154results about How to "Processing small amount of data" patented technology

The invention belongs to crossing field of computer vision and intelligent robots and discloses a variational mechanism-based large-area indoor scene reconstruction method. The method comprises the following steps: step 1, acquiring calibration parameters of a camera, and building an aberration correcting model; step 2, building a camera position and gesture depiction and camera projection model; step 3, utilizing an SFM-based monocular SFM (Space Frequency Modulation) algorithm to realize camera position and gesture estimation; step 4, building a variational mechanism-based depth map estimation model, and performing solving on the model; and step 5, building a key frame selection mechanism to realize three-dimensional scene renewal. According to the invention, an RGB (Red Green Blue) camera is adopted to acquire environmental data, and a variational mechanism-based depth map generation method is proposed through utilizing a high-precision monocular positioning algorithm, so that quick large-area indoor three-dimensional scene reconstruction is realized, and problems of three-dimensional reconstruction algorithm cost and real-time performance are effectively solved.

The embodiment of the invention discloses a three-dimensional map rendering and display method. The method includes the following steps that: a model file corresponding to a current scene model is called according to map position information of a current view point; data in the model file are cached in the form of an octree structure; based on the combination of the octree structure and the frustum of the current view point, visibility judgment of level of detail (LoD) is performed on all nodes inside the current scene model through adopting a preset visibility judging algorithm, so that visible nodes and child nodes of the model file which enter the view field range of the frustum of the current view point can be marked; and a scene which enters the view field range of the frustum of the current view point can be rendered and displayed according to the marked visible nodes and child nodes. The invention also discloses a three-dimensional map rendering and display system. With the three-dimensional map rendering and display method and system of the invention adopted, rendering efficiency can be improved, and network traffic can be reduced, and large-scale point cloud data can be displayed for users through Web three-dimensional technologies.

The invention discloses a terahertz communication synchronization method, which belongs to the field of communication signal processing. According to the invention, a baseband signal inserted into a pilot frequency is upconverted to a terahertz band and transmitted to a wireless channel; after a wireless signal passes through a low noise amplifier, the first time of analog quadrature mixing is carried out to filter out high frequency components; after filtering is carried out again through a narrowband filter, downsampling, squared demodulating and fast Fourier transform are carried out in order; the maximum peak position is searched to find the corresponding frequency offset; the frequency offset is fed back to a local oscillator, so that the carrier frequency of a receiving end and the carrier frequency of a received signal are dynamically synchronized; the baseband signal with offset is acquired; the second time of analog quadrature mixing is carried out to filter out high frequencycomponents; sampling and fast Fourier transform are carried out; the signal peak is observed to acquire the estimated value of the frequency and the phase offset; and the frequency and phase of the sampling clock of a high-speed analog-to-digital converter are compensated to realize terahertz communication synchronization. According to the invention, the signal processing difficulty and hardwareoverhead can be reduced.

The invention discloses a pseudo-code auxiliary carrier synchronization implementation method for terahertz communication, and belongs to the field of communication signal processing. The method comprises the following steps: inputting a data baseband signal with a low-power spread spectrum code pseudo-random sequence pilot inserted into a digital to analog converter DAC by a sending end, and after analog mixing is carried out on analog signals output by the DAC and a terahertz carrier, sending the analog signals to a wireless channel by an antenna; using local two paths of orthogonal terahertz carriers to mix with the received signals to obtain I and Q two paths of baseband signals by a receiving end; using a narrowband low-pass filter to filter the I and Q two paths of signals after orthogonal mixing to extract a pilot part; calculating the pilot carrier frequency offset, that is the data signal carrier frequency offset in real time and accurately by capturing and tracking the pilotpart; and continuously modifying the frequency of the local terahertz carrier by using the frequency offset to complete the carrier synchronization. The pseudo-code auxiliary carrier synchronization implementation method for the terahertz communication provided by the invention can reduce the signal processing calculation amount, and reduce the complexity of the hardware circuit design and implementation.

