685results about How to "Improve real-time performance" patented technology

The invention provides a visual localization and obstacle avoidance method and a visual localization and obstacle avoidance system for an unmanned plane. The method comprises the following steps that an unmanned plane-mounted camera acquires visual sensing information of the unmanned plane, and acquires inertial navigation data by an inertia measuring unit of the unmanned plane; a remote control system receives visual sensing information and inertial navigation data, and obtains obstacle information in a flight environment where the unmanned plane is positioned according to the visual sensing information; an external camera performs visual localization on the unmanned plane to obtain the position information of the unmanned plane; the remote control system plans the flight path of the unmanned plane according to the obstacle information and the position information of the unmanned plane, and generates a flight control command according to the inertial navigation data and the flight path; and the unmanned plane receives the flight control command to control the unmanned plane to avoid obstacles. According to the embodiment of the invention, the problems of visual localization and obstacle avoidance of the unmanned plane are solved, and the unmanned plane has the capability of completing visual obstacle avoidance by using the plane-mounted camera and a positioning camera.

The invention discloses a method and a device for processing a face image. The method disclosed by the invention comprises steps of positioning the face area characteristics of a human face, extracting the image of the human face area according to the area defined by the face characteristic points, performing conversion on the human face of an original image according to a face shape of an object image, performing conversion on the face image of the source image according to the skin color distribution of the face area of the object image, and obtaining a new human face through combining the face area image of the source image and the face area image of the object image. The embodiment of the invention realizes the face area image processing which has high similarity with the image of person, and the invention is simple in process, small in counting and high in instantaneity.

The invention provides a six-camera around looking-based cylindrical panoramic generation device and method. The device comprises a regular hexagonal prism enclosure; six cameras are horizontally and outwardly arranged at the centers of the six side surfaces of the regular hexagonal prism; and the six cameras are focus-fixed fisheye cameras with 90-degree horizontal view angles. The method comprises the following steps: capturing images of the six cameras simultaneously; carrying out distortion correction on the six images by using a reverse division model; carrying out cylindrical projection on the six corrected images; carrying out feature detection and matching on the adjacent images with superposed areas by using an SIFT feature point detection algorithm; calculating the image translation splicing parameters and determining the sizes of the superposed areas through the optimum matching point; carrying out image splicing and fusion simultaneously; and carrying out smooth transition on the public area by using a slow-in slow-out fusion method so as to finally obtain the naturally transitioned cylindrical panoramic images. According to the device and method, the natural panoramic images of 360-degree view can be efficiently generated, so that the device and method are suitable for panoramic monitoring.

The invention relates to a universal protocol adapting method comprising the following steps of: receiving an input message and judging the protocol type of the message; finding a protocol rule base according to the protocol type, analyzing the message and storing the content obtained through the analysis into a message parameter library; finding the protocol rule base according to a message record in the message parameter library, packaging a message frame, and sending the packaged target protocol message frame to a target address. The invention also relates to a universal protocol adapting device. According to the invention, a universal message structure is summarized from various field bus protocols, and a unified control and management interface is provided for various heterogeneous terminals and control networks, so that protocol conversion can be carried out between any two field bus protocols capable of being expressed through the universal message structure without depending on a concrete system platform, is more convenient and is better in real-time property.

The invention relates to the technical field of unmanned gyroplane autonomous tracking, in particular to an unmanned gyroplane moving target autonomous tracking device based on a pan-tilt camera and an airborne sensor. The unmanned gyroplane moving target autonomous tracking device comprises an aircraft and a ground station. The aircraft comprises an aircraft controller, an aircraft propeller motor connected with the aircraft controller, and a pan-tilt motor. The pan-tilt motor is connected with the pan-tilt camera. A wireless image transmission module and a wireless data transmission module are arranged between the aircraft and the ground station. The aircraft and the ground station are connected through the wireless image transmission module and the wireless data transmission module. Images obtained by the airborne pan-tilt camera serve as input of a system, coordinates of a target in the images are obtained on the ground station through a Struck algorithm, modeling is conducted on the coordinates of the target through a uniformly accelerated motion (CA) model, the control quantity of a pan-tilt is designed according to the coordinates, estimated through modeling, of the target, and accordingly tracking instantaneity and smoothness of the pan-tilt camera are guaranteed.

