Patents
Literature
Hiro is an intelligent assistant for R&D personnel, combined with Patent DNA, to facilitate innovative research.
Hiro

41results about How to "Precise target position" patented technology

Forward-looking radar imaging sea-surface target key point detection and recognition method

The invention discloses a forward-looking radar imaging sea-surface target key point detection and recognition method. The method includes the steps that radar echo data are quantized into a gray scale image; a region of interest is extracted from the gray scale image, partition is performed, and a target region partition image is obtained; the radar echo data and the region of interest of the image are used, and a target region peak point information matrix is obtained; information merging is conducted on the target region partition image and the target region peak point information matrix, the number K of peak points in a target region in a merging result is obtained through counting, the previous K peak points are selected and serve as target effective peak points and form a target effective peak point image in a binaryzation mode, target axial features are extracted, and a target position is determined; the center of gravity of target energy is calculated and serves as the target key point. According to forward-looking radar target features, multiple mode identification methods are used comprehensively, the inherent features of a target can be reserved, meanwhile interference factors such as artifacts and side lobes can be restrained, and the recognition accuracy and the positioning precision of radar imaging sea surface target key points are improved.
Owner:HUAZHONG UNIV OF SCI & TECH

Pedestrian detection method based on HOG and D-S evidence theory multi-information fusion

Provided is a pedestrian detection method based on HOG and D-S evidence theory multi-information fusion. The method is characterized by carrying out gray conversion on an original video frame image; then, carrying out detection on the converted image by utilizing a classifier based on HOG features to obtain a preliminary pedestrian detection result; based on the result, judging whether a region is supposed to have a pedestrian, but the region is not detected through an inter-frame relation matrix; then, detecting the region by utilizing a classifier based on LBP features to make up for the loss due to HOG features; and if the region is not detected under such condition, carrying out predication by utilizing historical data, that is, the data obtained through the inter-frame relation matrix, and by utilizing a Kalman filter to obtain the position of the pedestrian target, and enhancing accuracy of judgment by utilizing the D-S evidence theory and through fusion of multiple information of detection and tracking and the like. The method can detect the corresponding pedestrians accurately, and has a good effect under the condition of having partial shielding; and accuracy, robustness and anti-interference capability of pedestrian detection are enhanced.
Owner:安徽中科星驰自动驾驶技术有限公司

Cow face and cow face key point detection method based on Mask-RCNN

The invention discloses a cow face and cow face key point detection method based on Mask-RCNN. The blank of cow face and cow face key point synchronous detection is filled. The cow face and cow face key point detection method based on Mask-RCNN comprises the following steps: 1) collecting pictures with cow faces, and marking the cow faces and cow face key point data in each picture: (1) collectingthe pictures with the cow faces from a domestic network, and selecting the pictures with the cow faces from the collected pictures: a, selecting a video related to cow raising from domestic agricultural programs, intercepting one picture every five frames, and selecting the pictures with the cow faces from the pictures with the cow faces; b, collecting a video of the cow on site by using a camera, intercepting a picture every five frames, and selecting a picture with a cow face from the pictures; (2) marking the position of each cow face in the picture; 2) designing a cow face and cow face key point detection network structure, and 3) training a cow face and cow face key point detection network by using the marked data to finally generate a cow face and cow face key point detector.
Owner:JILIN UNIV

Text key information extraction method and device, computer equipment and storage medium

The invention relates to a text key information extraction method and device, computer equipment and a storage medium. The method includes: obtaining a current text of the key information to be extracted; performing word segmentation processing on the current text to obtain a target word corresponding to the current text; obtaining each target word vector corresponding to the target word, and combining the corresponding target word vectors in the order of the current text according to the target words, to obtain a target vector matrix. The target vector matrix is processed according to the trained target neural network model, and the target position corresponding to the key information to be extracted is obtained. Among them, the target neural network model performs feature extraction in units of target word vectors. The feature extraction direction of the target neural network model is consistent with the order of the target word vectors in the target vector matrix; the target word corresponding to the target position is extracted from the current text as key information. The above method improves the accuracy of extracting key information.
Owner:TENCENT TECH (SHENZHEN) CO LTD

