988results about How to "Meet practical application needs" patented technology

The invention relates to a visual computing and prewarning method of deviation angle and distance of an automobile from a lane line. The image processing and computer vision technologies are utilized, and the deviation angle and distance of the automobile from the lane line are computed in real time according to the road surface image acquired by a vehicle-mounted camera, thereby estimating the line crossing time for safety prewarning. The method comprises the following steps: detecting the lane lines of the road surface image to obtain a linear equation of partial lane lines; establishing a three-dimensional coordinate system by using the camera as the initial point, and recording the mounting height and depression angle of the camera; calibrating the focal length according to the lane detection result under the condition of a given deflection angle; computing the deflection angle and vertical distance of the automobile relative to the lane line according to a pinhole camera model; and estimating the deviation time from the lane according to the instantaneous running speed of the automobile, thereby obtaining the safety prewarning or intelligent control information of the running automobile.

The invention relates to a deep convolutional neural network-based traditional Chinese medicine tongue image automatic segmentation method and belongs to the computer vision field and traditional Chinese medicine tongue diagnosis field. According to the method of the invention, a convolutional neural network structure is designed; collected sample data are adopted to train the network, so that a network model can be obtained; and the model is adopted to automatically segment a traditional Chinese medicine tongue image. The method includes an offline training phase and an online segmentation phase. The method can be applied to both closed type and open tongue image acquisition environments and can effectively improve the accuracy and robustness of the automatic segmentation of the traditional Chinese medicine tongue image. The method of the present invention specifically relates to deep learning, semantic segmentation, image processing and other technologies.

The embodiment of the invention discloses a data distribution method, which is used for improving the performance of a distribution storage system. The embodiment of the invention comprises the following steps: a data distribution device receives a storage instruction of a user; to-be-stored data which is indicated to be stored by the storage instruction is divided into P data segments; a storage node group corresponding to each data segment is determined; and finally, each data segment is distributed to the main node of the corresponding storage node group. The embodiment of the invention also provides a relevant data storage method and a relevant device and system.

The invention discloses a multi-person human body posture estimation method, and relates to processing of a recording carrier for identifying graphics. The method is a multi-person human body postureestimation method based on depth features. According to the method, a depth feature human body key point detection model composed of a main body network and a fine tuning network is constructed; a depth feature human body key point detection model is trained by adopting a two-stage training method combining a bottom-up method and a top-down method; finally, the human body key points are detectedby using a depth feature human body key point detection model trained in two stages. Redundant key points which do not belong to a target person are removed through human body key point clustering processing, then a multi-person human body posture estimation result is output, and the defects that in the existing multi-person human body posture estimation method technology, under the condition thatcrowds are dense, the target human body shielding interference robustness is poor, and the human body key point detection accuracy is low are overcome.

The invention belongs to the technical field of membranes and in particular relates to a porous material-polymer gas separation composite membrane. The composite membrane comprises a separation layer and a supporting layer, wherein a porous material-polymer hybrid membrane is adopted as the separation layer; one of a porous flat plate and a tubular or hollow fiber membrane is adopted as the supporting layer; the porous material of the separation layer is metal-organic frame materials (MOFs) with large specific surface area and high gas adsorption. The MOFs are added into the separation layer so that the gas permeation and selectivity of the membrane are obviously improved; the separation layer and the supporting layer are compounded so that the membrane keeps good mechanical property, and the separation layer is allowed to exist in a compact or asymmetric structure. The composite membrane provided by the invention has high gas separation property, good mechanical property and wide industrial application prospect.

The invention discloses a reservable type human face identification door access visitor system and a method. The reservable type human face identification door access visitor system comprises a visitor equipment, an owner terminal, a passage gate and a cloud platform server, wherein the visitor equipment is connected with the cloud platform server to obtain a first human face information of a visitor and generate a first human face feature template; the owner terminal is connected with the cloud platform server to check the application of the visitor and send the check information to the cloudplatform server; the cloud platform server is connected with the passage gate to send the door opening authority of the visitor and the first human face feature template to the passage gate when thereviewing information is passing allowance; the passage gate is used for detecting whether the visitor stays in the identifying area or not; when the visitor stays in the identifying area, a second human face information of the visitor is obtained so as to further generate a second human face feature template; the first human face feature template and the second human face feature template are compared; when the first human face feature template is matched with the second human face feature template, and the passage gate obtains the door opening authority, the gate passing and releasing are executed.

