1177results about How to "Reduce cumulative error" patented technology

The invention discloses a positioning method and system based on visual inertial navigation information fusion. The method comprises steps as follows: acquired sensor information is preprocessed, wherein the sensor information comprises an RGB image and depth image information of a depth vision sensor and IMU (inertial measurement unit) data; external parameters of a system which the depth visionsensor and an IMU belong to are acquired; the pre-processed sensor information and external parameters are processed with an IMU pre-integration model and a depth camera model, and pose information isacquired; the pose information is corrected on the basis of a loop detection mode, and the corrected globally uniform pose information is acquired. The method has good robustness in the positioning process and the positioning accuracy is improved.

The invention discloses a machine vision and inertial navigation fusion-based mobile robot motion attitude estimation method which comprises the following steps of: synchronously acquiring a mobile robot binocular camera image and triaxial inertial navigation data; distilling front/back frame image characteristics and matching estimation motion attitude; computing a pitch angle and a roll angle by inertial navigation; building a kalman filter model to estimate to fuse vision and inertial navigation attitude; adaptively adjusting a filter parameter according to estimation variance; and carrying out accumulated dead reckoning of attitude correction. According to the method, a real-time expanding kalman filter attitude estimation model is provided, the combination of inertial navigation and gravity acceleration direction is taken as supplement, three-direction attitude estimation of a visual speedometer is decoupled, and the accumulated error of the attitude estimation is corrected; and the filter parameter is adjusted by fuzzy logic according to motion state, the self-adaptive filtering estimation is realized, the influence of acceleration noise is reduced, and the positioning precision and robustness of the visual speedometer is effectively improved.

The invention relates to an indoor positioning method of an intelligent mobile phone. The indoor positioning method is technically characterized by comprising the following steps of 1, in a walking process of a pedestrian, performing pedestrian dead reckoning, such as judging a moving state of the pedestrian, calculating the number of generated steps, calculating a walking distance and reckoning the direction change of the pedestrian through an acceleration sensor and a gyroscope which are arranged in the intelligent mobile phone, thus obtaining reckoned position information; and 2, constructing an indoor environment magnetic field map, performing magnetic field matching by a particle filtering algorithm according to information which is acquired by the indoor environment magnetic field map and a magnetic field sensor in real time, and correcting the reckoned position information so as to obtain a final position of the pedestrian. According to the indoor positioning method, the robustness of indoor positioning is improved from two aspects of a method for resisting various handheld mobile phones and a method for processing different step lengths; the matching precision is improved by improving the stability of a magnetic field matching judging and determining weight and reducing a system accumulation error, so that long-time accumulated errors are avoided; the system complexity is reduced by reducing the amount of external equipment.

The invention discloses an indoor scene 3D reconstruction method based on Kinect and solves the technical problem of the real-time reconstruction of an indoor scene 3D model and avoidance of excessive redundant points. The method comprises steps of: obtaining the depth data of an object by using Kinect and de-nosing and down-sampling the depth data; obtaining the point cloud data of a current frame and calculating the vector normal of each point in the frame; using a TSDF algorithm to establish a global data cube, and using a ray casting algorithm to calculate predicted point cloud data; calculating a point cloud registration matrix by using an ICP algorithm and the predicted point cloud data, fusing the obtained point cloud data of each frame into the global data cube, and fusing the point cloud data frame by frame until a good fusion effect is obtained; rendering the point cloud data with an isosurface extraction algorithm and constructing the 3D model of the object. The method improves the registration speed and the registration precision, is fast in fusion speed and few in redundancy points, and can be used for real-time reconstruction of the indoor scene.

The invention discloses a multi-source fusion unmanned aerial vehicle indoor and outdoor positioning method and system. The system comprises a sensing layer and a decision making layer, wherein the sensing layer comprises a plurality of sensors, the decision making layer comprises a visual odometer system, a laser odometer system, an IMU odometer system and two multi-source fusion systems. IMU information is selected, relatively good initial gesture and gravity direction are provided, and the scale problem of a visual odometer and the inherent defect of scale drift are made up; high accuracy information of a laser radar and the visual odometer are fused, the motion estimation error is reduced, and the system robustness is improved; and with the addition of GPS absolute position correction, and an accumulated error of the system is further reduced. By the method, accurate positioning information can be generated according to true motion track of a vehicle, the positioning accuracy is high, indoor and outdoor seamless switching can be achieved, the positioning stability of the vehicle in an indoor and outdoor environment is improved, high robustness is achieved, and the system is suitable for indoor and outdoor positioning of a pilotless automobile in a large scene.