Remote sensing target detection method based on boundary constraint CenterNet

The invention provides a remote sensing target detection method based on boundary constraint CenterNet, which is used for solving the technical problems of relatively low detection precision and recall rate of dense small targets in the prior art. The method comprises the following implementation steps: obtaining a training sample set; the method comprises the following steps: constructing a boundary constraint CenterNet network; obtaining a prediction label and an embedded vector of the training sample set; calculating the loss of the boundary constraint CenterNet network; carrying out the training of a boundary constraint CenterNet network; and obtaining a target detection result based on the trained boundary constraint CenterNet network. Through performing maximum pooling in the constrained pooling area through the corner constraint pooling layer, the fine features around the target are extracted, the detection precision and recall rate of dense small targets are effectively improved, meanwhile, the boundary constraint label generated by the boundary constraint convolutional network is utilized to constrain the prediction box, a more accurate target prediction box is obtained, and the detection precision of the target is further improved.
Owner:XIDIAN UNIV

Casting surface defect identification method based on deep convolutional neural network

The invention discloses a casting surface defect identification method based on a deep convolutional neural network. The method comprises the following steps: 1, collecting a casting surface defect image, marking the image, and establishing a data set of common casting surface defects; 2, constructing a deep convolutional neural network defect recognition model; 3, constructing a network loss function; 4, dividing the data set into a training set and a test set, and training the defect recognition network by using the training set; 5, inputting the test image into the trained network to identify the position, the type and the size of the defect. According to the invention, the recognition precision and recognition performance of the casting surface defects are improved, and the online, intelligent and automatic development of casting quality detection is promoted.
Owner:SOUTHEAST UNIV

Conveyor belt tracking method, robot and computer-readable storage medium

The embodiment of the invention relates to the field of robots, and discloses a conveyor belt tracking method, a robot and a computer-readable storage medium. According to the conveyor belt tracking method provided by the invention, an encoder and a robot are correspondingly arranged on a conveyor belt. The method comprises the steps that the robot acquires encoder values from the encoder at different times; the robot converts each encoder value into the displacement value of the conveyor belt; the robot calculates the relationship between the displacement value based on the conveyor belt and the displacement time according to each displacement value and the corresponding acquisition time; and the robot calculates the displacement value of the conveyor belt through the relationship. According to the conveyor belt tracking method, the robot and the computer-readable storage medium, the robot can acquire a more accurate target position when tracking the conveyor belt.
Owner:SHANGHAI SIGRINER STEP ELECTRIC

H5 page processing method and device, electronic equipment and storage medium

The invention provides an H5 page processing method and device, electronic equipment and a storage medium, and aims to solve the problem that a virtual keyboard called out from an H5 page shields an input box. The method comprises the steps that when it is detected that an H5 page comprising an input box is entered, sending a monitoring starting instruction to a Native layer, and the monitoring starting instruction is used for indicating the Native layer to monitor an operation event of the input box; receiving event information of an operation event of the input box returned by the Native layer when the Native layer monitors that the operation event of the input box occurs; and determining a target position of the input box according to the event information, and moving the input box to the target position. According to the method, the H5 page layer is set to delegate the Native layer to monitor the operation event of the input box, the event information is transmitted back to the H5page layer, the target position, determined by the H5 page layer according to the event information, of the input box is more accurate, a virtual keyboard can be prevented from shielding the input box, and the user experience is improved.
Owner:BEIJING CHENGSHI WANGLIN INFORMATION TECH CO LTD

Rib image detection method, computer equipment and storage medium

The invention relates to a rib image detection method, computer equipment and a storage medium. The method comprises: obtaining an original medical image, wherein the original medical image comprisesribs; carrying out expansion processing on a rib in the original medical image to obtain an expanded image of the rib; inputting the expanded image of the rib into a first neural network model for processing to obtain a candidate position of a region of interest; performing region division on the original medical image according to the candidate position of the region of interest to obtain a candidate image region, wherein the candidate image area is an image area including a candidate position of the region of interest on the original medical image; and inputting the candidate image region into a second neural network model for processing to obtain a target position of the region of interest. By adopting the method, the detection time can be saved and the detection accuracy can be improved.
Owner:SHANGHAI UNITED IMAGING INTELLIGENT MEDICAL TECH CO LTD