The present invention discloses a three-dimensional reconstruction method and system capable of maintaining sharp features. The method includes the following steps that: 1) a two-dimensional image filtering method extended to a three-dimensional space is adopted to perform smoothing de-noising on inputted roughly-registered point cloud; 2) an improved region growth method is adopted to perform outlier removal on the smoothed roughly-registered point cloud; 3) a kd-tree (k-dimensional tree) acceleration-based ICP (iterative closest point) algorithm is adopted to perform precise registration on the outlier-removed roughly-registered point cloud; 4) a neighborhood search and boundary point detection-based fusion method is adopted to fuse the precisely-registered point cloud; and 5) a feature point detection and adaptive step size update-based method is adopted to perform surface reconstruction on the fused precisely-registered point cloud. The three-dimensional reconstruction system is composed of a point cloud preprocessing module, a point cloud combining module and a surface reconstruction module. The three-dimensional reconstruction system which is realized based on the method of the invention can maintain the sharp features of the edge of a reconstructed model, and therefore, reconstruction speed is considered with accuracy ensured.

The invention discloses a Chinese word stock automatic generation method based on writing style modeling, and the method comprises the steps: carrying out the stroke track automatic extraction and wrong result screening operation of inputted handwritten Chinese character image, and obtaining training data; carrying out the learning and modeling of handwritten Chinese character stroke shape and style and stroke inter-frame structure style through a neural network, and obtaining handwriting stroke connection characteristics and stroke end part contour handwriting features through statistical analysis, and then generating a high-quality Chinese character form consistent with the handwriting style of a user. The method just needs a few of commonly-used Chinese characters as the input, does not need any manual intervention, can automatically generate a vector Chinese word stock comprising a large number of Chinese character patterns, can quickly and automatically generates a handwriting Chinese word stock with other handwriting style for a user, can remarkably improves the manufacturing efficiency of the Chinese word stock, and greatly reduces the production cost.

The invention provides a feature line fused laser point cloud and analog image registration method, and aims to solve the problems that a point cloud distance intensity image or point feature-based image and point cloud registration method in the prior art has many defects of manual point selection errors, point cloud data interpolation errors, need of many observation values and the like, and theprecision is very limited. The invention provides a laser point cloud and analog image registration method based on two-dimensional and three-dimensional fusion characteristic lines. The method comprises the steps of: establishing a direct transformation model between the image feature line and the point cloud feature line. The requirement on the initial value of the exterior orientation elementof the image is not harsh, the method has high robustness to noise, realizes automatic matching of two-dimensional and three-dimensional fusion features, improves the degree of automation, achieves higher real-time performance, has the advantages of no model error, no need of exterior orientation element initial values, only need of a small number of registration elements and the like, is high incalculation speed, and meets the registration requirements of analog images and ground laser point cloud data.

The invention discloses an automatic generation method of a handwritten Chinese character library, which comprises the following steps of: predicting a Chinese character skeleton flow field by adopting a two-stage convolutional neural network G through a writing track synthesis stage and a font style rendering stage on the basis of a deep stacking network and structure information guidance; firstly, the writing style is learned from a small amount of handwritten Chinese characters written by a user, so that the writing track of unwritten Chinese characters is synthesized; The handwriting styleof the target is rendered; and generating a complete GB2312 Chinese handwriting word stock file with the writing style of the user. The method can ensure the structural accuracy and style consistencyof the generated fonts at the same time, is simple, efficient, low in cost and high in quality, and can meet the actual application requirements of ordinary people for quickly making personalized handwriting fonts.

