52 results about "Visual localization" patented technology

An apparatus and method for visual localization of a visual camera system outputting real-time visual camera data and a graphics processing unit receiving the real-time visual camera data. The graphics processing unit accesses a database of prior map information and generates a synthetic image that is then compared to the real-time visual camera data to determine corrected position data. The graphics processing unit determines a camera position based on the corrected position data. A corrective system for applying navigation of the vehicle based on the determined camera position can be used in some embodiments.

There is provided a visual localization method comprising: (a) transmitting data representative of one or more detected visual features from a mobile device to a server; (b) estimating the location of the mobile device at the server based on the visual features 5 received from the mobile device; (c) transmitting reference data associated with the estimated location from the server to the mobile device; and (d) the mobile device determining its location based on the reference data received from the server.

The invention relates to the technical field of electronic security and protection, particularly to a tag matching method, an electronic fence boundary management method and device and a storage medium. A single base station is adopted to be combined with a monocular camera or an integrated structure formed by the two is adopted, through fusion of a vision localization technology and wireless distance measurement technology, accurate obtaining of information of a monitored target including location information, identity information and authority information of entering an electronic fence is realized, a circumstance that a target enters an electronic fence boundary or area without carrying a tag card and a circumstance that a target carries multiple cards can also be detected, and thus unattended monitoring and management of personnel and equipment of large industries, building yards, mining regions, prisons and the like. The system has comprehensive functions, is simple in deployment,is convenient and easy to operate, and realizes a better management effect with less resources, and cost required by the system is also greatly reduced.

The invention relates to a vision localization, measurement and control method for an automatic refueling dead zone of an unmanned aerial vehicle. According to the method, pattern recognition, calculation and localization of a fuel delivery taper sleeve on an unmanned aerial vehicle for fuel delivery are carried out by a visual machine inspection method, so that the relative position between a filler pipe and the fuel delivery taper sleeve is obtained; locating data are transmitted to a ground station; the ground station filters the data, and calculates a control command by a control algorithm; the control command is sent to an aircraft end for controlling the aircraft to achieve the purpose of automatic refueling. The invention also provides a relevant localization, measurement and control system and an experimental platform. The invention has the core idea that a refueled unmanned aerial vehicle collects a target image of the fuel delivery taper sleeve through a camera with an adjustable focus, calculates a relative position and size of a target in view by a projection method and a pattern recognition algorithm, then calculates the relative position between a refueled pipe and the fuel delivery taper sleeve, and the ground station controls unmanned aerial vehicles to dock and implement refueling operation. The vision localization, measurement and control method for a dead zone is fast and stable, high in precision, and can meet the application requirement for localization of an automatic refueling dead zone of an unmanned aerial vehicle.

The invention puts forward a virtual studio mobile robot visual localization method based on improved Apriltag tags, comprising the following steps: adding four uniformly-sized ring markers at the appropriate length of quadrilateral diagonal extension cords of a tag; calibrating a camera; acquiring tag images: preprocessing the acquired images; carrying out threshold processing on ROI region extraction; identifying elliptical markers; extracting the ROI region; carrying out threshold processing on the minimum rectangular area surrounding the four points which is the ROI tag area of the tag images with the coordinates of four ellipse centers on the improved tags as the benchmark; extracting tag corner points and fitting the extracted straight lines by the weighted least square method, calculating intersection points of straight lines to obtain subpixel-level corner point coordinate, and extracting image coordinates of the tag corner points; reading the coded message of the tags according to the extracted image coordinates of the corner points, obtaining world coordinates of the corner points, and obtaining the pose of a robot by coordinate transformation. The method of the inventionhelps reduce the data size of the identification tags and improve the real-time performance of identification.