Crossroad non-motor vehicle and pedestrian position detection method based on special-purpose short-distance communication

The invention discloses a crossroad non-motor vehicle and pedestrian position detection method based on special-purpose short-distance communication. The method comprises the following steps: disposing a road side special-purpose short-distance communication unit and a roadbed computer unit at a crossroad turning area, wherein the road side special-purpose short-distance communication unit and the roadbed computer unit are connected; the road side special-purpose short-distance communication unit obtaining crossroad area wireless signal data frames; and by taking the data frames as input signals, performing real-time signal spectrum analysis, window function weighing and multipath signal inhibition processing through the roadbed computer unit such that time sequence based non-motor vehicle and pedestrian position detection in the crossroad turning area is realized. The method provided by the invention has the advantages of being free of space-time change influences, effectively utilizing the wireless signal resources of the crossroad area, and improving the safety grade of a crossroad.
Owner:BEIHANG UNIV

Variable shelter template matching algorithm of video target tracking

The invention specifically relates to a variable masking template matching algorithm in video target tracking which belongs to the field of computer vision and pattern recognition technology. In video target tracking, the target is obscured often by other object partly. In such circumstances, using the traditional fixed masking template matching to determine target location will cause accuracy decreased significantly. In order to realize target location accurately under a condition that the target is obscured partly, the invention combines with block analysis technology, on the basis of the first non-exact match result using variable masking template matching algorithm to do correction, the template masking in the process of correction changes dynamically as the change of candidate target location, thus causing non-masking part of the target always can guide the matching algorithm effectively to find precise target location. Confirms effectiveness of the invention algorithm based on experimental result of real video streaming.
Owner:FUDAN UNIV

Video target detection method and device based on sparse foreground prior, and storage medium

The invention discloses a video target detection method based on sparse foreground prior, a storage medium and equipment. A foreground extraction method based on orthogonal subspace learning is adopted to calculate and obtain a sparse foreground prior graph corresponding to each frame in a video; utilizing the ResNet feature extraction network and the feature pyramid structure to obtain semantic enhancement feature maps of the video frame and the sparse foreground image thereof; cascading the semantic enhancement feature map of the sparse foreground prior map with the semantic enhancement feature map of the current frame, and performing convolution fusion operation to obtain foreground prior fusion features of the current frame; mapping on each pixel of the foreground prior fusion featuremap to generate a candidate anchor box; and inputting the foreground prior fusion features and all anchor boxes into the trained classification and regression sub-network to obtain the category and position coordinates of the target object. The sparse foreground prior of the video data is fully mined, and the target detection accuracy is improved.
Owner:XIDIAN UNIV

Positioning method and device, robot, storage medium and positioning system

The invention provides a positioning method and device, a robot, a storage medium and a positioning system. An identification code is arranged at a preset position point of the target area; the identification code is generated according to the position information of the position point in a preset map of the target area; the position information of the position point is matched with the scale granularity of the map; the identification code acquired by the mobile device is the accurate position information in the target area, and meanwhile, the depth image acquired by the mobile device comprises the imaging of the position point, so that the accurate position of the mobile device in the map of the target area can be obtained by combining the identification code and the depth image. Besides,the target position is determined by receiving the signal transmitted by the signal transmitting device preset at the target position, so that the target position is accurate, and the mobile device can reach the accurate target position according to the navigation route from the initial position to the target position.
Owner:北京云迹科技股份有限公司

Superpixel substantial object detection algorithm based on area energy

The present invention provides a superpixel substantial object detection algorithm based on the area energy. The algorithm comprises: decoding the input video information to be observed to be an independent image frame sequence in a RGB format; setting a substantial algorithm parameter, and extracting a substantial area through the substantial algorithm; setting an area energy algorithm parameter, determining a pre-estimation energy threshold value of a horizontal direction and a pre-estimation energy threshold value of a vertical direction, dividing the target area threshold value of the substantial area, and extracting the substantial area of the substantial detection object; calculating the energy concentration degree of the energy substantial area; setting a superpixel segmentation input parameter , performing superpixel segmentation of the image frame, and obtaining a superpixel segmentation diagram; and according to the energy concentration degree, extracting the superpixel areas satisfying the energy concentration degree to form a final substantial object result and realize a substantial object detection process. The superpixel substantial object detection algorithm based on the area energy can effectively eliminate the background information and extract the observation target so as to provide more effective characteristic information criteria for the target tracking and identification.
Owner:DALIAN NATIONALITIES UNIVERSITY