The invention discloses a method for preparing a high-performance high-temperature-resisting nanometer composite permanent magnet, relating to preparation techniques of permanent magnetic materials. The method mainly comprises the following steps of: (1) weighing each element raw material according to elements of NdFeB alloy, mixing the element raw materials, melting the mixed raw material in vacuum, and rapidly quenching to prepare thin belts; (2) preparing SmCo ally rapidly quenched belts through the step (1); (3) respectively carrying out high-energy ball milling on the NdFeB and the SmCo alloy rapidly quenched belt to prepare corresponding nanometer crystal alloy powder; (4) mixing the NdFeB and the SmCo nanometer crystal alloy powder according to certain proportion, adding a crystal boundary nanometer modifier to be uniformly distributed on the surfaces of the NdFeB and the SmCo powders to obtain composite powder with the uniformly mixed three; (5) pressing the composite powder to form moulded blanks; (6) carrying out discharge plasma sintering on the moulded blanks to prepare a nanometer composite magnet; and (7) carrying out thermal deformation on the nanometer composite magnet to improve the degree of orientation and obtain the high-performance high-temperature-resisting nanometer composite permanent magnet. The method has the advantages of simple process and easiness in operation, and is suitable for large-scale batch production.

The invention provides a YOLOV3-based smoke and fire automatic detection and early warning method. The method comprises the steps of S1, constructing a sample set of a training detection model; S2, establishing a deep learning target detection network architecture based on YOLOV3; S3, configuring training parameters, and training a detection model; S4, acquiring to-be-detected image information; acquiring image frames of an on-site video picture from monitoring equipment on a to-be-detected site, and processing frame-by-frame images by utilizing an image preprocessing method; S5, detecting smoke and flame targets, and sending the video image frames processed in the step S4 into a pre-trained detection model in the step S3 for target detection, and outputting a detection result; S6, carrying out post-processing on a detection result; and S7, continuously analyzing the detection results of the multiple frames of images, confirming that the target is valid and outputting an alarm. According to the YOLOV3-based smoke and fire automatic detection and early warning method, second-level detection and alarm can be realized, the fire early warning time is greatly shortened, timely notification and timely rescue are realized, and fire spreading is effectively prevented.

The invention provides a Chinese medicine tongue color and tongue coating color automatic analysis method based on a convolutional neural network, belongs to the field of computer vision and the field of Chinese medicine tongue diagnosis, and particularly relates to the technology of deep learning and image processing. The accuracy and the robustness of tongue color and tongue coating color automatic analysis can be effectively enhanced by the method. The classification accuracy is high. Compared with the conventional "feature extraction + classifier" scheme, the convolutional neural network has an end-to-end structure, and the two processes of feature extraction and classification are completed under the same framework so that the dependence on feature and classifier selection can be avoided. The method has obvious advantages in the aspect of classification accuracy in comparison with the conventional method so as to meet the actual application requirements.

The invention relates to a production method of ultra-fine crystal magnesium / titanium laminated composite panel, which comprises the following steps: firstly, cleaning the surfaces of the metal materials; secondly, stacking the metal materials; thirdly, annealing the stacked sheets; fourthly, rolling the sheets after annealing; fifthly, cutting the rolled sheets into two identical parts, trimming and deburring; repeating the first step to the fourth step for lamination and ply rolling; sixthly, annealing the composite panel obtained from lamination and ply rolling to obtain the ultra-fine crystal magnesium / titanium laminated composite panel. The production method solves the problems of complicated technology, high cost, poor corrosion resistant performance of material and poor interface binding performance of the magnesium / titanium laminated composite panel existing in the prior art. The production method has the advantages of simple technology, low cost as well as good corrosion resistant performance and good interface binding performance of the composite panel.

The invention discloses a preparation method of a titanium alloy cast ingot containing high-melting-point elements, which comprises the following steps: 1, weighing raw materials: proportionally weighing the raw materials according to nominal chemical components of a titanium alloy to be prepared, wherein the raw materials include fine-grained sponge titanium, a titanium shell coated on the outer side of an electrode block to be prepared and high-melting-point metal powder; 2, mixing: evenly mixing the weighed fine-grained sponge titanium and high-melting-point metal powder to obtain a uniform mixture; 3, preparing a consumable electrode: pressing the evenly mixed uniform mixture into the electrode block, coating a titanium shell layer on the outer side of the electrode block, and welding the electrode block to form the consumable electrode; and 4, smelting: smelting the prepared consumable electrode in a vacuum consumable electric arc furnace to obtain the titanium alloy cast ingot containing high-melting-point elements. According to the invention, the method is simple in steps, reasonable in design, low in input cost, simple to use and operate and good in use effect; and the prepared titanium alloy cast ingot is uniform in components and has no high-density inclusions.