The invention discloses a detection method for radar target tracks on basis of correlation speeds, and relates to the technical field of radars. The method includes the first step of using echo data to obtain amplitude values at corresponding resolution cells, the second step of setting a primary constant false-alarm detection threshold to carry out filtering on the echo data and obtaining the amplitude values after filtering is conducted, the third step of setting value functions of the resolution cells, the fourth step of calculating value functions of speed correlation track recursion accumulation of the accumulated echo data of each frame, the fifth step of setting a secondary detection threshold and obtaining the resolution cells where targets exceeding the threshold are located, the sixth step of conducting backtracking processing on the data to obtain the resolution cells where frames of all tracks are located, and the seventh step of determining true target tracks from all the tracks. A correlation speed algorithm is introduced, the processing data size is reduced, mistaken correlation of the target tracks is reduced, false tracks are reduced, and mechanical multi-target tracking under low an SNR is achieved.

The invention relates to a television set remote control method for separating a user interface. The method includes the following steps that an information coding module acquires user interface information from a television set and codes the user interface information; a first wireless communication module of the television set transmits the coded user interface information to a second wireless communication module of an intelligent terminal; an information decoding module of the intelligent terminal acquires the user interface information received by the second wireless communication module and decodes the user interface information; a user interface operating module of the intelligent terminal generates and displays the user interface by the utilization of the decoded user interface information; the user interface operating module carries out corresponding response and operation after receiving user operation; an instruction interpretation module of the television set interprets operation instructions received by the first wireless communication module, breaks the operation instructions up into one or a plurality of steps and instructs a core function module of the television set to execute the steps in sequence; the core function module of the television set executes the operation instructions in all the steps in sequence. A system using the television set remote control method for separating the user interface comprises the television set and the intelligent terminal, wherein the intelligent terminal is provided with a television set user interface generating program and can communicate with the television set wirelessly.

The invention provides a springback compensation method for creep age forming. The method comprises the following steps: Step S100, based on MSC.MARC commercial finite element software, constructing creep age forming finite element simulation; Step S200, weighting a springback iterative compensation process based on a molding surface of a finite element simulation result; and Step S300, considering a reverse compensation process of high temperature molding surface thermal expansion effect. The method considers the influence of the strength of a component on surface compensation, the influenceof a component feature structure on surface compensation, and reverse compensation of thermal expansion of a high temperature molding surface. A component formed by the model can still meet the engineering error requirement after being cooled.

The invention provides an image composition method and an electronic apparatus. The image composition method comprises the following steps that under N different focal lengths, N images of a first photographed surface of an object are respectively obtained, each image in the N images is corresponding to one focal length in the N different focal lengths respectively, and each image has S pixels; S pixels of which the positions differ from each other are determined in N * S pixels in the N images; P focal lengths corresponding to P images where the S pixels of which the positions differ from each other are located are obtained; based on the P focal lengths, S pieces of depth information of the S pixels of which the positions differ from each other are obtained; based on the S pieces of depth information, the depth image of the object at the first photographed surface is constructed; and the j-th pixel in the S pixels of which the positions differ from each other is obtained. The j-th pixel in the S pixels of which the positions differ from each other is obtained through the following steps that first feature information of the k-th pixel on each image in the N images is contrasted; and the j-th pixel of which the first feature information satisfies a preset condition is determined from N k-th pixels.

The invention discloses a method for shooting by rotating a viewing frame. The method comprises a step of acquiring a center coordinate of the viewing frame, a step of rotating a picture, a step of calculating a rotating angle beta, a step of calculating the polar coordinate system angles of four corners of the viewing frame which is rotated by the angle beta, a step of calculating the relative coordinates of a Descartes rectangular coordinate system of the four corners of the rotated viewing frame, a step of calculating the absolute coordinates of the four corners of the rotated viewing frame, a step of drawing a rectangular viewing frame, a step of calculating the minimum enclosing rectangle of the rotated viewing frame, a step of tailoring an image in the minimum enclosing rectangle for the first time, a step of rotating the image reversely by an angle of beta with the viewing frame as the center, and a step of tailoring the image for the second time with the center of the view frame as the center, with W as the width and with H as the height. According to the method, the viewing frame within a video area can be moved and rotated, the head or the body of a shooting target can be corrected in a dynamically tilting mode, and the purposes that the shooting area is accurately defined and the shooting process is accelerated are achieved.

