403results about How to "Solve drift" patented technology

The invention relates to a vehicle / vehicle flow detection method based on a three-axis magnetoresistive sensor, comprising the following steps: calibrating or calibrating the three-axis magnetoresistance sensor; obtaining the environmental magnetic field strength when there is no car, and obtaining the magnetic field strength when there is a car passing by; using Kaiser The FIR filter is designed to filter the measured magnetic field strength data to obtain the data sequence; the data sequence is binarized; whether there is a car is judged; when there is no car, interference is judged; when there is a car, interference is judged; the vehicle leaves the judgment; after the vehicle is judged The software forcibly resets the three-axis magnetoresistive sensor. It is characterized in that the method integrates the three-axis magnetic field information, which can fully reflect the disturbance information of the geomagnetic field when the vehicle passes by, the sensor can be placed arbitrarily, and the software reset method can effectively overcome the signal drift of the magnetic resistance sensor generated when the vehicle passes continuously Problems, through the judgment of vehicle entry, interference, and vehicle departure, it can ensure high detection accuracy and strong anti-interference ability. This method is easy to implement and is suitable for vehicle / traffic flow detection on traffic roads, bridges, squares, tunnels, etc. It can also be used in Vehicle detection in the parking lot for intelligent management.

The invention discloses a vehicle dynamic measuring device of a steel rail abrasion integrated parameter, and the vehicle dynamic measuring device comprises a visual sensor, a computer and a mileometer, wherein, a high speed image acquisition card and a measuring module are installed in the computer, the visual sensor comprises a steel rail section image system and an optical grating projector, the optical grating projector can project out more than one light planes vertical to a measured steel rail, the measuring module is used for obtaining the vertical abrasion of the steel rail, the lateral surface abrasion, the wave depth and the wave length of the wave abrasion; the invention also discloses a vehicle dynamic measuring method of a steel rail abrasion integrated parameter. The adoption of the device and the method described in the invention can improve a sampling intensity under the condition of not improving the performance of the image sensing and an acquisition hardware equipment, thereby meeting the high speed online dynamic measurement requirement of the wave abrasion, and obtaining the exact wave depth and the wave length of the wave abrasion.

The invention relates to a UAV (Unmanned Aerial Vehicle) man-machine interaction method based on binocular vision and deep learning. An embedded image processing platform and a binocular camera are carried on a UAV. The embedded image processing platform is connected to a flight controller through an interface. The UAV communicates with a ground station through the platform. The platform is provided with a Graphics Processing Unit (GPU), a convolutional neural network deep learning algorithm is run and parallel computing for an image is carried out during video capture of the camera. According to the invention, a convolutional neutral network is transplanted to the embedded platform with a dedicated GPU and the run speed becomes faster through parallel computing for speeding up the computation.

The invention discloses a pavement cracking detection system and a pavement cracking detection method based on line structured light. A near infrared laser vertically irradiates to a to-be-measured pavement, a camera takes a shoot on the light vertically irradiated to the to-be-measured pavement by the near infrared laser, light stripe images are collected and transmitted to a computer, and the computer calls an image transmission storage processing module to perform storage and pavement cracking detection, so that the problem that light stripes drift up and down along the scanning direction due to ups and downs of the pavement can be avoided, and the collection precision and the detection processing speed are high. In addition, the detection method comprises the steps that the crosswise contour line of the pavement is extracted from the light stripe image sequence, then the pavement depth image through which global three-dimensional contour information of the pavement is reflected is generated by using straight-line fitting and differentiating computation stitching, and the curve shape characteristics of cracks serve as the detection basis to be distinguished from other shape signals, thus cracking extraction is realized accurately, so that errors caused by ups and downs of the pavement or vibration of a vehicle can be effectively removed in real time, and the phenomena of false detection and missing of detection are decreased.

