289results about How to "Fast precision" patented technology

Systems and methods are disclosed for vehicle motion planning where a sensor, such as a ladar system, is used to detect threatening or anomalous conditions within the sensor's field of view so that priority warning data about such conditions can be inserted at low latency into the motion planning loop of a motion planning computer system for the vehicle. The ladar system can perform compressive sensing to target the field of view with a plurality of ladar pulses.

A ladar system can estimate intra-frame motion for an object within a field of view of the ladar system using a tight cluster of ladar pulses. For example, ladar pulses in a cluster can be spaced apart but overlapping with at least one of the other ladar pulses in that cluster at a specified distance in the field of view. A ladar receiver can then process the reflections from the cluster to computer intra-frame motion data, such as intra-frame velocity and intra-frame acceleration for an object.

The invention discloses a human face three-dimensional reconstruction and human face replacing video editing system and method based on deep learning, and the system consists of a deep learning module, a three-dimensional reconstruction module, a video preprocessing module, a two-dimensional image generation module, and a video editing module. The deep learning module obtains necessary parameters for the generation of a three-dimensional model through a training convolution neural network. The three-dimensional reconstruction module obtains a corresponding three-dimensional model according to an inputted human face image / video. The video preprocessing module carries out the processing of a target video, and obtains the human face feature points and expression parameters corresponding to each frame. The two-dimensional image generation module enables a three-dimensional model of a target human face to generate a two-dimensional human face image corresponding to each frame. The video editing module enables a new human face image to replace the target image, and carries out the smoothing and adjustment of the illumination conditions.

The invention discloses a face three-dimensional reconstruction method based on a single face image. The method is based on an end-to-end convolutional neural network. In order to give full play to capability of a deep convolutional neural network, a face three-dimensional model is coded to a two-dimensional coordinate graph, a new loss function relatively fitting with face reconstruction is provided, and learning and prediction can be carried out through direct utilization of a lightweight end-to-end network. According to the algorithm, face alignment also can be carried out, and accurate three-dimensional face key point coordinates are obtained. Contrast experiments on a plurality of public face data sets show that according to the method, the accurate face three-dimensional model can bereconstructed through utilization of the single face image. Compared with the existing face reconstruction method, the method provided by the invention has the advantages that precision and speed arerelatively greatly improved.

The invention discloses an indoor robot positioning navigation method and device. The method comprises using a camera on a robot to shoot an image containing a calibration plate positioned above the robot; using an image processing technology to recognize the calibration plate in the image, determining a current position and a current deviation angle of the robot according to the calibration plate in the image and known coordinates of the calibration plate, and positioning absolute positions of the robot; by taking the current position and the current deviation angle as a reference, navigating the robot by adopting an inertia navigation manner. By using the invention, problems of low positioning precision, high cost, inconvenient use and narrow use range of the existing positioning navigation method are solved.

The invention discloses an electrocardiogram signal feature detection algorithm based on wavelet transformation lifting and approximate envelope improving and belongs to a weak bioelectrical signal processing technology field. The current electrocardiogram signal detection technology applied clinically can not give consideration to both a detection precision requirement and a real time requirement. Electrocardiogram signal pretreatment algorithm based on wavelet lifting for improving semi-soft threshold denoising and approximate envelope improving and electrocardiogram feature detection algorithm based on a slope threshold are provided in the invention. Detection criterions are set based on waveform characteristics and time domain distribution characteristics of the electrocardiogram signals. Position detections of R wave, start-stop points of the QRS waves, P wave and T wave are carried out respectively to the electrocardiogram signals. The electrocardiogram signal feature detection algorithm provided in the invention is easy, quick and suitable for parallel processing, and occupies little memory space and is convenient for DSP chip realization. Even in strong noise and P / T wave interference circumstances, R point position can be accurately detected through the algorithm provided in the invention. An R wave false detecting rate of 105 data containing serious noise disturbance is only 0.27% compared with MIT-BIH Database annotation.