The invention provides a question classification method and a question classification device for an automatic question-answering system. A class reference table including fine granularity classes of noun and interrogative pronoun is preset. The method comprises the steps: calling a preset interface for executing the class marking operation to divide a received question into a plurality of word segmentations, and carrying out the fine granularity class marking for the word segmentations according to the class reference table to obtain a corresponding fine granularity class mark sequence; matching the fine granularity class mark sequence with a preset first-level classification mode so as to primarily determine the type of the question; if the primarily determined type is non-unique, matching the fine granularity class mark sequence with a preset second-level classification mode so as to secondarily determine the type of the question; if the secondarily determined type is non-unique, matching the fine granularity class mark sequence with a preset third-level classification mode so as to determine the type of the question at a third time, and determining the type to be the type of the question if the type determined at the third time is unique. By adopting the method, the question classification efficiency and accuracy can be improved.

The invention discloses a holographic polymer dispersed liquid crystal grating and a preparation method thereof. The holographic polymer dispersed liquid crystal grating is prepared through holographic record treatment on mixture including compound liquid crystal and photo-polymerizable monomer. The compound liquid crystal includes first kind liquid crystal compound and second kind liquid crystal compound, wherein the first kind liquid crystal compound is biphenyl cyanide liquid crystal compound, and the mass ratio of the first kind liquid crystal compound to the second kind liquid crystal compound in the compound liquid crystal is 3 / 26 to 15 / 10. The mass ratio of the compound liquid crystal to the photo-polymerizable monomer is 25 / 50 to 50 / 30. By means of improving a structure, composition and the corresponding preparation method of the key holographic polymer dispersed liquid crystal grating, a novel way for improving diffraction efficiency and a driving voltage requirement of the holographic polymer dispersed liquid crystal grating is provided.

The invention discloses an automatic identification method for a milking sow gesture on the basis of a depth image. The method comprises the following steps that: collecting original depth image data,carrying out preprocessing, and carrying out manual annotation to form a milking sow gesture identification dataset; designing and training a milking sow hybrid deformable component model based on animproved HOG (Histogram of Oriented Gradient) feature; constructing a milking sow gesture identification deep convolutional neural network, utilizing an annotation frame and annotated gesture category training set information, and combining with a random Dropout method to train the network; inputting the test set into the milking sow hybrid deformable component model to obtain the target area ofthe milking sow; and inputting a target area result into the milking sow gesture identification deep convolutional neural network to identity the milking sow gesture. By use of the automatic identification method for the milking sow gesture on the basis of the depth image, the problem that an RGB (Red, Green and Blue) image is likely to be affected by the changes of factors, including outside illumination, shades and the like is overcome, the problem that the milking sow gesture is difficult in identification at night is solved, and the practical application requirement of all-weather milkingsow gesture monitoring can be met.

The invention discloses a rod-shaped object recognition-based mobile robot self-positioning system. According to the invention, the positioning method of the system includes the following steps that:a rod-shaped object off-line map is constructed; a laser radar is utilized to obtain point cloud data around a moving object, and rod-shaped object information is extracted from the point cloud data;and self-positioning is performed according to the rod-shaped object information and the rod-shaped object off-line map, so that the positioning information of the moving object can be obtained. According to the rod-shaped object recognition-based mobile robot self-positioning system of the invention, the laser radar is utilized to identify and track a rod-shaped object; data processing and fusionare performed through a corresponding positioning algorithm on the basis of the information of an inertial measurement unit and a GPS, and therefore, the self-positioning of the moving object can berealized. With the method of the invention adopted, positioning accuracy of a decimeter level can be realized. Capable of meeting actual application requirements, the system or the method has the advantages of low cost, high precision, real-time performance and robustness.