The gas analyzer calibrating equipment includes zero gas generator, calibration system and various kinds of gas analyzers. One gas inlet of the calibration system has one total gas intake pipe connected and the total gas intake pipe is connected to several standard gas cylinders via solenoid valve; and the gas outlet of the calibration system has one total gas outlet pipe connected and the total gas outlet pipe has the other end connected to analyzer via one three-way solenoid valve, which has normally closed port connected to the total gas outlet pipe, the normally opened port connected to atmosphere, and common port connected to the air intake port of the analyzer. The computer controls these solenoid valves via D / A converter. The equipment may be used to obtain the output slope of the analyzer data for correcting the data the analyzer obtains and obtaining the practical value of the gas component in atmosphere, solving the drift problem of analyzer.

A miniature combined attitude measuring system for small satellite is composed of a combined inertial-magnetic measuring unit, a signal processing circuit and a satellite carried computer for storing data and program. Said measuring unit has 3 sensor sets based on microelectromachinal system. Each sensor set has microgyroscope, microaccelerometer and micromagnetometer. Its advantages are high real-time performance, high reliability and high anti-interference power.

The invention relates to a deep defect quantifying method based on an eddy rotating probe signal. The method comprises the steps of 1, acquiring a calibrated defect impedance variation vector signal; 2, performing differential treatment for the defect signal; 3, normalizing the signal; 4, forming a calibration quantifying curved line; 5, acquiring an actual signal; 6, performing deep defect quantifying for the actual signal through the calibration quantifying curved line. With the adoption of the method, the problem of failure of quantifying of a rotating probe can be solved; the algorithm transformation is performed to solve the drifting problem of an absolute signal.

The present invention belongs to the field of laser precision detection technologies and laser micro-nano machining monitoring technologies, and relates to a laser micro-nano machining differential confocal online monitoring integration method and device which can be used for laser micro-nano machining and online detection of a complex micro-nano structure. A laser differential confocal axial monitoring module with high axial resolution and a femtosecond laser machining system are organically integrated, a differential confocal system curve zero point is employed to perform nanoscale monitoring of a sample axial position and measurement of a sample axial machining size to achieve real-time fixed focus of the sample axial position and high-precision measurement of the micro-nano structure size after machining, solve the drift problem and the high-precision online detection problem in the measurement process, improve the controllability of micro-nano femtosecond laser machining precisionand sample machining quality. Prior to machining, a microimaging module is utilized to perform rough alignment and horizontal position identification of the sample.

The invention relates to a dynamic control system and a control method for a downhole directional power drilling tool face. The dynamic control system is characterized in that the dynamic control system comprises a set of tool face decoding system, a set of dynamic tool face control system, a set of top drive / turntable servo and brake system, a set of top drive / turntable angle sensing system and a set of top drive / turntable angle signal processing system; and the input end of the tool face decoding system is connected with the pressure signal sensor of a measurement while drilling (MWD) system in a well drilling system, the output end of the tool face decoding system is connected with the dynamic tool face control system, the dynamic tool face control system is preset in a monitoring computer, the output end of the dynamic tool face control system is connected with the top drive / turntable servo and brake system, the top drive / turntable servo and brake system is connected with a top drive / turntable and the top drive / turntable angle sensing system, the output end of the top drive / turntable angle sensing system is connected with the top drive / turntable angle signal processing system, and the output end of the top drive / turntable angle signal processing system is connected with the dynamic tool face control system. The dynamic control system and the control method can be widely used for the dynamic adjustment control of the downhole directional power drilling tool face.

The invention provides a double FP-cavity optical fiber acoustic sensing probe and a sensing system thereof. By arranging two FP cavities whose length difference is odd times of an one-eighth working light wavelength, and eliminating the problem of operating point drift by processing two orthogonal signals acquired through utilizing the two FP cavities, the double FP-cavity optical fiber acoustic sensing probe and the sensing system thereof are characterized by good stability, high sensitivity, electromagnetic interference resistance, and the like; and the double FP-cavity optical fiber acoustic sensing probe and the sensing system thereof can be widely applied to the fields of voice communication, environmental noise monitoring, noise source locating, acoustic signal pickup in complex electromagnetic environments, and the like.