The invention discloses a compressed sensing nuclear magnetic resonance imaging method based on a deep neural network. Through the method, a high-quality nuclear magnetic resonance image can be reconstructed with k-space low sampling data collected by a nuclear magnetic resonance imaging device. The method mainly comprises three steps: constructing an alternating direction multiplier deep neural network, training the parameters of the network, and applying the trained network to compressed sensing nuclear magnetic resonance imaging. A nuclear magnetic resonance image reconstructed with multiple pairs of sampling data at low sampling rate and corresponding full-sampling data is used as a training data set, and the model parameters of the alternating direction multiplier deep neural network are trained so that the output image of the deep neural network when sampling data at low sampling rate is used as input is as close as possible to an image reconstructed with full-sampling data. In application, k-space sampling data at given low sampling rate is input to a trained alternating direction multiplier deep neural network, and the output of the network is a reconstructed nuclear magnetic resonance image.

The invention relates to a method for macrosegregation numerical simulation of a casting, aiming at solving the problems of long computation time, low precision and unsuitability for predication of castings with large size or complex shapes in the traditional macrosegregation. The method comprises the following steps of: performing mesh generation on a selected dendritic crystal growth microcosmic computation field; computing dendritic crystal growth shapes during alloy solidification under different maximum nucleation densities from a microcosmic scale, and outputting a solid phase fraction-average dendritic crystal solid phase feature diameter curve; performing macro-scale mesh generation on the casting; computing a mass, momentum, energy and composition conservation equation from a macrocosmic scale; computing a solid phase fraction by using a Newton downhill method; and computing a pasty zone permeability model from a macrocosmic scale by adopting a linear interpolation technologybased on the solid phase fraction-average dendritic crystal solid phase feature diameter curve under a given nucleation density. The method disclosed by the invention is suitable for macrosegregationnumerical simulation in sand casting and metal mold casting of various sizes and complex shapes.

Disclosed herein are examples of ladar systems and methods where data about a plurality of ladar returns from prior ladar pulse shots gets stored in a spatial index that associates ladar return data with corresponding locations in a coordinate space to which the ladar return data pertain. This spatial index can then be accessed by a processor to retrieve ladar return data for locations in the coordinate space that are near a range point to be targeted by the ladar system with a new ladar pulse shot. This nearby prior ladar return data can then be analyzed by the ladar system to help adapt a shot selection for use by the ladar system with respect to new ladar pulse shots.

The invention discloses an additive and subtractive manufacturing based metal part repair method. The additive and subtractive manufacturing based metal part repair method is used for solving the technical problem of the low material utilization rate of an existing metal part repair method. According to the technical scheme, firstly, after to-be-repaired model three-dimensional data are obtained through three-dimensional scanning, subtractive machining is carried out on the worn part of a part, that is, the part is worn through dry high-speed cutting, and a regular surface is obtained; and then additive manufacturing is carried out through an electron beam fuse wire additive process, the deposition additive repair process of the part is completed, meanwhile, the part surface is machined through dry high-speed cutting machining, and a repair part with high surface quality is obtained. According to the additive and subtractive manufacturing based metal part repair method, error accumulation brought by workpiece clamping, taking and placing during original multi-platform machining is avoided, meanwhile, the interference problem happening in the subtractive process of parts of complex structures is solved, and the mechanical performance of the damaged part achieves the level of normal parts. According to the method, the repair process is carried out in a vacuum environment, workpiece clamping only needs to be carried out once, and the material utilization rate is high.

The invention discloses a projection positioning device, comprising a projector, a camera and a processing device, wherein the projector is used for projecting regularly-set strip patterns on the surface of an object to be detected; the camera is used for shooting a modulated and deformed strip image according to the shape of the surface of the object to be detected; and the processing device is used for determining a space position and / or a position change of the object to be detected according to the deformed strip image. Furthermore, the invention discloses a corresponding projection positioning method, an interaction system and an interaction method. By projecting the strip patterns on the surface of the object to be detected and capturing and analyzing the deformed strip image modulated by the evenness of the surface of the object to be detected, the higher-precise position and the faster-speed real-time trace are carried out on the object to be detected, such as a human hand with complex three-dimensional postures, so that the accurate and smooth motion sensing or touch operation are realized.

