2610results about How to "Meet real-time requirements" patented technology

The invention relates to a flight path planning method based on a sparse A* algorithm and a genetic algorithm and belongs to the technical field of flight path planning of an unmanned aerial vehicle (UAV). According to the characteristics of flight path planning, the method comprises the following steps: planning an initial reference flight path by the utilization of the sparse A*search (SAS) algorithm, wherein constraint conditions are combined into algorithm search, so that useless nodes in a search space can be effectively deleted and the search time is shortened; and when emergent threat occurs during real-time flight of the UAV, performing real-time flight path planning by the utilization of the genetic algorithm and generating a flight path with local optimum or approximate optimum until the threat disappears and the UAV returns the original global optimum reference flight path and continues flying. The method provided by the invention has high real-time performance and rapidity; the searched flight path is closer to the actual UAV optimal flight path; and the method can be applied to the technical fields of robot path planning, urban vehicle path planning and the like under complex environments.

The invention discloses a real-time panoramic image stitching method of aerial videos shot by an unmanned plane. The method comprises the steps of: utilizing a video acquisition card to acquire images which are transmitted to a base station in real time by an unmanned plane through microwave channels, carrying out key frame selection on an image sequence, and carrying out image enhancement on key frames; in the image splicing process, firstly carrying out characteristic detection and interframe matching on image frames by adopting an SURF (speeded up robust features) detection method with good robustness; then reducing the series-multiplication accumulative errors of images in a frame-to-mosaic image transformation mode, determining images which are not adjacent in time sequence but adjacent in space on a flight path according to the GPS (global positioning system) position information of the unmanned plane, optimizing the frame-to-mosaic transformation relation, determining image overlapping areas, thereby realizing image fusion and the panoramic image construction and realizing real-time effect of carrying out flying and stitching simultaneously; and in image transformation, based on adjacent frame information in a vision field and adjacent frame information in airspace, optimizing image transformation to obtain the accurate panoramic images. The stitching method has good real-time performance, is fast and accurate and meets the requirements of application occasions in multiple fields.

The invention discloses a logistics tracking method based on a block chain. The method comprises steps that a user sending node uploads parcel information to a main node; the main node generates a digital fingerprint for the parcel information uploaded by the user sending node, writes the digital fingerprint and a corresponding time stamp into a block chain, broadcasts the block chain in all nodes; each logistics node generates parcel state information successively according to the digital fingerprint, the node information of the logistics node and the node information of a next logistics node in accordance with a parcel physical space transfer path; a corresponding user receiving node signs for the parcel, generates parcel receiving information and uploads the parcel receiving information to the main node in order for the main node to write the parcel receiving information into the block chain. The logistics tracking method applies the block chain technology to the logistics field, guarantees the traceability of the authenticity and the state of logistics information by using the unchangeable characteristic of the block chain, saves the logistics information and update state information fast and efficiently according to the characteristic of a private chain in order to satisfy a requirement for logistics instantaneity.

The invention relates to a fast multilevel imagination and reality occlusion method at an actuality enhancement environment, comprising the following steps of: acquiring videos of a real scene by using a two-path video camera; extracting a pair of keyframes from a two-path video stream at set intervals to carry out the resolving of a dense depth map, building a three-dimensional model of a real object participating in occlusion, and extracting sparse feature points; tracking the sparse feature points of all middle frames of the two-path video stream, and estimating the posture of a current camera by combining with the positions of the sparse feature points in an image; acquiring the three-dimensional information of the real object according to a model built for the last time; moving and rotating the model of the real object according to the posture of the current camera; carrying out depth relation comparison by utilizing the three-dimensional information of the recently regulated real object with a registered three-dimensional virtual object; and respectively processing the three-dimensional tracking of the middle frames and the three-dimensional reconstruction of the keyframes in parallel at different threads. The invention is suitable for unknown and variable environments without modeling in advance and can meet the requirements for real time.

The invention discloses an industrial robot workpiece positioning grabbing method based on visual guidance. The method includes the following steps that 1, a system parametric model is established, and calibration of a camera is conducted; 2, a workpiece feature template is determined; 3, workpiece template living examples are sought for, and the positions of workpieces are determined according to coordinate information of the living examples; 4, the speed of a target workpiece is calculated; 5, the pose, in a robot basic coordinate system, of the workpiece is predicted when the workpiece is located at a station to be grabbed; 6, a robot moves according to a planned path to approach and grab the workpiece and place the workpiece at the position of a target point. The invention further discloses a system for implementing the industrial robot workpiece positioning grabbing method based on visual guidance in the patent claim 1. The system comprises a conveying belt, an optoelectronic switch, a camera, an industrial control computer, the robot and the target workpiece. The system has the advantages of being high in positioning accuracy, high in working efficiency, high in automation degree and the like.