Liquid crystal panel layering detection equipment and method

The invention provides (a) liquid crystal panel layering detection equipment and method. The equipment comprises a base, a movable carrying table, a natural light source camera assembly, a first ultraviolet light source camera assembly, a second ultraviolet light source camera assembly, an upper light source assembly and a backlight source assembly, the movable carrying table is an air floating platform and is arranged above the base, the natural light source camera assembly, the upper light source assembly and the first ultraviolet light source camera assembly are all located at one side of the movable carrying table, the first ultraviolet light source camera assembly and the second ultraviolet light source camera assembly are symmetrically arranged with the liquid crystal panel as the axis, and the second ultraviolet light source camera assembly and the backlight source assembly are both located on the other side of the liquid crystal panel; the upper light source assembly comprises a first natural light source assembly and a first ultraviolet light source assembly, and the backlight source assembly comprises a second natural light source assembly and a second ultraviolet light source assembly. The defects on the surface of the liquid crystal panel and the defects in the liquid crystal panel can be automatically distinguished, and the detection efficiency is improved.
Owner:HEFEI SHANGJU IND EQUIP +2

Solar photovoltaic module hot spot identification method based on infrared and visible light images

ActiveCN113076816ASolve the problem of not being able to quickly analyze the cause of hot spotsAvoiding the Problem of Insufficient Detection CapabilitiesCharacter and pattern recognitionNeural architecturesImaging dataMaterials science
The invention relates to a solar photovoltaic module hot spot identification method based on infrared and visible light images. The method comprises the following steps: establishing a photovoltaic panel infrared image data set and a photovoltaic panel visible light image data set; constructing a hot spot detection model, and training the hot spot detection model by using the photovoltaic panel infrared image data set; constructing a hot spot cause analysis model, and training the hot spot cause analysis model by using the photovoltaic panel visible light image data set; and obtaining an infrared image and a visible light image of the same position of the photovoltaic panel to be identified, sending the infrared image into the hot spot detection model, obtaining hot spot positions and drawing the hot spot positions in the visible light image, analyzing an image in the hot spot positions of the visible light image by using the hot spot cause analysis model, and obtaining hot spot causes corresponding to the hot spot positions. Compared with the prior art, the method effectively obtains the hot spot position in the infrared image, analyzes the hot spot cause in the visible light image based on the hot spot position, and improves the recognition accuracy and recognition efficiency.
Owner:SHANGHAI UNIVERSITY OF ELECTRIC POWER

Motion capture method and device, electronic equipment and computer readable storage medium

The embodiment of the invention relates to the technical field of attitude detection. The invention discloses a motion capture method and device, electronic equipment and a computer readable storage medium, and the method comprises the steps: obtaining the angular speed of a first target limb of a first motion object at a current moment through inertial motion capture equipment, and predicting thefirst position information of the first target limb at a target moment according to the angular speed; then, determining second position information of the first target limb at the target moment based on a first preset optical mark point of the first target limb through optical motion capture equipment; then, according to the first position information and the second position information, determining the target position of the first target limb at the target moment so as to capture the action of the first target limb; according to the method provided by the embodiment of the invention, errorsand offsets introduced by inertial motion capture equipment can be reduced as much as possible, and the motion capture accuracy is greatly improved.
Owner:BEIJING BYTEDANCE NETWORK TECH CO LTD