The invention discloses a self-propelled rotary face milling cutter with a slicking blade, comprising a cutter body, a tool apron slot, a cutting tool apron, a slicking tool apron, a positioning surface, a bearing, a cutting blade and a slicking blade. The cutter body is in an odd-number five-slot shape. The tool apron comprises the cutting tool apron and the slicking tool apron. The cutting tool apron utilizes a cutter shaft supporting structure of combining the angular contact bearing and the composite bearing. The slicking tool apron utilizes a cutter supporting structure of combing enclasping up-and-down sliding bearing ring and the composite bearing. The tool apron slot utilizes an over-positioning structure of repeated limiting of the conical surface and the planar surface, and the tool apron slot inclines in both axial direction and the radial direction, so as to enable the self-rotary blade to obtain the cutting edge inclination angle and ensure the continuous self-rotation during processing. The blade utilizes a tangential structure. The slicking blade finally finishes processing the surface. The milling cutter has the advantages of reasonable design, compact structure, good cutting performance, fast cutting speed and high processing precision. The milling cutter also has the advantages that the cutting speed is increased by 50%, the processing precision is improved by 1-2 levels, the surface roughness is up to Ra0.8 micrometers, cooling lubricant is replaced by drying milling, environmental pollution is avoided, and the cutting and slicking functions complement with each other.

The invention relates to micro-pipe internal flaw measurement and optical measurement. The invention provides a micro-pipe internal flaw measurement and appearance measurement device and a method, having the performance characteristics of flexibility, automation, rapidness, online and the like, so as to meet the measurement requirements of modern industry on small size feature inner bore or micro-pipe internal flaw and appearance. For achieving the purposes, the invention adopts the technical scheme that: the micro-pipe internal flaw measurement and appearance measurement device comprises a micro-pipe internal flaw measurement and appearance measurement sensor, wherein the sensor consists of an optical transmission part, a lighting source and a camera; the sensor is fixedly connected with a tail end joint of an industrial robot; pipe internal 360-degree cylindrical surface annular panoramic picture data is transmitted by the optical transmission part, shot by the camera and is collected by an image acquisition card to be processed by a computer. The micro-pipe internal flaw measurement and appearance measurement device and the method are mainly applicable to micro-pipe internal flaw measurement.

A laser detecting instrument is disclosed that comprises body, laser head, mechanical arm, delivery device, material receiving device, detecting clamping fixture, laser controller. Said laser head is mounted on the laser lifting device which can make the laser head move along the direction of the Z axis, the laser lifting device is mounted on the body supporting plate, the delivery device, the material receiving device and the detecting clamping fixture are also mounted on the body supporting plate, the mechanical arm includes the mechanical fingers, the Y axis supporting device, the Z axis supporting device, and the mechanical fingers mounted on the Y axis supporting device which makes the mechanical fingers move along the direction of the Y axis, the Y axis supporting device is mounted on the Z axis supporting device which is mounted on the body supporting plate to support the whole mechanical arm. The laser detecting instrument includes the control device which controls the movement of said laser head, the laser controller, the mechanical arm, the delivery device, the material receiving device, the detecting clamping fixture. The invention has high precision in detection, fast speed, flexible setting for parameters, simple and convenient operation, and it has the large market and good economic benefit.