A joint search method for UAV multiobjective path planning in an urban low altitude environment first constructs a static safety index map based on static known obstacles. Meanwhile, based on proactively detected obstacles that are not marked on a geographic map by a UAV, the method constructs online a dynamic safety index map. Second, a multiobjective path planning problem is solved using a joint offline and online search method. Moreover, this method first plans offline the least cost path from a starting point to an end point and then invokes the online search scheme to replan online a changed path when the UAV detects unknown obstacles. Thus, the UAV can avoid dynamic obstacles effectively. The online search scheme has a small search space and can quickly replan a safe path for the UAV, thus satisfying the requirement of UAV on the real-time path planning.

The invention provides a wood defect detection method based on deep learning. The method comprises the following steps of collecting images; segmenting the images into image blocks with a same size; selecting defect image blocks with different types and non-defective image blocks as a training sample set; using the training sample set to train a deep learning algorithm in an off-line mode; and using a trained deep learning algorithm to detect and identify defects of a wood image in an online mode and the like. In the invention, through the powerful deep learning algorithm, defects on different complex texture wood surfaces are detected and identified in high precision and online modes and a problem that a traditional image processing algorithm can not solve is solved. The invention also provides a wood defect detection system based on deep learning. Through cooperation among an image acquisition module, a deep learning algorithm processing module and a control execution module, a detection speed can be effectively accelerated and practicality is increased.

The invention discloses a multiline laser radar-based 3D point cloud segmentation method. The method comprises the following steps of: 1) scanning 3D point cloud data in a 360-degree range by utilizing multiline laser radar, establishing a Cartesian coordinate system OXYZ, converting the 3D point cloud data under the Cartesian coordinate system, and pre-processing the 3D point cloud data under the Cartesian coordinate system so as to determine a region of interest in the 3D point cloud data; 2) filtering suspended obstacle points in the region of interest by utilizing the statistic characteristics of adjacent points; 3) constructing a polar coordinates grid map, mapping the 3D point cloud data, the suspended obstacle points of which are filtered, into the polar coordinates grid map, and segmenting non-ground point cloud data from the 3D point cloud data in the polar coordinates grid map; and 4) voxelizing the non-ground point cloud data by utilizing an octree and carrying out clustering segmentation by adoption of an octree voxel grid-based region growing method. The method can improve the operation efficiency, is high in detection precision and strong in reliability, and can be widely applied to the technical field of vehicle environment perception.

The invention relates to a virtual-reality occlusion handling method based on depth image data flow. The virtual-reality occlusion handling method comprises three parts of construction of a scene point cloud model, three-dimensional space registration and virtual-reality occlusion handling and rendering. Firstly filtering and other processing operation are performed on depth data acquired by a depth camera and the normal vector of each point is calculated; then the camera attitude is calculated by using an iterative closest point algorithm according to the point cloud carrying the normal vector and the point cloud obtained by projection from a three-dimensional scene model through the last frame of camera attitude; then the point cloud of the current frame is fused into the three-dimensional scene point cloud model; when the scene is reconstructed, the color image feature points acquired by the depth camera are calculated in real time and three-dimensional space registration is performed by matching with the template image feature points; and then the space position relation and the occlusion relation of the virtual object and the three-dimensional scene are processed by the obtained camera attitude and rendered in real time. The method can be operated in real time on the present mainstream equipment, and the great virtual-reality occlusion effect can also be obtained when the resolution of the input data is low.

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

The invention provides a solar silicon wafer defect detecting system. The solar silicon wafer defect detecting system comprises a visual image acquisition system, an image processing system and a sorting system; the visual image acquisition system is used for automatically acquiring images of silicon wafers on an automatic silicon wafer production line; the image processing system is used for analyzing the acquired silicon wafer images, automatically identifying the silicon wafer images and sending identification signals to the sorting system; and the sorting system is used for executing the signals of the image processing system, sorting out defective silicon wafers and sorting the silicon wafers with the same defects into same sorting boxes to realize accurate sorting of the silicon wafers. The invention also provides a detecting method based on the solar silicon wafer defect detecting system disclosed by the invention; the silicon wafers can be conveyed rapidly and rapid defect detection of the silicon wafer images can be carried out; and a high-speed and high-precision visual detection requirement of the production line can be met.