The present invention provides an improved closed-loop detection algorithm-based mobile robot vision SLAM (Simultaneous Location and Mapping) method. The method includes the following steps that: S1,Kinect is calibrated through a using the Zhang Dingyou calibration method; S2, ORB feature extraction is performed on acquired RGB images, and feature matching is performed by using the FLANN (Fast Library for Approximate Nearest network); S3, mismatches are deleted, the space coordinates of matching points are obtained, and inter-frame pose transformation (R, t) is estimated through adopting thePnP algorithm; S4, structureless iterative optimization is performed on the pose transformation solved by the PnP; and S5, the image frames are preprocessed, the images are described by using the bagof visual words, and an improved similarity score matching method is used to perform image matching so as to obtain closed-loop candidates, and correct closed-loops are selected; and S6, an image optimization method centering cluster adjustment is used to optimize poses and road signs, and more accurate camera poses and road signs are obtained through continuous iterative optimization. With the method of the invention adopted, more accurate pose estimations and better three-dimensional reconstruction effects under indoor environments can be obtained.

Embodiments of the invention provide a character recognition model training method, device and system, a character recognition method, device and system, and a storage medium. The character recognition model training method comprises the steps of acquiring at least one sample image and labeled data corresponding to the at least one sample image, wherein the labeled data corresponding to each sample image is used for indicating the position of an actual character region of the sample image and actual characters in the actual character region; for each sample image in the at least one sample image, inputting the sample image into a region positioning network in a character recognition model to determine the position of a predicted character region of the sample image; inputting image information corresponding to the position of the predicted character region into a character recognition network in the character recognition model, so as to identify predicted characters in the predicted character region; and training the character recognition model by at least using the labeled data corresponding to the at least one sample image and the predicted characters. The accumulative error of character recognition can be effectively avoided.

The invention relates to an indoor positioning method based on fuzzy theory and multi-sensor fusion. The indoor positioning method is technically characterized by comprising the steps of according to multiple sensors arranged in an intelligent mobile phone, acquiring data of an acceleration sensor, a gyroscope and a barometer of the intelligent mobile phone firstly, and then the fuzzy theory is applied to the acquired data, speculating the motion state, step number and step length of a pedestrian so as to obtain the calculated position information; meanwhile, acquiring the user height from the information of the barometer; and establishing an indoor environment magnetic field map, acquiring information in real time by utilizing the indoor environment magnetic field map and magnetic field sensors, matching magnetic fields through the fuzzy theory and correcting the calculated position information to obtain the final position of the pedestrian. According to the method, the indoor positioning robustness is promoted through two aspects of multiple intelligent mobile phone carrying methods and processing of different step lengths. Through the matching of the fuzzy theory of the magnetic field, the accumulated errors can be reduced, the accuracy is improved, and the system complexity can be lowered.

The present invention provides a reconstruction method and device of a three-dimensional scene. The method comprises the following steps that: the first camera attitude of a current frame is obtained according to the depth image of the current frame and the depth image of a reference frame; the second camera attitude of the current frame is estimated according to the gray-scale image of the current frame and the gray-scale image of the reference frame; the fused camera attitude of the current frame is obtained according to the first camera attitude of the current frame and the second camera attitude of the current frame; and a three-dimensional scene model corresponding to the current frame is generated according to the depth image of the current frame and the fused camera attitude of the current frame. According to the reconstruction method of the three-dimensional scene provided by the embodiments of the invention, depth information-based first camera attitude estimation and gray-scale information-based second camera attitude estimation are integrated, so that the cumulative error in the camera attitude estimation can be reduced, and therefore, the reconstruction method can conduct more stably in different scenes, and the reconstructed three-dimensional scene is more accurate.

The invention discloses an inertial vision integrated navigation method based on an optical flow method, relates to an inertial vision integrated navigation method, and has the purpose of solving theproblems that the navigation information is inaccurate due to cumulative errors caused by the existing inertial navigation working for a long time, a mask interruption or an excessive dynamic error may happen when the existing satellite navigation is indoor or shielded by a building, and the algorithm of the existing vision navigation is complex and susceptible to camera pose, illumination changes, image noise and the like. The method comprises the following processes: 1, defining a coordinate system; 2, carrying three sensors of an IMU (Inertial Measurement Unit), a camera and an altimeter onthe drone to obtain two-dimensional optical flow data; 3, obtaining the position of the drone in the world coordinate system; 4, performing inertial navigation according to the IMU measuring information to calculate the position and attitude of the drone in the world coordinate system; and 5, obtaining the position and attitude information of the merged drone in the world coordinate system. The method is used in the technical field of autonomous navigation of drones.