The disclosed embodiment provides a video shooting method, an apparatus, an electronic device and a computer-readable storage medium. The method includes: receiving a video shooting trigger operationof a user through a video playing interface of an original video; in response to the video shooting triggering operation, superimposing a video shooting window on the video playing interface; receiving a video shooting operation of a user through thevideo playing interface; in response to the video shooting operation, playing the original video is ed while the user video is shot, and displaying the user video through the video shooting window; and synthesizing the user video and the original video to obtain a co-shot video. With the scheme of this embodiment, the user only needs to carry out the related operation of shooting the user video on the video playing interface to obtain the function of co-shooting video, and the operation process is simple and fast. Because the user's feelings tothe original video can be reflected through the user video, the user can conveniently display the feelings to the original video through the scheme, and the user's interactive experience is improved.

The invention provides a near-infrared universal model detection method for quality indexes of fruits with similar optical and physical properties. According to the near-infrared universal model detection method, a near-infrared universal model is established; the near-infrared universal model is utilized for determining two parts of fruit quality indexes; the near-infrared universal model is established based on the similarity of near-infrared spectrums between the different types of the fruits based on the similar optical and physical properties. The basic modeling concept is that after the spectrums are collected and are pre-processed, a common characteristic wavelength range of all the types of the fruits is screened by using a moving window partial least square method; common characteristic wavelength points are further extracted in the common characteristic wavelength range by using an SPA (Super Pairwise Alignment) algorithm; finally, a PLS model or an MLR (Multiple Linear Regression) model is established by using existing software. According to the method, the determination accuracy is high and the feasibility is strong; the disadvantages that the different types of the fruits need to be detected by classes in an existing near-infrared detection technology are overcome; the modeling cost is reduced, the working efficiency is improved and the number of the modeling wavelength points is few; the near-infrared universal model detection method is applicable to a common optical filter type near-infrared instrument.

The invention discloses a Chinese medicinal material microscopic image feature extraction and recognition retrieval method. A PCNN time matrix information processing method is introduced, image feature information in a transform domain is extracted through combination of PCNN and image Fourier transform and decimal power exponent filtering phase, and a Chinese medicinal material microscopic image information rapid storage and recognition retrieval method is put forward on the basis of mahalanobis distance with combination of comprehensive similarity measure of a Pearson product moment correlation method so that a holographic Chinese medicinal material microstructure image and pollen image feature extraction and recognition retrieval system is constructed, objectivity, accuracy, repeatability and intelligent degree of Chinese medicinal material quality evaluation can be further enhanced, and a new approach is provided for modernization of Chinese medicinal material detection and analysis.

The invention relates to the technical field of information processing, and discloses a logistics monitoring method and system. The logistics monitoring method comprises the following steps: obtaining the position information of a logistics vehicle, wherein the position information comprises a current position; judging whether the current position deviates from a preset logistics path or not; and if the current position deviates from the preset logistics path, generating early warning information. The invention also provides a logistics monitoring system. By use of the logistics monitoring method and system, a merchant can monitor the current position in a commodity transportation process in real time, and can know accident situations in commodity transportation in time through the early warning information which is automatically generated so as to adopt a response measure in time to avoid risks.

The embodiment of the invention provides an augmented reality-based information providing method and device, equipment and a storage medium. The method comprises the steps of obtaining a scene image of a reality scene collected by an image collection device; according to the scene image, obtaining to-be-displayed augmented reality AR content and a display target of the AR content in the scene image; fusing the AR content with the display target so as to provide the fused content for the user; wherein the AR content is determined by the following modes: identifying a scene image to obtain scene information of the scene image; and determining AR content corresponding to the scene information based on the scene information. Based on the scheme provided by the embodiment of the invention, the information associated with the scene where the user is located can be provided for the user, and the information is displayed to the user based on the AR technology, so that the immersion of the user is effectively improved, and the actual application requirement can be better met.