The invention discloses an adjacent panel misplacement unit glass wall system, comprising a unit keel, wherein the unit keel comprises a male vertical column, a female vertical column, an upper transverse beam, a middle transverse beam and a lower transverse beam; the male vertical column, the female vertical column, the upper transverse beam, the middle transverse beam and the lower transverse beam form a frame structure; the unit keel is sequentially provided with a galvanized steel sheet, insulation rock wool and an aluminum back plate from bottom to top; outwards convex glass or inwards concave glass is arranged on the male vertical column or the female vertical column; the male vertical column and the female vertical column are respectively connected with an adjustable connecting piece; and an adjusting bolt is arranged on the adjustable connecting piece. According to the curtain system disclosed by the invention, the problem of needing excessive opening moulds, caused by the factor that panels are not on one plane, is solved, so that the engineering cost is reduced, the production efficiency is improved, and the construction period is shortened. The concave-convex model of the panel can be completed in the process of producing plate blocks, so that the installation precision of the plate blocks and the engineering effect are ensured; and in comparison with the traditional construction technology processing, the curtain system has the characteristics of simpleness for manufacture, high production efficiency, convenience for installation, and concise and practical structure.

The invention discloses a rapid detecting device and a detecting method of the precision of blades of a gas turbine. A detected blade is positioned by a blade positioning device; and profile measuring devices are symmetrically arranged at two sides of the blade positioning device and used for acquiring profile data of the section of the blade. By adopting the rapid detecting device and the detecting method, the workload for acquiring blade conic data is greatly reduced; influence from personal factors is little and the data acquiring speed is high; contact type measurement by a measuring pen is adopted with high measurement precision; the arithmetic of the profile data of a discrete section is simple and mature, is high in precision and is easy for subsequent error analysis calculation; the height and the quantity of the profile data is selective and adjustable and has high flexibility; the mechanical structure is simple, the cost is relative low and the measurement precision is ensured.

The invention discloses a linear array camera and area array camera combined large-aperture super clean smooth surface defect detection device. When a super clean smooth surface is detected, the device provided by the invention firstly uses a linear array camera with low magnification factor to quickly scan full aperture and acquires a dark field image of a sample surface, extracts position information of defects, and then, locates each defect through an area array camera configured with a high fixed-times optical system and further performs data processing analysis for the image, and thus, obtains a high-precision detection result. Such detection mode is not only fast, but also highly accurate and sensitive. Moreover, the device provided by the invention also carries a distance sensing system for leveling of a to-be-detected super clean smooth surface and focusing of an imaging system, and has more fast adjustment accuracy and stability. The device provided by the invention solves a difficult problem of various large-aperture super clean smooth surface defect automated detection, and can be used for fast detection on surfaces of various smooth glass, wafers, metals and so on.

The invention provides a stainless steel electrolytic etching process, which mainly comprises the following steps of: (a) pre-treatment: cleaning the surface of a stainless steel workpiece by washing, deoiling, etc.; (b) application: applying a photosensitive adhesive on the surface of the workpiece and carrying out exposure, developing, hardening, etc.; (c) electrolytic etching: disposing the workpiece in an electrolysis tank containing a stainless steel electrolytic etching solution and electrifying for electrolysis, wherein the stainless steel electrolytic etching solution mainly contains copper chloride 200 to 400 g / L, hydrochloric acid 20 to 80 g / L, thiourea 1 to 5 g / L and potassium sodium tartrate 10 to 20 g / L; and (d) demolding: removing the residual adhesive film on the surface after electrolysis and cleaning to obtain the product. Accordingly, the stainless steel electrolytic etching process provided by the invention can achieve very high etching speed and higher accuracy under the low pressure condition, and has the advantages of simple process, low power consumption, high etching speed and wide application prospect.

The invention belongs to the technical field of building deformation measurement and particularly relates to a high tower tube bending degree detection method based on three-dimensional laser scanning. The method comprises the steps of (1) arranging an instrument through retention points which are symmetrically arranged around a tower tube as a center, and obtaining point cloud data of a tower tube body, (2) constructing a tube body TIN model, selecting tower tube cross sections, and carrying out fitting of a tower tube central axis, (3) randomly selecting n points or selecting n points with a set interval along a counter-clockwise (or clockwise) direction at the edge perimeters of the cross sections and carrying out reduction of barycentric coordinates of the tower tube cross sections, and (4) calculating a tower tube bending displacement component, a bending displacement resultant and a bending average curvature according to an adjacent cross section through the obtained barycentric coordinates of the tower tube cross sections. According to the method, the disadvantages of difficult field operations, difficult precision control, time and labor consumption, poor accuracy of target data and a large data limitation in the traditional detection of a high tower tube bending degree are solved, the high tower tube bending degree detection can be rapidly and precisely realized, and the method has important practical significance for similar engineering detection and monitoring.