The invention discloses a method for achieving an Internet of things intelligent data gateway. The gateway achieves data collection in a device and the data are sent to a server at regular time or the server obtains the data at a certain fixed time interval. According to the gateway device, the defects that a traditional gateway device is complex in upper computer software programming, low in data collection speed, low in real-time performance and integrity of data collection and the like are overcome.

The invention discloses a convolutional neural network accelerator based on calculation optimization of an FPGA. The convolutional neural network accelerator comprises an AXI4 bus interface, a data cache region, a pre-fetched data region, a result cache region, a state controller and a PE array. The data cache region is used for caching feature map data, convolution kernel data and index values read from an external memory DDR through an AXI4 bus interface; The pre-fetched data area is used for pre-fetching feature map data needing to be input into the PE array in parallel from the feature mapsub-cache area; The result cache region is used for caching a calculation result of each row of PE; The state controller is used for controlling the working state of the accelerator to realize conversion between the working states; And the PE array is used for reading the data in the pre-fetched data area and the convolution kernel sub-cache area to carry out convolution operation. The accelerator utilizes the characteristics of parameter sparsity, repeated weight data and an activation function Relu to end redundant calculation in advance, so that the calculation amount is reduced, and the energy consumption is reduced by reducing the access memory frequency.

An electronic image stabilization method based on a visual attention mechanism includes the following steps: conducting foreground moving area detection on a reference frame, marking foreground moving sub blocks, extracting overall situation remarkable characteristic points in the reference frame, matching characteristic point pairs, removing error-matched characteristic point pairs, acquiring moving parameters, conducting moving filtering, conducting rapid moving compensation, and rebuilding undefined boundary information to acquire a panoramic image. By means of the method, compensation parameters are acquired through extraction, matching, checking and moving parameter calculation of the overall situation remarkable characteristic point pairs and self-adaptation filtering smooth movement, vision stability and definition between video frames are improved, instability of video sequences is removed or reduced, and observation effects of a video monitoring or tracking system is improved.

The invention relates to a method and a system for real-time publishing and sharing files in a mobile communication terminal, belonging to the technical field of mobile value-added application. The method and the system are characterized in that a content server stores published and shard files, the only extraction code of each file, a mobile communication terminal identification number of a user and subscription information. A short message push gateway and a content server interface realize the sending of bulks of short messages with an agreed format. The mobile communication terminal can analyze and generate the short messages with the agreed format, utilizes the analyzed short message content to automatically generate an HTTP request containing a content server domain name and the extraction code, and downloads the files from the content server. The method and the system can utilize the real-time characteristic of the mobile communication terminal short message and the convenience of network HTTP downloading to realize the sharing and publishing of remote files and especially large-capacity media files.

The invention provides a visual target identification method, comprising the following steps: inverting a collected original image to a binary image; calculating the enclosure frame of a visual target on the basis of the binary image; and searching characteristic points in the enclosure frame. The invention also discloses a visual target tracking method. The method comprises: pretermitting that the size of the search window of a zeroth frame is the same as that of the image; identifying a first frame image to obtain the enclosure frame; and predicting the search window. In the invention, an image processing method is used to calculate the enclosure frame and the characteristic points in the enclosure frame. Meanwhile, on the basis of a target tracking technology, the invention also provides a method based on a predictable search window, so as to predict and track the motion of the target, thereby obviously reducing the search range and improving the instantaneity of the target tracking method.

The invention relates to an ROS based service robot and indoor navigation method thereof. The service robot includes a mobile chassis, a housing, a support frame, an upper layer PC and a bottom layer PC, wherein the mobile chassis includes a wheel unit, a motor, a motor control panel, a gyroscope, an encoder, an electronic compass, a power source, an anti-collision module and a data collection module. The housing is equipped with a touch screen. The upper layer PC and the bottom layer PC are both equipped with a robot operation system ROS. The navigation method includes that the service robot acquires indoor environment information and creates an indoor two-dimensional map; a target position is set after the map is loaded in a visible window and the position of the robot itself is determined; an overall path is planned by utilizing an overall path planning interface provided by ROS and unknown barriers which are namely movable objects or unknown areas appearing in a navigation process are avoided by utilizing a partial path planning interface provided by ROS.