The invention discloses a method for quickly measuring topography by using a ground laser scanner based on CORS (Continuous Operational Reference System) and ICP (Iterative Closest Point) algorithms. The method comprises the following steps: 1, coaxially connecting a GPS (Global Position System) antenna with the scanner; 2, mounting the scanner at the top of a vehicle; 3, scanning in 360 degrees through the scanner during operating; 4, synchronously measuring a topocentric geodetic coordinate of the scanner by using a GPS carrier phase dynamic real-time difference method on the basis of CORS; 5, partitioning large measuring regions into blocks along roads and rivers; 6, roughly registering measurement blocks by using public planimetric points between measuring stations; 7, carrying out multi-station cloud accurate splicing in the measurement blocks by using an iterative closest point algorithm from a point to a tangent plane; 8, converting point cloud of the measurement blocks to the geodetic coordinate by using the topocentric geodetic coordinate; 9, dividing a three-dimensional point cloud into ground points and non-ground points and generating a contour line by using the ground points; 10, measuring and drawing a terrain landform line graph by using a method of collecting man-machine interaction details and slicing point cloud based on the three-dimensional point cloud; and 11, carrying out field investigation and plotting, mending and measuring, and editing and finishing a terrain line graph to obtain a topography result.

The invention discloses a method and system for performing three-dimensional reconstruction of an indoor complete scene based on a consumer-level depth camera. The method comprises: a depth image is obtained and adaptive bilateral filtering is carried out on the depth image; with the filtered depth image, visual odometer estimation is carried out, segmentation is carried out on an image sequence automatically based on a visual content, and closed-loop detection is carried out between segments, and global optimization is carried out; and according to the optimized camera trajectory information, weighted volume data fusion is carried out to reconstruct a three-dimensional model of an indoor complete scene. According to the method and system disclosed by the invention, edge keeping and denoising of a depth image are carried out by using an adaptive dual-side filter algorithm; because of the automatic segmentation algorithm based on a visual content, an accumulation error in visual odometer estimation can be reduced and the registration precision can be improved; and on the basis of the weighted volume data fusion algorithm, the geometric details at the surface of the object can be kept. Therefore, a technical problem of improving three-dimensional reconstruction precision in an indoor scene can be solved and thus a complete, accurate, and refined indoor scene model can be obtained.

The invention a MEMS gyroscope zero-bias self-calibration method based on driving/detecting modal inversion, and the MEMS gyroscope zero-bias self-calibration method is capable of solving problems inthe prior art that drift of fully-symmetrical MEMS gyroscopes is caused and power-on zero-bias of a plurality of times is different. According to the MEMS gyroscope zero-bias self-calibration method,gyroscope working characteristics are taken into consideration, a gyroscope dynamic model is taken as a base, the change rules of two working modes before and after driving modal and detecting modal inversion are given, switch of the circular gyroscope between the two working mode is realized using a signal processing circuit. The difference of the detection signals at the adjacent working modes is obtained, so that MEMS circular gyroscope zero-bias self-calibration is realized at last. The MEMS gyroscope zero-bias self-calibration method is capable of extracting zero-bias signals effectively,inhibiting zero-bias drift, and increasing the precision of the MEMS circular gyroscopes.

The invention discloses a light screen device with sending and receiving functions and a test method of the light screen. A light source and a receiving light device are installed on a circuit board to form a detection device main body with sending and receiving functions, and the light is reflected by an illuminator to form a detection screen. The detection device and the reflecting device are both provided with angle regulatory mechanisms with the same structure, the angle regulatory mechanism can adjust the angle of an incidence screen and a reflecting screen to ensure the angle between two screens by cooperating with the light screen. The main light axis of the light-emitting device is ensured to be at the same plane with the main light axis of the receiving device. The light screen device has the advantages of simple structure, fewer links, small cumulative error, high measurement precision, and convenient assembling; the light screen device not only can be used for measuring the speed of a pellet, but also used for carrying out the denseness degree test and the measurement of pellet flying angle and direction; and the light screen device broadens the measurement scope of the pellet.

The invention discloses a fast image splicing method of matching strategy fusion and low errors. The method includes: S1, acquiring a plurality of to-be-spliced sequence images, and carrying out preprocessing and calibration of overlapped areas between the images thereon; S2, extracting SIFT feature points In the overlapped area, and calculating SURF feature descriptors; S3, completing feature rough-matching, utilizing rough-matching information to complete sequence image sorting, then adopting a random sample consensus algorithm to refine matching results, and fitting homography transformation matrices between the images; S4, selecting an image of a middle position of the sequence images to use the same as a reference image, and establishing a projection transformation model of any otherimage to the reference image; and S5, using the projection transformation model and combining multiple threads to project all the images to reference-image coordinate space after brightness correction, and synthesizing a high-quality panoramic-image after image fusion. The method greatly increases a correct-matching rate, can fast synthesize the high-quality panoramic-image of low errors, and hashigher practical values.