The invention relates to a fibre optic gas detector for the fire hazard of a coal gab area, comprising a light source, a probe, a photodetector and a signal processing unit, wherein light emitted by the light source is led into the photodetector by the probe so as to be converted into electrical signals, and the electrical signals are led into the signal processing unit to be processed. The fibre optic gas detector is characterized in that an optical circulator is arranged on a light path between the light source and the probe, the light source is connected with the first port of the optical circulator, the probe is connected with the second port of the optical circulator, the third port of the optical circulator is connected with a photoelectric converter, the probe is of a multi-time reflection type white cell structure, and a terminal optical fiber collimator and a fibre reflecting mirror are arranged on the exit end of the white cell.

The invention discloses a method for correcting a synchronous phasor under the condition of frequency shift of voltage of a power grid, and a current signal. By a uniform approximation method in a fonctionelle, an expression is reasonably corrected, a formula is corrected by the same method in an amplitude-phase space, and a quantized correction expression is finally given. A calculated amount is greatly simplified by an amplitude-phase characteristic relation approximation ellipse. By the uniform approximation method, the correction formula of amplitude is solved, and a practical function expression of engineering is given. Compared with the conventional algorithm, the method mainly solves the problem of 0 / 0 data shift, which is caused by conventional amplitude correction and the problem of sudden change of phase angle, which is caused by discontinuity of a tan function. The research result has certain engineering application value.

The invention discloses a path trajectory adjusting method and device. The method comprises the following steps: obtaining a real-time coordinate value of each target position in a plurality of target positions of a current path and a neighbor coordinate value of a neighbor position adjacent to each target position; adjusting the real-time coordinate value by utilizing a target filter algorithm and the neighbor coordinate value to obtain a first target coordinate value of each target position; and carrying out adjustment on the current trajectory of the current path according to the first target coordinate value of each target position to obtain a first target trajectory of the current path. Through the technical scheme, the real-time coordinate value of each target position can be subjected to initial adjustment through the target filter algorithm, and then, real-time adjustment is carried out again on the first target trajectory through one or more algorithms of a wavelet transform algorithm and a Kalman algorithm to obtain a higher-accuracy higher-precision trajectory of the current path, thereby solving the problem of GPS longitude and latitude coordinate drift.

The invention discloses an anti-blocking target tracking method based on partitioning. In combination of the quick tracking speed and the good tracking effects of a kernel correlation filter (KCF) tracking algorithm, the real-time performance of a partitioning method is ensured. The method uses a high-confidence model updating strategy, a model drift problem of the KCF in the case of blocking can be solved, the tracking speed and the tracking precision are quickened, and besides, when the partitioning method is used, after a part is blocked during the target tracking process, the position information and the reliability of the unblocked part are used reasonably, the position information after the target is blocked is determined accurately, and the robustness of tracking towards blocking is thus improved. As the method can meet the real-time performance, the method can be applied to a real scene.

The invention provides a method and device for determining movement tracks. The method comprises the steps: positions of n original track points P<1>-P<n> in a movement process of a mobile terminal are acquired, wherein the time when the mobile terminal passes through the i original track point P is earlier than the time when the mobile terminal passes through the i+1 original track point P<i+1>; the n original track points P<1>-P<n> are subjected to filtering processing, wherein filtering processing is used for deleting and / or updating the positions of at least one original track point in the n original track points P<1>-P<n>; and according to target track points obtained after the n original track points P<1>-P<n> are subjected to filtering processing, the movement tracks of the mobile terminal on at least one road in a road map are determined. According to the method and device for determining the movement tracks, a map matching process for a high-noise data acquisition mode such as base station positioning is more accurate, and the obtained movement tracks better accord with the actual condition.