The invention discloses a sensor network real-time routing method based on data-driven link estimation, which comprise the following steps of: 1, perceiving geographical position information by a current node, and informing the geographical position information of a neighbor node; 2, when data are sent, generating a data message and a destination node, and when the data message is forwarded, acquiring the data message and the destination node from a media access control (MAC) layer; 3, calculating the distance from the current node to the destination node; 4, calculating the distance from the current node to the destination node in a neighbor set to generate a hopping candidate set; 5, calculating path speed from the current node to the node in the hopping candidate set; 6, selecting one node as a hopping node; 7, writing a node clock value into a data frame of the MAC layer; 8, sending the data message to the MAC layer; and 9, updating the hop delay, delivery rate and energy consumption of the corresponding nodes in the neighbor set. By the method, the real time of the transmission of network data is improved, and the energy consumption of the network is reduced.

The invention provides a comprehensive performance test rig of a vehicle electric power steering system. The comprehensive performance test rig includes a rack part, an electric control part and a computer control system; and the rack part comprises a base, a rotary table, a fixture clamp plate, a C-EPS fixture clamp, an input terminal driver, a C-EPS output terminal loading device, a P / R-EPS output terminal loading device, a fixture clamp bench, and an adverse impact part. The comprehensive performance test rig of the invention can comprehensively test three types of electric power steering systems, wherein the three types include a C-EPS, a P-EPS and an R-EPS. Compared with a conventional electric power steering test rig, the control system has good real-time performance, can provide the real-time vehicle dynamics simulation environment, and can provide the comprehensive high-precision multiple performance tests of three types of the EPS.

The invention discloses a method and device for determining a face area. The method comprises the following steps that: receiving a positioning request, wherein the positioning request is used for requesting to position a face area in a target picture; through a convolutional neural network, carrying out a face positioning operation on the target picture to obtain a positioning result, wherein the convolutional neural network is used for calling a graphics processor to carry out a convolutional operation on the target picture, and the face positioning operation comprises the convolutional operation; and under a situation that the positioning result is used for showing that the face area is in the presence in the target picture, returning a positioning result. By use of the method, the technical problem of poor face detection instantaneity in a relevant art is solved.

The invention provides a real-time data storage and query method. The method comprises the steps that when storage is conducted, partition is conducted on data to be stored by adopting consistency Hash sharding, primary index data carried with index information of the consistency Hash sharding are obtained from at least one distributed node of a distributed storage system, merger processing is conducted on at least two primary index data of one same distributed node to form a distributed file block corresponding to the distributed node, and a clustered index is established for the primary index data in the distributed file block; secondary index data carried with clustered index information are obtained from the distributed file block corresponding to the distributed node, and the secondary index data in the distributed file block are written in a disk in which the corresponding distributed node is located; when query is conducted, by means of the index information of the consistency Hash sharding established when the data to be inquired are stored and the clustered index information established in the distributed file block, a needed result can be located accurately and rapidly.

An aerial photography transport video vehicle rapid detection method comprises the steps that: step 100, the space-based coding part adopts the global motion estimation to determine the global motion vector of a background; step 200, the residual difference value is calculated according to the global motion vector to divide background regions and motion regions; step 300, whether the regions are the background regions or not is judged, an image of which the regions are all the background regions is changed to the next frame, and the step 200 is executed; otherwise, the motion regions of the image execute the step 400; step 400, the ground part firstly determines a self-adaptive gradient threshold which is lower and can correctly divide various objects to carry out the primary marker extraction; then two parameters of an area and a water collection depth are introduced for further screening of the extracted marker, thus determining a final marker point; then the marker point is taken as the region minimum value for carrying out the VS watershed division; finally, the regions are merged according to the texture information of the regions; step 500, the shadow is detected in an HSV color space, thus filtering out phony targets and finally detecting vehicles. The method of the invention solves the problems that the calculation amount during the decompression of the aerial photography image and the detection of motion targets is greater, and the real-time property and the robustness requirement are difficult to be satisfied at the same time.