The invention provides a method and a device for detecting a lane line and belongs to the technical field of image processing. The method comprises the steps of obtaining an image in front of a vehicle; performing edge detecting on the image in front of the vehicle to obtain an edge image; determining the detecting region of the initial lane line in the edge image according to the position of the lane line obtained in the front frame; dividing the detecting region of the initial lane line into N small detecting regions, wherein N is larger than or equal to 2; determining a certain small detecting region as the precise detecting region of the lane line according to the edge image numerical number of each small detecting region and the total quantity of pixels of each small detecting region; obtaining the lane line of the current frame from the precise detecting region of the lane line. The method and device for detecting the lane line are simple and the detected lane lines are more accurate.

Methods, devices and circuits are provided for power-on-reset circuits with low static power consumption. One such circuit (100) includes a detector (104) that draws current from a supply voltage. The detector detects that the supply voltage has exceeded a trip-point voltage level and then disables current draw from the detector. The detector responds to an enable signal by enabling current draw from the detector. A pulse generator (106) generates a reset signal in response the supply voltage transitioning from a voltage below the trip point voltage level to above the trip point voltage level. A monitor (108) detects that the supply voltage has dropped and provides, in response thereto, the enable signal to the detector to enable current draw from the portion of the detector.

The invention discloses a QGA-MMRVM-based self-adaptive inversion method of material parameters of a rockfill dam. The method comprises the following steps: firstly, introducing a mixed kernel function on the basis of a multi-output relevance vector machine (M-RVM) to build MMRVM with higher precision; then using parameter-fixed QGA to optimize MMRVM kernel parameters to realize self-adaptive calculation of an MMRVM model; and then exerting global-search capability of the QGA, and using actually measured settlement data of the dam as a goal to obtain dam construction material constitutive-model parameters by inversion to realize self-adaptive calculation of an inversion model. The method solves the problems of low computational accuracy, low computational speed, insufficient small-sample pertinence, poor inversion-model adaptability and the like existing in the prior art; and can be widely applied to other engineering and inversion projects.

The invention relates to a testing device and the method for the fabric moisture-penetrability. The device is made up of the column canister, the base, the waterproof moisture film, the pressure ring and the testing board. The shallow groove is set on the base which the two ends are connected with the air inlet and the air outlet carved on the base two sides. The waterproof moisture film is the bottom of the column canister to form the saturate water vapor resource; the dry quick flowing N2 is the carrier to carry the water vapor through the fabric away. So it can measure the relative humidity of the N2 flow to get the moisture content of the fabric. The invention has the high precision and good repetition, the testing time is short and the cost is low; so the device can be used in the textile mill to control the product moisture-penetrability quality.

The invention discloses a workpiece pose estimation method based on point cloud, which is mainly used for solving the problem of workpiece pose estimation in the field of automatic assembly, especially in the field of small part assembly. The method comprises the steps of point cloud data preprocessing, point cloud virtual view extraction and pose estimation. Wherein the pose estimation step further comprises the steps of performing iterative operation by adopting a particle filter algorithm based on a dynamic model of an iterative closest point algorithm, outputting the pose of the effectiveparticle in a weighted mean mode if an iterative stop condition is met, and calculating the pose of the target workpiece relative to the camera coordinate system through inversion operation. The invention also relates to an apparatus comprising a memory and a processor implementing the above method steps when the processor executes a program stored in the memory.