The invention relates to an airline routing algorithm of an unmanned plane in a large-range complicated three-dimensional space. The algorithm comprises the following steps: first dividing a three-dimensional space on the basis of a 2.5-dimensional grid (each grid point includes longitude, altitude and elevation information); improving the A* cost function by comprehensively considering the influence factors such as airline height, detected probability, airline distance under the constraint conditions of radar, disaster weather and forbidden areas, and determining an initial airline on the basis of an algorithm search flow; serially processing the initial airline to obtain a final flying airline for meeting the performance constraint of the unmanned plane (such as minimal step length, turning radius, climbing rate, safety height). The algorithm is good in stability and astringency, high in efficiency, capable of meeting the wide-range low-altitude airline routing requirement and applicable to relevant fields such as military, national defense, emergency relief.

The invention discloses a cloud computing task scheduling method facing real-time demand change and belongs to the technical field of cloud computing. According to the difference of demands of tasks required by users on real-time performance, the priority of the required tasks is divided into three levels according to values calculated according to a task classification function, tasks of different levels are distributed to corresponding virtual machines by adopting different scheduling strategies, and the demands of the tasks required by the users on the real-time performance are met. Meanwhile, a non-idle virtual machine cluster and a virtual machine buffering queue are built, the non-idle virtual machine cluster is used for normal task scheduling, and the virtual machine buffering queue achieves buffering of the non-idle virtual machine cluster. Under the condition that execution task load is large, a new virtual machine can be added into the non-idle virtual machine cluster at any time, and computing tasks are executed. When execution task load is small, a part of virtual machines can be closed and exit the virtual machine cluster, and efficient use of virtual machine resources is achieved.

The invention provides a news recommendation system and method based on an FOLFM model. Based on a content recommendation method, a news content model is expressed abstractly through a latent class model and content characteristics, and an individual latent class preference model is built for each user. Real-time training is carried out on a real-time behavior record of a user to obtain preference, on certain latent class news, of the user, whether the news is recommended to the user is determined through calculation, and a final news recommendation list is obtained after a series of processing processes. The news recommendation system and method based on the FOLFM model deeply excavate the interest of the user, improve recommendation accuracy and satisfaction of the user, avoid a cold starting problem of the news, and guarantee performance under the condition that the recommendation effect is improved as much as possible. The experiment shows that the news recommendation system and method based on the FOLFM model not only guarantee the requirements for high accuracy and high speed, but also realize visual real-time recommendation for the user.

The invention discloses an image type fire flame identification method. The method comprises the following steps of 1, image capturing; 2, image processing. The image processing comprises the steps of 201, image preprocessing; 202, fire identifying. The fire identifying comprises the steps that indentifying is conducted by the adoption of a prebuilt binary classification model, the binary classification model is a support vector machine model for classifying the flame situation and the non-flame situation, wherein the building process of the binary classification model comprises the steps of I, image information capturing;II, feature extracting; III, training sample acquiring; IV, binary classification model building; IV-1, kernel function selecting; IV-2, classification function determining, optimizing parameter C and parameter D by the adoption of the conjugate gradient method, converting the optimized parameter C and parameter D into gamma and sigma 2; V, binary classification model training. By means of the image type fire flame identification method, steps are simple, operation is simple and convenient, reliability is high, using effect is good, and the problems that reliability is lower, false or missing alarm rate is higher, using effect is poor and the like in an existing video fire detecting system under a complex environment are solved effectively.

The invention discloses a dual-card dual-standby terminal and a data communication method. The dual-card dual-standby terminal comprises a first SIM card, a second SIM card, a first protocol stack, a second protocol stack, a data service module and a TCP/IP protocol stack. The TCP/IP protocol stack comprises a first virtual WAN transmission interface and a second virtual WAN transmission interface. The first SIM card resides in a PS domain and a CS domain of a network through the first protocol stack. The second SIM card resides in a PS domain and a CS domain of a network through the second protocol stack. The first virtual WAN transmission interface and the second virtual WAN transmission interface are used for transmitting data of the first protocol stack and the second protocol stack respectively. The data service module is used for selecting the network where the first SIM card or the second SIM card belongs to establish a PS data service link so as to carry out data transmission. The terminal and the method have the following advantages that the SIM card can resides in a CS voice service and a PS data service simultaneously; transmission interruption and other problems generated because one network is abnormal can be avoided; a resource utilization rate is increased and a user experience is improved.