The invention relates to an oil-electric hybrid power multi-motor aircraft and belongs to the technical field of unmanned aircrafts. The oil-electric hybrid power multi-motor aircraft is mainly composed of an upper main rotor 1, a lower main rotor 2, a multi-rotor aircraft electric control panel 4, motor-driven auxiliary rotors 5, motors 6, a fuel oil engine 7, a rotating speed pairing gearbox 8, a battery 11 and a throttle control stepping motor 24. Flying power is provided for the aircraft through the main rotors and the auxiliary rotors jointly, the main rotors provide main lifting power, and the auxiliary rotors provide a small part of lifting power; the main rotors are composed of the upper main rotor 1 and the lower main rotor 2, the size of the upper main rotor is the same as the that of the lower main rotor, the rotating speed of the upper main rotor is completely the same as that of the lower main rotor, and the rotating direction of the upper main rotor is opposite to that of the lower main rotor, so that the positive torque and the reverse torque are balanced, and the aircraft is prevented from rotating in the air. The fuel oil engine 7 is fixed to an aircraft arm 16, rotary shafts of the main rotors penetrate through the geometric center of the multi-motor aircraft, and the gravity center of an oil tank 15 is located in the geometric center of the multi-motor aircraft.

The disclosed embodiment provides a video generation method, an apparatus, an electronic device and a computer storage medium. The method includes: receiving a video shooting operation of a user through a video playing interface of an original video; in response to a video shooting operation, a video shooting window is superimposed on a video playing interface, a user video is captured, and the user video is displayed through the video shooting window; the user video and the original video are synthesized to obtain a co-shot video. In the embodiment of the present disclosure, a video shootingwindow can be superimposed and displayed on the video playing interface based on the video shooting operation of the user on the video playing interface of the original video, and the shooting of theuser's video can be completed through the video shooting window on the basis of the original video. Through the scheme, the user video and the original video can be easily and quickly synthesized intoa co-produced video, and the operation process is simple and fast, which can better meet the practical application needs of users and improve the interactive experience of users.

The invention discloses a machine vision based technical inspection and detection system and method for railway trains and belongs to the field of technical inspection on railway freight trains. According to the system, a light-duty rail, a portable moving device, a laser device, an optical image acquisition device and a wireless transmitting device are arranged outside a train inspection room; a wireless receiving device, a communication management server, a control server and a central server are arranged in the train inspection room. According to the system and the method, the automation and the intellectualization of technical inspection on the railway trains are achieved through controlling the portable moving device running on the light-duty rail, automatically acquiring image and video data on the bottom of a vehicle, carrying out data processing and outputting a current state of each component of the vehicle, an air testing test result and a repair processed result. According to the system and the method, manpower can be replaced for outdoor vehicle guarding operation, and the high quality, the accuracy and the reliability of train inspection can be guaranteed; the system and the method can adapt to various climatic conditions and natural environments and high-strength operation, so that the work efficiency of train inspection is increased.

The invention discloses a micro-nano manufacturing method of a compound eye structure micro lens array, wherein a hard material for manufacturing a micro lens array template is fixed at a three-dimensional precision displacement platform, so that an ultra-short pulse laser acts on the surface of the material to ablate a desired crater array; the hard material on which the crater array is ablated is put into hydrofluoric acid diluent to be corroded; a PMMA (Polymethyl Methacrylate) chloroform solution is coated on the surface of the template uniformly by using a plane casting process, and subsequently dried and solidified naturally; a small ball is heated; and the back plane of a PMMA membrane is uniformly extruded by a ball crown until the membrane wraps the half ball to form a semispherical structure, thus obtaining the compound eye structure micro lens array. The high repeatability and the mass production with low cost of the compound eye structure micro lens array are realized; and meanwhile relevant parameters are flexible and adjustable.

The invention discloses a preparation method and an application of a single-ion conductive polymer electrolyte membrane. The preparation method and the application are characterized by adopting heat initiated free radical polymerization, a polymerization system comprises lithium-containing monomers, poly(ethyleneglycol)methacrylate and a cross-linking agent, and a specific plasticizer is added toa reaction system before the reaction. The preparation method of the single-ion conductive polymer electrolyte membrane comprises steps as follows: the lithium-containing monomers, poly(ethyleneglycol)methacrylate, the cross-linking agent and a thermal initiator are dissolved in the plasticizer, and after heat initiated polymerization, the transparent self-supporting electrolyte membrane with certain mechanical strength is obtained. The preparation method and the application have the advantages as follows: a one-pot method is adopted for synthesis, the operation is convenient, and the steps are simple; the prepared electrolyte membrane has good mechanical strength, heat stability and electrochemical stability, and further has high ionic conductivity and large transport number of lithium ions; an assembled lithium metal battery shows excellent cycle performance.