The invention provides a single-fiber ultrahigh-precision time transmission method and aims to solve the problems that slow change in fiber transmission delay is uncontrollable and the extremely slow delay drift and drift accumulation cannot be solved in the existing fiber communication network time frequency synchronization technology. A single-fiber bidirectional optical transceiving module, a delay processing module and a master time synchronization device or slave time synchronization device are respectively connected to two ends of a same fiber. A cesium atomic clock is connected to the side of the master time synchronization device. Fiber delay between the master and slave time synchronization devices and a change value of the fiber delay are precisely measured and compensated. The method has the advantages that the problems that drift and drift accumulation caused by the extremely slow change in fiber transmission delay along with time, temperature, fiber ageing and the like cannot be eliminated in the prior art are solved; time synchronization precision is controlled in the range of + / -5 nanoseconds, networking can be in any form, and the method has promising market prospect and significant economic benefit.

The invention discloses a video synthesis method based on lip language synchronization and miracle adaptation effect enhancement. According to the method, the portrait and the audio stream to be synthesized are directly and integrally coded; a cyclic decoder network retaining original face information is used to decode converted abstract features into an image sequence, and then five discriminatornetworks are used to carry out adversarial training on the synthesized image sequence according to a real image sequence, so that a total reconstruction error is minimized. Compared with an existingvideo synthesis method, the method has the advantages that the continuity of face change between the front frame and the rear frame is guaranteed, the definition of face pictures in the frames is improved, meanwhile, under the action of the lip language synchronous discriminator and the miracle adaptation discriminator, the synthesized video appears more natural, and the authenticity of the visualeffect is greatly enhanced. The method has high practical value in the aspect of improving the user experience of virtual live broadcast and man-machine interaction.

The invention discloses a correcting system and method of a sunlight reflecting device based on an artificial light source. The correcting system comprises a correcting vehicle capable of freely advancing, a laser emitting device installed on the correcting vehicle, a gesture locating system of the laser emitting device, a light-sensitive array receiving screen used for receiving rays reflected by the sunlight reflecting device, a gesture locating system of the light-sensitive array receiving screen, a laser emitting device supporting mechanism, a receiving screen supporting mechanism, a distance measuring and locating system and a data processing system. According to the invention, defects such as high cost, time spending, complexity in image processing, difficulty in searching light spots and the like of a sunlight combined solar light spot imaging correcting scheme are overcome, problems of high cost and searching of the correcting system such as a high-power tower-type artificial light source are solved, and the correction of the sunlight reflecting device is realized by using a method with low power, low cost and high precision.

The invention relates to a human-machine interaction method for drones based on binocular vision and deep learning. The drone is equipped with an embedded image processing platform and a binocular camera. The embedded image processing platform is connected to the flight controller through an interface. The man-machine communicates with the ground through the platform, which is equipped with a graphics processor, runs the convolutional neural network deep learning algorithm, and performs parallel calculation on the image while the camera captures the video. The invention transplants the convolutional neural network to an embedded platform configured with a dedicated graphics processing unit (GPU), and accelerates the running speed through parallel computing.

The invention discloses a method using a mechanical arm to inhibit the attitude interference of a spacecraft pedestal. The method comprises the following steps that a many-body dynamic model is established according to configuration of a space robot, and a momentum conservation problem of the mechanical arm, an antenna and the pedestal is analyzed on the basis of the model; a track of the spatialmechanical arm is designed by considering the task priority on the basis of a dynamic balance control and zero-reaction space theory, so as to coordinate a tail-end executor task of the spatial mechanical arm and inhibit an attitude interference task of the satellite pedestal; and on the basis of the analysis, closed-loop inverse kinetic control of the spatial mechanical arm is designed, and the problem that error accumulation is increased gradually in pedestal attitude inhibition is solved. Concepts of dynamic balance control and zero reaction space are combined, and the track of the kineticredundancy mechanical arm is constructed to inhibit aircraft pedestal attitude interference caused by other rotation mechanisms.