The invention discloses a method for identifying virtual malicious nodes and a virtual malicious node network in social networks. The method includes the following steps: (1) obtaining the attribute data, behavioral data and communication data of unidentified accounts from target social networks; (2) calculating the creditworthiness of each unidentified account according to the credibility model feature vector calculated using the extracted data; (3) making a comparison between the user behavioral habit statistic data of each unidentified account with the credibility being lower than a set threshold value and that of a normal user, and judging whether the unidentified account is a virtual malicious node or not; (4) sorting the virtual malicious node set and establishing association for the virtual malicious node in each sort result, so as to form a virtual malicious node network; then evaluating by using the Bayesian network algorithm to determine the final virtual malicious node network. The method provided by the invention can be used to effectively identify highly-disguised malicious nodes and a collaborative virtual malicious node network.

The invention provides a multi-feature united sparse represented target tracking method. The multi-feature united sparse represented target tracking method comprises building a primary dictionary; performing partitioning process on target modules; extracting candidate particles; extracting target characteristics; confirming the number of image characteristics and the number of block categories; performing nucleating process on the characteristics; performing block sparse representation on candidate samples in the dictionary; performing nucleus expansion; solving sparse problems; performing residual calculation on blocks; building likelihood functions; and updating template bases. The multi-feature united sparse represented target tracking method is analysis and improvement of utilized target characteristics and the traditional sparse coefficient solving method through a sparse encoding tracking device. According to the multi-feature united sparse represented target tracking method, stability of target tracking is maintained and accuracy of the target tracking device is improved under complex conditions that the illumination influence is large and the target is seriously shielded, and the accuracy of the algorithm and stability of the tracking are improved.

The invention discloses the technical field of wind power prediction, particularly, relates to an ultrashort-term power prediction method. The method comprises the following steps of: firstly, acquiring the wind speed, the wind direction and the wind power of a wind power farm to form a sample set; then, preprocessing the data of the sample set; reducing dimensions of the preprocessed sample set by a depth autocoder network; and finally, training a relevance vector machine regression model by the sample set with reduced dimensions, and predicting the ultrashort-term wind power through the trained relevance vector machine regression model. The method reduces the training time of a prediction model, satisfies the requirements on precision and real-time property in system status estimation, and enables the prediction method to be more accurate.

The invention provides a method and system for achieving integrated camera automatic focusing, and provides a zooming tracking method based on a relationship type, namely for a shot object at a fixed object distance, according to the initial zooming multiple, the value of an initial focusing position and a plurality of zooming tracking curves stored in advance, the mapping relation between the object distance interval and the focusing position is constructed, and the position to be focused is calculated in real time. Then, a gradient control function obtained by the combination of the statistical law of a lot of experimental data is utilized to achieve mobility control over a zooming motor and a focusing motor, the value of the position to be focused is improved and amended to a certain degree, the value of the position to be focused is closer to a real zooming tracking curve, and an automatic focusing effect is more accurate. Meanwhile, the zooming tracking curves are not needed to be stored in advance, all the zooming tracking curves corresponding to different object distances are not needed to be calculated point by point, an algorithm is simple, it is possible that the zooming speed meets a real-time property, and the focusing effect can meet the requirement for clearness in the whole process.

The invention provides a method and a device used for video enhancement. The method comprises the steps as follows: a nonlinear tensioning curve mapping table is arranged; for each frame of image, the histogram balance mapping table of the frame of image is calculated according to the brightness data; subsequently, the mapping value of at least each brightness value of a brightness value set of the pixel point of the frame of image in the nonlinear tensioning curve mapping table and the histogram balance mapping table is weighted and averaged; furthermore, a brightness mapping table is used for recording the corresponding relationship between each brightness value and the mapping value after weighted-average; subsequently, according to the brightness mapping table, the brightness value of the pixel point of the frame of image is updated as the mapping value after weighted average corresponding to the brightness value. The device of the invention comprises a brightness enhancement module which further comprises a nonlinear curve tensioning table memory unit, a histogram balance mapping table calculation unit, a brightness mapping table calculation unit and an image brightness enhancement mapping unit. The method and the device of the invention can improve the visible effect of the video images under the condition of low illumination and normal illumination.