Target tracking method and system based on composite convolutional network

PendingCN114565639ATracking is accurate and robustRich feature representationImage enhancementImage analysisData setEngineering
The invention discloses a target tracking method and system based on a composite convolutional network. The method comprises the following steps: constructing a twin composite convolutional feature extraction sub-network Siam-Co-CNNs; inputting feature maps of the template branch and the detection branch extracted by the Siam-Co-CNNs into a region proposal sub-network to form a target tracking network based on a composite convolutional network; performing offline pre-training on a target tracking network based on the composite convolutional network by using a video frame sequence in the training data set; and converting the test video into an image frame input system, marking a to-be-tracked target on a first frame, carrying out online tracking, and obtaining and outputting a target tracking result. Experiments prove that the target tracking method provided by the invention can improve the accuracy and success rate of target tracking.
Owner:XIDIAN UNIV

Full-autonomous lunar surface navigation and data communication method based on spherical satellite

ActiveCN112394381AMeet the mission requirements of manned moon landing activitiesEffective coverageSatellite radio beaconingLocation information based servicePhysicsData transmission
The invention discloses a full-autonomous lunar surface navigation and data communication method based on a spherical satellite, and the method comprises the steps: selecting a lunar circular orbit asa navigation satellite operation orbit, effectively covering the lunar surface, and guaranteeing that any point of the lunar surface can receive an effective navigation satellite signal at any moment; under the condition that the initial state of the spherical satellite is known, an orbit prediction value within a long time can be obtained. Besides, inversion of kinetic model parameters is carried out through inter-satellite distance measurement, a high-precision kinetic model can be obtained, and high-precision forecasting of the constellation ephemeris is achieved so as to support long-termfull-autonomous operation of the constellation. Moreover, the spatial position of the lunar surface user is obtained by measuring and solving the distance between the lunar surface user and the plurality of spherical satellites observed by the lunar surface user, so that accurate target position and speed can be provided for lunar surface activities, efficient data communication service can be implemented, and navigation and positioning of any point of the lunar surface and data transmission between any two points of the lunar surface can be realized; and the task requirements of manned lunar-landing activities can be well met.
Owner:NAT INNOVATION INST OF DEFENSE TECH PLA ACAD OF MILITARY SCI

Light control method, movable platform and light control system

PendingCN112106448AAccurate Lighting ControlReal-time controlElectrical apparatusEngineeringComputer science
The embodiment of the invention provides a light control method, a movable platform and a light control system, and the light control method comprises the steps: obtaining a movement track of the movable platform; obtaining at least one waypoint of the movement track; obtaining light parameter information associated with any waypoint in the at least one waypoint; determining that the movable platform moves to any waypoint in the at least one waypoint; and performing light control on at least one light device of the movable platform according to the light parameter information associated with any waypoint. According to the embodiment of the invention, accurate light control can be carried out on the movable platform in the movement process.
Owner:SZ DJI TECH CO LTD

Position determination method and related equipment

The embodiment of the invention provides a position determination method and related equipment, and can calculate more accurate message delay, thereby obtaining a more accurate vehicle position. The method comprises the steps that a first moment when target information is read is determined, the target information comprises a second moment when a target object is detected and attribute informationof the target object, the target information is a message which is sent by a detection device and assembled according to a preset format, and the target object is an object of which the position is to be determined; determining message delay according to a target clock offset, the first moment and the second moment, wherein the target clock offset is the clock offset of the terminal equipment; and determining a target position of the target object at the first moment according to the message delay and the attribute information of the target object.
Owner:TENCENT TECH (SHENZHEN) CO LTD

A key point detection and recognition method for forward-looking radar imaging sea surface targets

The invention discloses a forward-looking radar imaging sea-surface target key point detection and recognition method. The method includes the steps that radar echo data are quantized into a gray scale image; a region of interest is extracted from the gray scale image, partition is performed, and a target region partition image is obtained; the radar echo data and the region of interest of the image are used, and a target region peak point information matrix is obtained; information merging is conducted on the target region partition image and the target region peak point information matrix, the number K of peak points in a target region in a merging result is obtained through counting, the previous K peak points are selected and serve as target effective peak points and form a target effective peak point image in a binaryzation mode, target axial features are extracted, and a target position is determined; the center of gravity of target energy is calculated and serves as the target key point. According to forward-looking radar target features, multiple mode identification methods are used comprehensively, the inherent features of a target can be reserved, meanwhile interference factors such as artifacts and side lobes can be restrained, and the recognition accuracy and the positioning precision of radar imaging sea surface target key points are improved.
Owner:HUAZHONG UNIV OF SCI & TECH