The invention discloses an RANSAC algorithm-based visual localization method which belongs to the field of visual localization. The traditional RANSAC algorithm has more iteration times, large calculation amount and long computation time, so that the visual localization method implemented by this algorithm has the problem of low localization speed. The RANSAC algorithm-based visual localization method comprises the following steps: calculating feature points of images uploaded by a user to be localized by an SURF algorithm and feature point description information; selecting one picture with the most matching points from a database, performing SURF matching on the obtained feature point description information of the images and the feature point description information of the pictures, defining each pair of images and pictures for matching as one pair of matching images, and obtaining a group of matching points after matching each pair of matching images; eliminating mistaken matching points in the matching points of each pair of matching images by the RANSAC algorithm of matching quality, and determining four pairs of matching images with the most correct matching points; calculating a position coordinate of the user by an epipolar geometric algorithm based on the obtained four pairs of matching images, so as to complete the indoor localization.

Provided is a method of processing the flight path of an unmanned aerial vehicle. An unmanned aerial vehicle is fixedly equipped with a binocular camera. The binocular camera consists of a left camera and a right camera. The method comprises the following steps: acquiring a current frame image of the binocular camera; detecting whether the current frame image is similar to the key frame image of a recorded path; if the current frame image is similar to the key frame image of the recorded path, calculating the relative pose of the binocular camera according to the current frame image and the key frame image, calculating the relative pose of the unmanned aerial vehicle according to the relative pose of the binocular camera, and correcting the relative pose of the unmanned aerial vehicle; if the current frame image is not similar to the key frame image of the recorded path, judging whether the current frame image is a key frame image; and if the current frame image is a key frame image, and storing the current frame image as a key frame image. As the flight path of the unmanned aerial vehicle is stored according to the key frame image, the amount of computation of a visual localization algorithm is reduced greatly. The recorded path enables the unmanned aerial vehicle or other unmanned aerial vehicles equipped with a camera to fly by path. Moreover, by correcting the position drift of the unmanned aerial vehicle in the process of flight by path, a precise tracking effect can be achieved.

The invention discloses a robot moving method and device. The method comprises the following steps of performing sound localization through a sound collecting device, and determining the direction ofa target; enabling a camera to face the target, and detecting the target through the camera; when the target is detected, performing visual localization through the camera, and determining the position where the target is located; moving to the position where the target is located. According to the robot moving method, the target direction is determined through sound localization so as to enable the camera to face the target, then, the camera is used for performing visual localization to determine the accurate position of the target, the robot can be rapidly and accurately moved to the front of the user (the target), the technical problem that the robot cannot be accurately moved to the front of the user is solved, the purpose of accurately adjusting the orientation relation and the distance relation between the robot and the user is achieved, it is ensured that the robot can better receive a voice instruction of the user, and the user experience is greatly improved.

The invention discloses a joint positioning method based on a RGB-D and a laser speedometer. In the present invention, visual localization is based on the RGB-D, motion estimation is carried out according to line feature matching, and an EKF-SLAM is used in a laser method to carry out motion estimation, when the two algorithms are successfully positioned at the same moment, the system has two poses, and optimal estimation is obtained by carrying out combined filtering on the two poses; and if positioning failure is caused by certain scenes such as low textures and the like, the pose obtained by laser radar positioning is taken as the optimal pose of the global filter. Therefore, a combined positioning mode is adopted, advantages and disadvantages of a radar system and a depth camera systemare integrated, and visual SLAM based on line features and the EKF-SLAM are combined, so that the positioning and mapping precision of the system is greatly improved.

The invention discloses a CT scanning visual localization method, which comprises the following steps: (1) setting a scanned site, scanning parameters and human body parameters of a patient; (2) acquiring one or more localization reference human body characteristics related to the scanned site; (3) in a process of moving a scanning bed, continuously acquiring human body images passing a scanning region and implementing human body characteristic recognition in real time, and acquiring positions, corresponding the localization reference human body characteristics, of the scanning bed; and (4) inaccordance with the acquired scanning bed positions, in combination with anthropometry knowledge, calculating an initial scanning position and an ending scanning position, corresponding to the scanned site, of the scanning bed. According to the CT scanning visual localization method, automatic localization can be implemented in a moving process of the scanning bed in accordance with the selectedscanned site, so that a relatively high processing speed and accurate localization are guaranteed, and the method is friendly to the patient. Meanwhile, the invention also discloses a CT system for implementing the CT scanning visual localization method.