A voltage stabilization on-line monitoring method based on a wide area measuring system is characterized by comprising the steps of defining voltage stabilization indexes, perfecting the voltage stabilization indexes, processing load and electric generator collinear nodes, building a basic framework and a processing flow for voltage stabilization on-line monitoring, and the like. Voltage weakness nodes in the system can be recognized accurately, and high robustness is achieved. The load and electric generator collinear nodes are processed by adopting an electric generator inside potential expanding method, influences on system voltage stabilization caused by electric generators and loads on the collinear nodes are taken into consideration effectively, the defect that power injection of load nodes in the collinear nodes is not taken into consideration in a traditional processing method is overcome, model precision is high, and calculation errors are small. Real-time performance and accuracy of wide area measuring information and reliability and stability of EMS network topology information are integrated, system voltage stabilization information can be obtained only by collecting phasor measuring information of the system once, the problem that repeated collection leads to identification parameter drift is solved, and voltage stabilization on-line monitoring accuracy is improved.

The invention provides a method for calibrating microscopic images, which is used for solving the problem of image drift at different multiplying factors of an optical microscope. The calibration comprises the following steps: 1, providing a calibration swatch with graduation marks and a center mark; 2, aligning the center mark of the calibration swatch with the center mark of an object lens at a certain multiplying factor; 3, switching the optical microscope to another multiplying factor and calculating the offset of the center mark of the object lens relative to the center mark of the calibration swatch; 4, locating the test point of a test piece at the multiplying factor in step 2; and 5, switching the multiplying factor to the multiplying factor in step 3, and moving the microscope stage placing the test piece according to the offset calculated in step 3. By adopting the method of the invention, the test point of the test piece can be rapidly located even under the condition that the microscopic image drifts, therefore, the problems of the low availability of a testing machine equipped with an optical microscope and the test process delay of the testing machine caused by the drift of the microscopic image are avoided.

The invention discloses a mileage statistic method based on a GPS (global positioning system). The mileage statistic method is used for acquiring information of moving speed of equipment in every second and calculating moving mileage of the equipment every second, wherein the mileage statistic method comprises the following steps: a, confirming whether the equipment is under a moving state or a static state; b, accumulating the travelled distance of the equipment under a motion state; and c, and making mileage compensation to the displacement generated when the equipment is under a static state. By continuously and repeatedly judging the moving speed of the equipment and conforming the moving state of the equipment, mileage accumulated errors caused by the phenomenon of position shift since the GPS obtains coordinate when the equipment is under the static state can be effectively avoided. Compared with the 10 percent of mileage statistic errors in the prior art, the method disclosed by the invention has the advantages that people can obtain more accurate mileage statistic information, and the errors are reduced to about 5 percent.

The invention provides a wavelength circulation tuning compensation interferential sensor operating point drift method and belongs to the field of optical fiber sensing technology. The method is characterized in that stabilization of quiescent operating points of a two-beam interferential sensor is achieved by means of a bandwidth-limited wavelength tunable laser. According to the method, the circulating tunable wavelength of the laser is controlled through a sensor feedback signal, the quiescent operating points of the sensor are tracked and compensated in real time, and long-term stable operation of the sensor is achieved. The method has the advantages that the distributed feedback semiconductor laser is adopted as a light source, a circulation tuning compensation method is used for solving the problem of operating point drift of the interferential sensor, the compensation problem that the operating points drift on a large range when the sensor is under a severe environment is solved, and system cost is reduced compared with other methods. The method is suitable for detecting vibration and acoustic waves of complex environments by means of the optical fiber sensing technology.

