application is obviously improved, and the high-precision face recognition is realized under the condition that an image has a great number of noise pixels.

A system includes an optical power meter operable to generate an optical power signal corresponding to the optical power of a received output signal generated by an optical IQ-modulator. The system further includes a processor operable to receive the optical power signal and determine, based on a minimization algorithm and the received optical power signal, a first bias voltage to be applied to a first sub-modulator of the optical IQ-modulator and a second bias voltage to be applied to a second sub-modulator of the optical IQ-modulator. The system may also include a peak power meter operable to generate a peak power signal corresponding to the peak power of the received output signal generated by the optical IQ-modulator, wherein the processor is further operable to determine, based on a minimization algorithm and the received peak power signal, a third bias voltage to be applied to a phase shift component of the optical IQ-modulator.

The invention relates to a resource joint optimization method for the multi-target velocity estimation of a distributed MIMO radar system. The method comprises the following steps that: a target is designated, with minimizing key target velocity estimation error as an objective function, a resource joint optimization model including transmitting array elements, receiving array elements, transmitting power and signal time length is established; an optimization problem is transformed into a second-order cone programming problem, a cyclic minimization algorithm is adopted to solve four optimization variables sequentially; and after the algorithm converges, circulation is terminated, array element selection variables are dualized, an optimal array element is selected, transmitting power and signal time length are re-allocated, so that the result of resource joint allocation is obtained. The method of the invention has high flexibility for the number of tracking targets. With the method adopted, fewest transmitting array elements can be selected under a condition that the requirements of different speed estimation of a plurality of targets are satisfied, key target tracking performance can be improved, and multi-target overall tracking accuracy error is minimum. The method has a high application value.

An image data processing component (122) includes algorithm memory (212) including one or more image domain only iterative de-noising algorithms (214) based on the Huber roughness penalty minimization and a processor (206) which de-noises reconstructed image data solely in the image domain based on at least one of the Huber roughness penalty iterative minimization algorithms.

The invention provides a compression sensing and positioning method, which comprises the steps of step 1, sampling wireless signals in the space of a Wifi environment to obtain a sparse sampling matrix; step 2, recovering the sparse sampling matrix to obtain an RSSI matrix; step 3, acquiring the wireless signal information of the current location of a wireless access point, and comparing the wireless signal information with the RSSI matrix to obtain the current location. During the step 1, a sampling model is established and the environment space is sparsely sampled to obtain the sparse sampling matrix. During the step 2, the sparse sampling matrix is recovered based on the minimization algorithm L1 to obtain the RSSI matrix. According to the technical scheme of the invention, a plurality of random reference positions in the space environment are sampled in all directions for multiple times, so that the sparse sampling matrix in the space environment can be obtained. Meanwhile, the RSSI matrix that meets the positioning precision is obtained through recovering the sparse sampling matrix based on the minimization algorithm L1. Therefore, the time period for constructing the RSSI matrix is shortened.

The invention relates to a multi-label image annotation result fusion method based on a rank minimization optimization algorithm. The method comprises the following steps: (1), extracting various feature representations of a training set image, wherein the training set image has a semantic annotation word which is given in advance; (2) in the various feature representations, training each supervised learning image annotation model; (3) for a new image without a semantic annotation word, using the same method for extracting various features of the image, and using the features for being respectively input to the corresponding supervised learning image annotation models to predict multi-label results; (4) utilizing the rank minimization algorithm for fusing the multi-label results output by the various models to obtain a more accurate annotation result. According to the multi-label image annotation result fusion method based on the rank minimization, the complementarity of the image annotation models under the various feature representations is fully utilized, the rank minimization algorithm is utilized for reducing the number of prediction mistakes in the fused annotation result, and therefore the final image annotation result is more accurate.