The invention discloses a gymnastics auxiliary training method based on human body action skeleton information, which comprises the following steps: acquiring gymnastics action skeleton data made by professionals and learners by adopting an optical action capture system, storing the gymnastics action skeleton data in a sequence form, and establishing a reference database by taking the professionals' data as a reference sequence; and comparing the action skeleton information of the learner with the reference action skeleton information to obtain action similarity and wrong postures. According to the method, the action information of the human joint part can be quickly recognized; the gymnastic training feedback system can perform real-time action evaluation and gymnastic training feedback,can perform scientific and considerable evaluation on actions of gymnastic learners, discover action problems, find and correct wrong postures, optimize and improve action standards, obtain an improved direction, evaluate and guide the actions of the gymnastic learners, and improve the gymnastic learning effect of the learners. The method provided by the invention has a certain engineering application value.

The invention discloses a method for reconstructing vertical profiles of airborne multi-scan meteorological radar targets. The vertical profiles of the meteorological targets at the fronts of aircrafts can be displayed by the aid of meteorological echo information acquired by airborne multi-scan meteorological radar, and altitude-dimension intensity change and distribution ranges of reflectivity factors of meteorological target radar can be quickly, accurately and comprehensively reflected. Besides, at least eight layers of meteorological target information are acquired by the aid of the airborne meteorological radar in multi-scan work modes, and metrological target reflectivity factor change rule statistical models are innovatively built by means of statistics by the aid of metrology in information reconstruction procedures; the vertical profiles of the meteorological targets right ahead routes of the carrier aircrafts can be truly and completely represented by the aid of the multi-scan echo information. The method has the advantages of simplicity and convenience in operation. In addition, the quantities of processed data are low as compared with the traditional method, and the method has low requirements on systems and is high in reconstruction precision.

The invention discloses an evaluation method for searching and optimizing circumferential equal division hole position degree based on polar angle dichotomy, and the method comprises: processing original position information acquired by a measuring device, fitting to obtain the actual hole center position of a part, and calculating to determine the initial pole angle difference of each hole and the rotatable angle range on the premise that the position degree error requirement is met; searching and optimizing is performed through dichotomy, optimal position degree errors of all equal divisionholes meeting the principle of the minimum condition can be obtained rapidly and accurately without tangential reference selection under the condition of undetermined tangential offset directions of the circumferential equal division holes, meanwhile, the whole searching and optimizing process is small in processing data size, and the calculation efficiency is greatly improved.

The invention discloses an automatic site selection method and device for large-scale emergency facilities and a computer readable storage medium. The method comprises the following steps: based on urban road network data of a to-be-planned area, calculating the shortest passing path among all road network nodes in the to-be-planned area; encoding the shortest path and a pre-constructed facility site selection model, wherein the facility site selection model indicates that emergency facilities set based on an emergency facility layout rule need to meet site selection conditions in the site selection process; and based on the coded shortest path, calculating the optimal solution of the facility site selection model by using a genetic algorithm to obtain the set total number and the set position information of the emergency facilities to be subjected to site selection in the to-be-planned area, and taking the set total number and the set position information as the site selection information of the emergency facilities for automatic output. On the premise that the setting number of the emergency facilities does not need to be determined in advance, site selection can be conducted onthe emergency facilities which need to balance the contradiction between the safety and economy of the facilities and are large in scale, and the layout of the urban emergency facilities is optimized.