The invention discloses a method for carrying out panoramic positioning and orientation by utilizing a mobile phone device sensor. The method comprises the following steps: collecting a rotation angles, on three axles of an equipment coordinate system, of a mobile phone body relative to an own initial position through a gyroscope built in a mobile phone, and acquiring data through a direction sensor built in the mobile phone so as to obtain direction data, in a word coordinate system, of the mobile phone body, and weighting and fusing the rotation angles and the direction data to carry out correction; generating corresponding textures according to the corrected visual direction data, rendering and displaying the corresponding textures to the inner surface of a VR ball model, putting a camera model at the center of the ball mode, taking the camera model as a human eyeball, taking the inner surface of the VR ball model as a display screen, and synchronously changing the camera orientation according to correction direction of mobile equipment. According to the method disclosed in the invention, the problem of mobile phone gyroscope drifting is solved, the problem of upside-down rotation of virtual VR world is solved, furthermore, the orientation of equipment can be smoothly and accurately positioned, and an immersion type impression effect of panoramic videos can be obtained.

The invention relates to a single-substrate multi-chipset high-power LED (Light-Emitting Diode) encapsulation one-step bonding method, belonging to the technical field of multi-chipset high-power LED encapsulation. The invention solves the problem that the whole bonding time is too long as multiple chipsets are sequentially spliced in the traditional multi-chipset high-power LED encapsulation technology. The method comprises the following steps of: cleaning a substrate, a chip carrying platform, a multi-chip suction nozzle and a supporting platform; placing the substrate on the supporting platform; determining positions of chips to be bonded on the front of the substrate, and placing multiple chips to be bonded on the chip carrying platform corresponding to the positions of the chips to be bonded; uniformly coating soldering paste on the positions of the chips to be bonded by adopting screen printing, and then simultaneously picking up multiple chips to be bonded with the multi-chip suction nozzle; heating the back of the substrate, enabling multiple chips to align with the positions to be bonded and then attached to the soldering paste, continuously heating for 5-10s, naturally cooling to room temperature, and removing the multi-chip suction nozzle, thus the bonding process is completed. The invention provides an encapsulation bonding method.

The invention discloses a learning target tracking algorithm based on a twin network. The method comprises the following steps: 1, setting a function f to compare the feature map similarity of a template image x and a candidate image y, the maximum response value corresponding to a target position, the x being generally an image with a target as the center in a first frame, and the y being an image searched by taking an (n-1) th frame of image target as the center in an nth frame; substituting the two inputs into a convolutional neural network, learning a network parameter rho by using a convolution processing function, generating two cross-correlation mappings, and performing exhaustive search on x at y to respond to the position of a target corresponding to the maximum value; wherein animproved discriminant correlation filter layer is added between input x and cross-correlation operation in the twin network, 3, performing tracking by measuring the similarity between a template imagex and a search image y, and online fine adjustment is carried out by means of a discriminant correlation filter module during online tracking so as to ensure the tracking accuracy.

The invention provides a robust long-term tracking method based on correlation filtering and target detection, and belongs to the field of target tracking. According to the method, the credibility ofa tracking result is obtained by utilizing a depth feature vector, and a preset credibility threshold value is used for deciding whether a detector is activated or not. When the detector is activated,it will select all targets in the current frame. The most reliable result in all candidate results is obtained by utilizing a multi-stage screening mechanism. Once a new target is obtained, the confidence template is to be updated. According to the method, the interference of an environment object in tracking can be solved by regularly updating the template in tracking.

The invention claims for protection of a spatiotemporal context visual tracking method with fusion of particle filtering. According to the method, first frame of image information is initialized, experimental parameters are set, and then a rectangular region in which a first frame of target is located is automatically selected; whether the target is blocked is determined by using a Bhattacharyya coefficient as a base; and when the target is blocked, a location and a motion track of the target at a follow-up image frame are estimated and predicted by introducing a particle filter algorithm, thereby realizing precise tracking of the target. Therefore, the method is suitable for visual target tracking under complicated background like illumination changing, target rotation, and background region interference; the robustness of target blocking is high and the real-time requirement is met.