Visual tracking method and system for regional suggestion prediction of multi-view expert group

PendingCN113762256ASolve the problem of single discriminant feature representationImprove tracking performanceImage analysisCharacter and pattern recognitionPositive samplePredictive learning
The invention discloses a visual tracking method and system for regional suggestion prediction of a multi-view expert group. The method comprises the following steps: acquiring a to-be-tracked video; wherein a first frame of image in the video to be tracked is provided with a target area frame of the object to be tracked and an area frame which is twice as large as the target area of the object to be tracked; based on the region frame of the first frame image, training a classifier to obtain a trained classifier; based on a multi-expert model, performing feature extraction on the ith frame of image of the to-be-tracked video to obtain an optimal feature; wherein the value range of i is 2-N; n is a positive integer; inputting the optimal feature into the trained classifier, and outputting to obtain a rough positive sample candidate frame; and obtaining an optimal candidate frame based on the rough positive sample candidate frame. And a network prediction learning tracker and a multi-view multi-expert repair scheme are suggested by a bridging region, so that the tracking performance can be cooperatively improved.
Owner:SHANDONG INST OF BUSINESS & TECH

Liquid crystal panel delamination detection equipment and method

The invention provides a liquid crystal panel delamination detection device and method, the device comprising: a base, a mobile platform, a natural light source camera assembly, a first ultraviolet light source camera assembly, a second ultraviolet light source camera assembly, an upper light source assembly and a backlight Assemblies, the mobile carrier adopts an air-floating platform, and the mobile carrier is set above the base. The natural light source camera assembly, the upper light source assembly and the first ultraviolet light source camera assembly are all located on one side of the mobile carrier. The first ultraviolet light source camera assembly and the second The two ultraviolet light source camera components are symmetrically arranged with the liquid crystal panel as the axis, and the second ultraviolet light source camera component and the backlight source component are located on the other side of the liquid crystal panel; wherein, the upper light source component includes a first natural light source component and a first ultraviolet light source component, The backlight source assembly includes a second natural light source assembly and a second ultraviolet light source assembly. The invention can automatically distinguish the defects on the surface of the liquid crystal panel and the defects inside the liquid crystal panel, thereby improving the detection efficiency.
Owner:HEFEI SHANGJU IND EQUIP +2

Target position determining method, determining device and determining system

The invention relates to the field of intelligent driving, in particular to a target position determining method, a determining device and a determining system, and the method comprises the steps: obtaining the sight direction of a first user in a cabin; determining a target position pointed by the sight line direction of the first user according to the sight line direction of the first user and an estimation model; wherein the estimation model is used for representing the probability that the sight line direction of a user in a cabin points to a target position, the estimation model is obtained by training according to the corresponding sight line direction when a second user executes a driving operation and a preset target position corresponding to the driving operation, and the preset target position comprises the target position. According to the invention, the target position of the sight line of the user can be accurately determined in the cabin in the vehicle with complex light, and the problem that the target position of the user is difficult to accurately obtain due to complex light in the cabin in the vehicle is solved.
Owner:HUAWEI TECH CO LTD

Vehicle detector for microwave and video fusion detection and installation method thereof

The invention provides a vehicle detector for microwave and video fusion detection. The vehicle detector comprises a shell main body, wherein two sides of the shell main body are respectively providedwith a panel fixing frame and a rear cover, a microwave protection panel is arranged between the panel fixing frame and the shell main body, a camera, a microwave detection module, a power panel anda processing module are arranged in the shell main body, a round hole is formed in a microwave protection panel corresponding to the camera, and the power panel is electrically connected with the camera, the microwave detection module and the processing module, the processing module is in data connection with the camera and the microwave detection module respectively, a mounting bracket is arranged on the outer side of the shell main body, the detection area of the vehicle detector is large, accurate target position, speed information and field image information can be provided, and stopped targets can be detected.
Owner:舟山千眼传感技术有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Eureka Blog
Learn More
PatSnap group products