The present invention discloses a self-adaption fault diagnosis method based on permutation entropy (PE) and manifold-based dynamic time warping (MDTW), enabling a bearing fault diagnosis process to be systematic and raising handleability and real-time performance of the diagnosis method. Firstly, a nonlinear and nonstationary bearing vibration signal is decomposed into a plurality of single-package components by applying an adaptive time-frequency analysis method; the adaptive time-frequency analysis method may be selected from empirical mode decomposition, local mean decomposition and local characteristic-scale decomposition methods; and then, extracting the PE of each single-package component as a fault signature. The PE can reflect complexity of the signal and has high robustness and rapidity. The MDTW method is provided by the present invention so as to rapidly and accurately measure distance test data and training data, thereby determining the current fault state and realizing bearing fault diagnosis, and the method has excellent practical engineering application values.

The invention relates to a linear test machine of a capacitance touch screen, which comprises a working platform, a movable touch and scratch device and a control system, wherein the capacitance touch screen to be measured can be placed at the top surface of the working platform, and the control system is connected with the movable touch and scratch device by a data line and can receive the output of the capacitance touch screen to be measured by a serial port. The linear test machine of the capacitance touch screen is characterized in that the movable touch and scratch device comprises an X-direction movement device, a Y-direction movement device, a Z-direction movement device and an induction pen device, wherein the Y-direction movement device is arranged on the working platform and can carry out linear back-and-forth movement in a Y direction on the working platform; the X-direction movement device is arranged on the Y-direction movement device and can carry out linear back-and-forth movement in an X direction; the Z-direction movement device is arranged on the X-direction movement device and can vertically slide up and down; and the induction pen device is arranged at the X-direction movement device and can touch and scratch the capacitance touch screen to be measured. The invention has high test accuracy and real-time property, simple structure and convenient use and can carry out accurate and comprehensive linear test on the capacitance touch screen.

The invention provides a dynamic page displaying method and a dynamic page displaying device. According to the embodiment, layout configuration data of a dynamic page of an application is obtained, wherein the layout configuration data includes native assembly data and layout data; and the layout configuration data is further subjected to analytical processing to obtain an analytical result, so that a dynamic page is output according to the analytical result. All assemblies displayed by the dynamic page are native assemblies, and page resources of the dynamic contents do not need to be loaded or rendered in a WebView assembly mode, so that the page displaying speed is accelerated; the time required by the page displaying is shortened; the page displaying real-time performance is effectively improved; the interaction effect is beautified; the user experience is greatly improved; and meanwhile, the memory consumption can also be reduced.

The invention discloses a visual system for realtime and accurate observation and analysis of table tennis rotating and a system operating method, wherein the system is composed of the following modules: an image acquisition and table tennis target identification and location module, a table tennis space motion KALMAN filter module, a pan-tilt motion plan and control module, a table tennis mark recognition and three-dimensional reconstruction module, a table tennis rotating motion particle filter module and a three-dimensional virtual scene reproduction module. The biggest characteristic of the system for the realtime and accurate observation and analysis of the table tennis rotating disclosed by the invention is that the system can observe and analyze the rotating information of the table tennis in real time, to obtain a rotating state of the table tennis on the basis of the space motion of the table tennis, so that abundant and complete table tennis track information is obtained; and the system is still quite high in realtime performance, to provide abundant and accurate information for athlete training and teaching, game realtime judgment and live broadcast, and timely response of an online decision-making automatic system. By introduction of observation and analysis on internal motion of the target, the accuracy of the target motion state estimation and predication is improved, so that the system can be popularized in other operation projects, civil, military and other similar scenes in a relatively scientific manner.

The invention relates to a co-phasing control device and control method for a spliced telescope.The device comprises a spliced telescope body, a first beam splitter, a wavefront detector, a second beam splitter, a translation error detector and a wavefront controller, wherein the first beam splitter receives emergent rays and generates first divided beams and second divided beams; the wavefront detector detects the first divided beams and obtains system aberration data and wavefront distortion aberration data of transmission paths of all spliced sub mirrors; the second beam splitter receives the second divided beams and generates third divided beams and fourth divided beams; the translation error detector detects the third divided beams and collects dispersion interference fringes between each pair of spliced sub mirrors; the wavefront controller is connected with the wavefront detector, the translation error detector and a driver of a secondary separating deformable mirror.By means of the co-phasing control device and control method for the spliced telescope, real-time detection and compensation and correction of translation, inclination and higher order aberration can be effectively achieved, and compared with the technology in a current similar system, better accuracy and instantaneity are achieved.