The invention belongs to the technical field of computer image processing, and in particular relates to an image edge detection method and system for power facilities, including: acquiring image data of power facilities, image acquisition for image preprocessing, including image grayscale, image enhancement, histogram Equalization, median filtering and improved threshold segmentation algorithm are used for image segmentation, and then edge detection is performed through a double-structured structure forest improved algorithm; the combination of threshold segmentation algorithm and traditional edge detection operator can quickly and Clearly detect the ice-covered edge of the transmission line, and can ignore the texture inside the transmission line; use structured learning input for the image information, which can not only ensure the reduction of the amount of processed data, but also effectively distinguish the Whether the information is edge information, this not only improves the speed of image edge detection, but more importantly, improves the accuracy of the edge detection algorithm.

The invention discloses a robotized coal machine fault audio recognition and diagnosis method, which comprises the following steps: preprocessing collected coal mine equipment sound information to obtain preprocessed sound information with a plurality of short-time sound frames; obtaining a spectrogram containing a plurality of spectrogram frames and corresponding to the preprocessed sound information; and inputting the spectrogram into the trained hybrid neural network model, and outputting a fault diagnosis result, wherein the hybrid neural network model comprises an ALEXTet network model, an LSTM network model and a Softmax classification layer. According to the invention, the abnormal sound of coal mine equipment is recognized through a hybrid neural network model, wherein an ALEXTet network model simplifies a traditional CNN convolutional layer and reduces the algorithm complexity; and a LSTM network model is adopted to circularly collect the image sequence, learn and memorize the sequence correlation information, and the single image information and the sequence correlation information are combined to carry out discrimination, so that the accuracy and robustness of abnormal sound recognition of the underground coal mine equipment are improved.

The invention discloses a defect positioning-oriented code searching method. The method comprises the following steps of: firstly, constructing a source code file abstract syntax tree corresponding toa bug report in a software history library; secondly, extracting all code diff files related to bugs, defining continuous code modification lines in the code diff files as chunks, and establishing achunks relation graph of each code diff file; then, obtaining an importance score detree of each node in the chunk relation graph; then, the relation between the bug report and the chunk is established; Bug-is established by using a knowledge base establishing tool; A chunk knowledge base; then, for a new bug report, obtaining the new bug report and the new bug-, wherein the similarity sim of thebug report in the chunk knowledge base; and finally, generating a chunk list by combining the importance score detree and the similarity sim of the chunk to realize defect positioning. The processingobject takes the chunk as a unit, and is oriented to a continuous code block, so that the processing data volume is reduced, the chunk is a modified code line, the pertinence is better in the calculation process, and the pertinence and the accuracy of bug positioning are improved.

The invention relates to a CT image based automatic extraction method of a 3D liver bounding volume. Volume measurement is carried out on liver segmentation data of N abdominal CT, the 3D CT slice anda liver mask layer in which the volume approaches the average liver volume most are selected, and a rectangular liver area frame corresponding to the CT slice is determined; a CT image of the determined rectangular liver area as a standard CT image is registered with a CT image of a non-determined liver bounding box roughly, the step and iteration frequency of an optimizer in rough registering are changed, and a transformation matrix is output; and geometric transformation is carried out on a binary image of the rectangular liver area frame, the matrix after transformation is placed in a 2D coordinate system, coordinates of all points whose grayscale values equal 1 are searched from the matrix, and a rectangular boundary is determined according to X and Y maximums and X and Y minimums, and serves as a liver area rectangle. The data bulk of image processing is reduced, rapid and accurate liver segmentation is guided, and the method has great significance in guiding segmentation of a liver tumor and provides basis for subsequent researches.

The invention discloses an image super-resolution reconstruction method, an image super-resolution reconstruction device, electronic equipment and a computer readable storage medium. The image super-resolution reconstruction method comprises the steps: obtaining a preview image; if a to-be-processed target exists in the preview image, segmenting the preview image to obtain a to-be-processed imagecontaining the to-be-processed target and a scene image not containing the to-be-processed target, wherein the to-be-processed target is a target meeting a preset condition; carrying out super-resolution reconstruction on the to-be-processed image to obtain a processed image; and fusing the processed image with the scene image to obtain a new preview image. Through the scheme of the invention, thedefinition of the shooting target can be improved in a targeted manner, and meanwhile, the processing data volume of the electronic equipment can be reduced.