The invention discloses a rapid and high-precision post-stack seismic impedance inversion method. The method comprises the steps of performing odd-even decomposition on a reflection coefficient sequence in a seismic convolution model to obtain a reflection coefficient sequence decomposed seismic convolution model; constructing an inversion objective function constrained by a longitudinal wave low-frequency model by use of the reflection coefficient sequence decomposed seismic convolution model and a Lagrangian multiplier method; performing inversion on the inversion objective function synchronously constrained by a longitudinal wave impedance low-frequency model by use of an L1-2 minimization algorithm to obtain odd-even component coefficients; and obtaining the longitudinal wave impedanceby use of the odd-even component coefficients.

The invention, which belongs to the technical field of microelectronics, provides a bandwidth mismatching optimization method for a multi-channel time-interleaved analog-to-digital converter. The method comprises: bandwidth parameters of signal transmission paths of all sub channels are extracted, calculation is carried out to obtain bandwidth mismatching values of all channels relative to a reference channel, a vector is formed and is used as an initial bandwidth mismatching vector, the vector is introduced into an objective function, and calculation is carried out to obtain an objective function initial solution; and elements in the bandwidth mismatching vector change randomly to generate a new bandwidth mismatching vector, the new bandwidth mismatching vector is introduced into an objective function, calculation is carried out to obtain a solution of the objective function, a minimum solution of the objective function is updated by using an energy state minimization algorithm until the updating frequency of the minimum solution of the objective function in the energy state minimization algorithm reaches a preset number of times, and then the minimum solution of the objective function is outputted. Therefore, optimization of a dynamic linear performance SFDR of the multi-channel time-interleaved analog-to-digital converter can be realized; no influence on other performances of the device is caused; and no circuit cost increase.

The invention discloses an automatic historical fitting method for a multi-scale fractured medium of a tight oil and gas reservoir. According to the method, differentiated processing is carried out oncracks of different scales, the multi-scale cracks are characterized by combining an explicit large-scale crack description method and an approximate small-scale crack region description method, namely parameterization is achieved, and an iteration minimization algorithm is adopted on the basis of a proposed parameterization method to form an automatic historical fitting method. According to themethod, the utilization efficiency of the dynamic production data of the tight oil and gas reservoir production well can be improved, the prediction precision of future production of the tight oil andgas reservoir production well is improved, and a foundation can be laid for optimization of a development strategy of the tight oil and gas reservoir.

The invention discloses a Low Rank based dynamic PET image reestablishment method. The Low Rank based dynamic PET image reestablishment method reestablishes a PET image by establishing a mathematical model of the reestablishment problem and converting an equivalence problem model based on a Low Rank method. When the Low Rank model is utilized to perform PET image reestablishment, the equivalence problem model adopts an alternating minimization algorithm in the solving process. Therefore, the Low Rank based dynamic PET image reestablishment method effectively utilizes the Low Rank method and solves the problems including low resolution and noise interference of results produced in the dynamic PET image reestablishment process of a computer. Compared with experiments of existing reestablishment methods, the Low Rank based dynamic PET image reestablishment method can obtain good reestablishment effect.

The invention discloses a weighted nuclear norm minimization algorithm based on enhanced block matching precision. The invention discloses a desert earthquake low-and-intermediate-frequency noise suppression method based on a weighted nuclear norm minimization algorithm. The noise suppression method comprises the following steps: step 1, collecting desert earthquake low-frequency noises; determining an original reference data block and a plurality of common data blocks to obtain a filtered reference data block and a plurality of common data blocks; step 2, calculating the similarity between the filtered reference data block and each common data block, selecting first m filtered common data blocks as blocks similar to the filtered reference data block to form a data matrix by combination; step 3, acquiring an approximate similar block data matrix of an ideal pure similar block data matrix from the data matrix; and S4, acquiring approximate pure earthquake low-frequency noise data afternoise suppression. In addition, the invention further discloses application of the desert low-and-intermediate-frequency noise suppression method to suppression of desert earthquake low-and-intermediate-frequency noises.