The invention discloses a trolley capable of automatically following a mobile terminal to walk. The trolley comprises a master control chip, an electronic compass module used for measuring azimuth information of the trolley, a bluetooth module used for transmitting azimuth information of the mobile terminal to the master control chip, an ultrasonic ranging sensor used for measuring distance difference between the trolley and the mobile terminal, a DC deceleration motor used for controlling the trolley to move forward or backward according to the distance difference and a steering engine used for controlling the steering angle of the trolley, wherein the electronic compass module, the bluetooth module, the ultrasonic ranging sensor, the DC deceleration motor and the steering engine are respectively connected with the master control chip. The steering angle of the trolley is calculated by the master control chip according to the difference value of the azimuth information of the trolley and the mobile terminal, and the calculation result is outputted to the driving circuit of the steering engine. Speed of the trolley is controlled by the DC deceleration motor after ultrasonic ranging, the azimuth information difference between the trolley and the mobile terminal is calculated after the electronic compass module acquires the azimuth information and finally the steering angle of the trolley is controlled by the steering engine so that the trolley has advantages of being high in real-time performance, easy to realize and low in cost.

The invention discloses a remote and safe intelligent household monitoring method and a device based on visual analysis. The method comprises steps: an indoor video image is acquired; one frame in the video image is extracted at an interval for household equipment recognition, and the operation state is judged; according to the acquired video image, whether a moving object exists is analyzed, in the case of existence, the moving object is tracked and photographed, and a high-definition picture of the object is acquired; the acquired high-definition picture is actively sent to a remote terminal, the high-definition picture is analyzed locally and the type of the moving target is recognized; the type is judged, if the moving object is a human being, and whether the human being is the resident is judged through face recognition; and according to the judgment result, a danger level is judged and corresponding information or alarm information is sent to the remote terminal. The device comprises a camera, an image processor, a controller, a data transmitting and receiving device and the remote intelligent terminal, wherein the remote terminal is connected with an information transmission device via Internet. The method and the device have the advantages of high real-time performance, wide application range, and easy mounting and maintenance.

The invention discloses a sulfur hexafluoride gas-filled type current transformer insulation state on-line monitoring system and method. A first switch solenoid valve of the current transformer is connected to a pressure stabilizing valve. An output terminal of the stabilizing pressure valve is connected to an infrared sensor and an electrochemical sensor. The electrochemical sensor is connected to a pressure pump through a second switch solenoid valve. The pressure pump is connected to a pressure regulating valve. The pressure regulating valve can communicate with a gas supplementing port of the transformer. The first switch solenoid valve can communicate with the gas supplementing port of the transformer. The method comprises steps that: a first hand-operated valve and a second hand-operated valve are opened, and the first switch solenoid valve is opened; when the pressure in the pressure stabilizing valve reaches one time the atmospheric pressure, the first switch solenoid valve is closed; SF6 purity and SO2 content in a gas sample are measured; the second switch solenoid valve is opened; the pressure pump is opened; the pressure regulating valve is opened, and the gas is delivered back to the current transformer; the second switch solenoid valve is closed; and the pressure pump is closed. With the monitoring system and the method provided by the invention, on-line monitoring is realized with good real-time performance. Tested gas sample is delivered back to the gas-filled type current transformer, such that gas loss and atmosphere pollution are reduced.

The invention discloses a malicious information transmission preventing method and system of a mobile intelligent terminal. The method comprises the following steps of: judging, namely judging whether an information transmission object is in a preset blacklist or whitelist or not when the application program of the mobile intelligent terminal is about to transmit information; and processing, stopping the transmission of the information if the information transmission object is in the preset blacklist and allowing the transmission of the information if the information transmission object is in the preset whitelist, wherein the blacklist is a list of application software untrusted by the mobile intelligent terminal, and the whitelist is a list of application software trusted by the mobile intelligent terminal. By applying the method and system disclosed by the invention, the transmission of the malicious information can be efficiently prevented.