The invention discloses a novel double-loop control method for a dual-active bridge DC / DC converter based on double phase-shift control. Under a double phase-shift control mode, a transfer function from the internal and external phase-shift angles to output voltage in different operating modes are obtained by using a small signal modeling technique in order to guide the inner-loop design of the dual-active bridge DC / DC converter. By establishing a relation between a loss model and peak current in different operating modes, the internal and external phase-shift angles in different operating modes during the optimal peak current are obtained so as to guide the outer-loop design of the dual-active bridge DC / DC converter. The use of a double-loop control compensation system can increase the cutoff frequency of the system, can greatly improve the dynamic response of the system, can reduce the static error of the system, and can realize the fast response speed and precision. The control method of replacing the optical loss with the optical peak current can simplify controller design and realize the real-time control of the converter.

The invention discloses a cold-rolled sheet shape closed-loop control method and system. The cold-rolled sheet shape closed-loop control method comprises the steps: evaluation functions synergistically controlled by cold-rolled sheet shape executing mechanisms are built according to cold-rolled sheet shape regulating and controlling efficacy coefficients; k-time solution vectors of the executing mechanisms are determined according to the evaluation functions; whether the k-time solution vectors are within feasible regions of the cold-rolled sheet shape executing mechanisms or not is judged, and if yes, the k-time optimal regulating variables are obtained; if not, evaluation functions harmoniously controlled by the cold-rolled sheet shape are determined, and gradient informationof the evaluation functions and the regulating directions of the cold-rolled sheet shape executing mechanisms are determined according to parameters of the cold-rolled sheet shape executing mechanisms; final values of the regulating variables of the executing mechanisms are determined according to the gradient information and the regulating directions; the evaluation functions are subjected to order reduction according to the final values, evaluation functions subjected to order reduction are obtained, and (k+1)-time solution vectors are determined, wherein the relation of k=k+1 is made to be met; and the cold-rolled sheet shape is adjusted according to the k-time optimal regulating variables of all the cold-rolled sheet shape executing mechanisms. According to the cold-rolled sheet shape closed-loop control method and system, the sheet shape regulating speed can be increased.

The invention relates to a method for manufacturing a water turbine model runner combining mechanical processing and additive manufacturing. The upper crown and the lower ring of the model runner are manufactured by mechanical processing of steel plates. On the formed crown, the electric arc additive manufacturing technology is adopted, and the cold metal transition surfacing method is used to import the mathematical model of the model runner into the computer, and the blade slicing and welding path planning are completed through the software, and the robot system is controlled to complete the blade augmentation. Manufactured from high-quality materials to form near-net-shaped blades, and stress-relief heat treatment is performed on the manufactured upper crown and the overall parts of the blade. The blade is machined in the CNC machining center, the upper crown and the blade are polished as a whole, and the lower ring is finally assembled to complete the manufacture of the turbine model runner. The method has the advantages of high processing precision, short cycle time and low cost.

The invention provides a multi-class pneumonia screening deep learning system based on CT images. The method comprises the following steps of a CT image data preprocessing part preprocesses a CT imageto obtain a preprocessed CT image; the slice-level pneumonia dichotomy part analyzes the basic depth features to obtain a dichotomy result, the weakly supervised lesion positioning part obtains a lesion position map according to the class labels, and the slice-level pneumonia four-class part detects four classes of pneumonia to obtain a four-class pneumonia classification task result; the case-level pneumonia classification part performs analysis to obtain a multi-class pneumonia classification task result; and the pneumonia diagnosis comprehensive evaluation part outputs a pneumonia diagnosis comprehensive evaluation result and a focus positioning distribution map based on the slice-level four-classification result, the focus position map and the case-level multi-classification pneumoniaclassification task result. The method is advantaged in that the COVID-19 and other pneumonia diseases can be quickly and accurately distinguished on the premise of no medical staff besides obtainingthe common pneumonia diseases, which is helpful for screening the epidemic situation of the COVID-19; the COVID-19 and other pneumonia diseases can be quickly and accurately distinguished.