The invention discloses a machining process of middle-long thin-wall blades, which belongs to the technical field of machining of blades. The machining process is implemented by a detailed process of integrated rough machining and integrated fine machining, wherein the integrated rough machining comprise the specific steps of rough machining of each surface of a blade tip, each surface of a blade root and a molded surface of a steam passage; and the integrated fine machining comprises the specific steps: semifinishing of each surface of the blade tip and each surface of the blade root, fine machining of each surface of the blade tip, each surface of the blade root and each residue, semifinishing of the steam passage and fine machining of the steam passage. According to the machining process of the middle-long thin-wall blades, the operation is simple and convenient, the use is convenient and fast, the cost is low, the machining efficiency is high, and the operation is easy and labor-saving; and the deformation of the middle-long thin-wall blades is controlled within 0 to 0.8mm, the normal machining of a subsequent mold correction process is facilitated, and the machining of each blade is finished once.

The invention discloses a splicing detection device of a circular Hartmann-Shack wave-front sensor based on a minor caliber. The device is characterized by comprising the Hartmann-Shack wave-front sensor, an x-z 2D electric control translation stage, a step motor controller, a computer, a mirror surface to be detected and a data acquisition card; the Hartmann-Shack wave-front sensor is positioned behind the mirror surface to be detected to detect the mirror surface to be detected, and a facula lattice is formed on the Hartmann-Shack wave-front sensor, and the data acquisition card acquires the facula data and transmits the data to the computer for storing; the computer sends an instruction to the step motor controller, and controls the 2D electric control translation stage to move along an x axis and a z axis to scan and detect the mirror surface to be detected; the data acquisition card sequentially acquires the facula data of a wavelet surface of each frame, and then a wave surface to be detected is obtained by a centroid algorithm, a splicing method and a restoration algorithm. The device and the method help improve a formula by ignoring a defocus error in theoretical analysis, eliminate a principle error of a translation error, improve splicing precision and lower computation complexity of solving splicing parameters.

The invention discloses a prediction method of a crack propagation life of an aircraft body. The prediction method comprises the following steps: obtaining a random loading spectrum of an aircraft in flight according to the practical flight situation of the aircraft; dividing the obtained random loading spectrum into a plurality of sections, and respectively performing rain-flow counting and constant amplitude on each section, so that each section is approximated as a block loading spectrum; according to an Miners progressive damage theory, calculating a cycle life lambda effected on a crack of each block loading spectrum under the state of a corresponding stress ratio R, performing a fatigue test under the suitable stress ratio, and obtaining a corresponding stress-life curve; according to an equal defect theory, selecting the correspondence life curve under the suitable stress ratio R in the stress-life curve, and obtaining the equivalent stress Seq corresponding to the same life by a point tracing method; then, according to the R and the Seq, obtaining a maximum stress value sigma max and a minimum stress value sigma min of the equivalent cyclic stress; according to parameters of the equivalent cyclic stress, calculating the propagation rate of the crack when the obtained block loading spectrum is acted on the aircraft body.

The invention relates to a six-freedom-degree industrial robot with ball screw pair transmission. Waist whole circle rotating is achieved through a waist swing mechanism, a shoulder swing mechanism and an elbow swing mechanism form a two-freedom-degree parallel mechanism to achieve two-dimensional horizontal moving, and ball screws are driven by motors arranged on a swing platform respectively. Three mutual-orthometric rotating shafts connected through a front arm rotating mechanism, a wrist swing mechanism and a tool clamping base swing mechanism are driven independently by three motors respectively, and three-freedom-degree gesture adjusting of an executing tail end tool is achieved. The advantage of the large working space of a serial mechanism is kept, ball screw transmission is used, the rigidness of the mechanism is guaranteed, the positions of the driving motors are lowered, good dynamic characteristics are obtained, bearing capacity is large, high-speed or superspeed operation can be easily achieved, and the six-freedom-degree industrial robot has the advantages that side gaps do not exist, rigidness is high, micro-motion feeding can be conveniently achieved, accumulated errors are small, motion accuracy is high, positive and athwart kinematics solving is convenient, and control is convenient.

The invention discloses an automatic thickness control system and method of a blown film. The automatic thickness control system of the blown film comprises a film thickness detection module, a film thickness control module, a film cooling and forming mold with a plurality of brushless direct-current motors as well as a control panel. According to the invention, an FPGA (field programmable gate array) control chip of a programmable logic device is utilized to achieve a fuzzy control algorithm so as to obtain the speed control quantity of the brushless direct-current motors; by combining a double-closed-ring control algorithm for the brushless direct-current motors a control signal output by the FPGA control chip is changed and adjusted correspondingly when the thickness of the film is changed; the thickness uniformity of the films at different circumferential parts can be controlled by acquiring horizontal thicknesses of the films at controlled subzones and the positions of air flue openings corresponding to the films via transmission and decoupling control, so that the control requirement on horizontal thickness uniformity of the films can be met; the automatic thickness control system and method of the blown film are high in integrity, small in volume, good in robust performance, stable and reliable in operation, high in expansibility, high in control precision, low in accumulative error, and high in adaptability.