The invention discloses a terahertz wave imaging system. Large-range two-dimensional multi-column scanning on the measured objected is realized by swinging a frame scanning plane mirror in the small-angle range, so the scanning imaging speed is increased, and the practicability of the system for scanning and imaging a human body in a real-time security check system is improved; the swinging range of the frame scanning plane mirror is small, so the stability of the whole system is improved and the imaging quality is improved; the measured object is irradiated by a terahertz radiation source, so that the imaging capability and the imaging quality of the system are improved; and a data receiving processing device corrects the image by using factors 1/cos beta, so that image distortion during acquisition is avoided and the imaging quality is improved.

The invention relates to an on-board data management system self fault-tolerance method. A on-board data management system self fault-tolerance mechanism is divided into three levels: data fault tolerance, part level fault tolerance and system level fault tolerance, and the operations are sequentially carried out during the fault tolerance. The data fault tolerance comprises three aspects: data protocol verification, 1553B bus interface fault tolerance and RS422 bus fault tolerance. The part level fault tolerance comprises three aspects: random-access memory (RAM) chip fault tolerance, software watchdog fault tolerance and signal clock fault tolerance. The system level fault tolerance comprises two aspects: data management computer self switching fault tolerance and remote placing unit self switching fault tolerance. The method provided by the invention provides the fault tolerance at the data level, the part level and the system level, and the fault tolerance processing can be carried out on the data, the main function module and the single machine equipment. Through the layered fault tolerance processing, the respective fault tolerance on faults at different harm degrees can be realized. Through the layered fault tolerance processing, the cost caused by faults at a low harm degree can be reduced, and the reliability is also improved because corresponding fault tolerance measures are respectively adopted for various faults.

The invention discloses a real-time scenic spot visitor flow counting method based on signaling information. The real-time scenic spot visitor flow counting method includes the steps that cell identities of all mobile network cells covering a scenic spot are obtained; position updating information of mobile users is obtained; the number of all the mobile users in the scenic spot is counted; the resident population of the scenic spot is excluded; the passing population of the scenic spot is excluded; the scenic spot visitor flow volume is defined. The number of the mobile users in the mobile network cells covering the scenic spot is monitored through signaling in a communication network, so that the number of visitors in the scenic spot is counted in real time; the real-time scenic spot visitor flow counting method is simple in system deployment, low in cost, convenient to maintain, high in realtime performance and accuracy and capable of counting the visitor flow volume of large-scale scenic spot or open type scenic spots.

The invention discloses a human body action real-time identification method based on an infrared image, and relates to the technical field of human body action recognition, and the method comprises the following steps: constructing an infrared image human body posture extraction network and an action recognition network based on skeleton features; acquiring an infrared pedestrian image data set and preprocessing the infrared pedestrian image data set, training an infrared image human body posture extraction network based on the preprocessed infrared pedestrian image data set, and obtaining a human body dynamic skeleton feature image; segmenting a region-of-interest sequence of the human body dynamic skeleton feature image, and training an action recognition network by utilizing a segmentation result; and acquiring a to-be-recognized infrared image, and performing action classification and prediction on the preprocessed to-be-recognized infrared image based on the infrared image human body posture extraction network and the action recognition network. According to the invention, the problems of poor real-time performance and low recognition rate of an existing behavior recognition method generally aiming at a visible light environment and performing human body behavior action recognition through an infrared image in a dark environment at night or in a severe weather environmentare solved.

The invention discloses a static target detection method based on two-background differencing. The method comprises the steps as follows: acquiring a video image; constructing a background model; detecting a moving target; establishing a pure background video image; detecting a foreground target; detecting a static target. Background modeling is performed by use of an improved Gaussian mixture model, and the method can be applied to accurate modeling of backgrounds having slow illumination change or repeated moving objects in complex scenes, particularly to conditions with slow illumination and weather change or higher moving object speed; by reducing the number of Gaussian distribution functions established for each pixel, the calculation amount is reduced, and the real-time performance is improved. Foreground object extraction is performed with a background differencing method, the principle and algorithm design are simple, the obtained result directly reflects the position, size and shape of a foreground object, and more accurate foreground object information can be obtained. A two-background model algorithm is adopted for static object detection, so that the method is lower in complexity and easy to implement.