The invention provides a localization method and system based on stroke sensing and visual fusion, a terminal and a storage medium. In a preparation stage, visual data and stroke data are acquired; incremental corresponding relationship conversion is conducted, wherein an initial position is determined, and a conversion relationship for an extracted increment in the stroke data is obtained by stroke sensing equipment between a later moment j and a previous moment i; constraint is conducted, wherein visual sensing equipment is constrained with the conversion relationship for the extracted increment of the stroke sensing equipment as a constraint condition. The stroke sensing equipment is adopted for sensing the length and heading angle information of the movement of a vehicle body, the length or trajectory information sensed by the stroke sensing equipment serves as the trajectory constraint of a camera, the problem of scale uncertainty of the monocular camera is solved, and visual localization error accumulation and scale drift can also be prevented. In addition, the good localization effect can be achieved on non-planar scenes instead of being limited to planar scenes simply by adopting the length constraint of stroke sensing.

The present invention relates to a visual positioning system and method for industrial robots, wherein the system includes a plurality of industrial robots, a plurality of unmanned aerial vehicles corresponding to the industrial robots and a host computer, and each of the industrial robots passes through the The drone communicates with the host computer to exchange position data and request data. The industrial robot is equipped with a first battery and a first power management module, and the drone is equipped with a second battery and a second power management module. . The present invention can obtain position information through binocular recognition of industrial robots, and obtain position data through GPS positioning through drones, and correct and transmit data to the positions of landmark buildings in the map stored in the host computer. Multiple industrial robots It can cooperate with each other, and is especially suitable for the unified management of large-scale group robots. The upper computer can realize map update and path planning. The structure is simple, the application is convenient, and it is suitable for large-scale promotion and application.

An apparatus and method for visual localization of a visual camera system outputting real-time visual camera data and a graphics processing unit receiving the real-time visual camera data. The graphics processing unit accesses a database of prior map information and generates a synthetic image that is then compared to the real-time visual camera data to determine corrected position data. The graphics processing unit determines a camera position based on the corrected position data. A corrective system for applying navigation of the vehicle based on the determined camera position can be used in some embodiments.

The invention discloses random fluctuation vision localization research in multi-medium propagation. In engineering application, light is prone to deviation in multi-medium propagation, media are mixed, and random fluctuation causes inaccurate vision localization under multimedia. The binocular camera vision localization technology under the multimedia is explored in random fluctuation vision localization research. The random fluctuation vision localization research includes the following steps 1, the distortion coefficient in an image coordinate system is determined; 2, the image coordinate system is converted into a camera coordinate system; 3, relative poses among cameras in a word coordinate system are determined; 4, the camera coordinate system is converted into the work coordinate system; 5, vision localization under medium conversion is carried out; 6, vision localization randomness is described based on a cloud model.

A method for locating a vehicle, including at least one vision sensor and at least one item of equipment from among an inertial navigation unit, a satellite navigation module and an odometry sensor. The method including carrying out vision-localization from image data supplied by the at least one vision sensor, to produce first location data, and applying a Bayesian filtering via a Kalman filter, taking into account the first location data, to the data derived from the at least one item of equipment and the data from a scene model, to produce second data for locating the vehicle.

A visual localization support system is provided. The visual localization support system includes one or more guidance indicators place on a road surface of a roadway, wherein the one or more guidance indicators each include a matrix barcode that uniquely identifies a location by latitude, longitude, and altitude, and describes an affine shape of the guidance indicator.

The invention discloses a vision localization method of laser marking software. The method comprises the following steps that S1, a product is placed within the work range of a marking table freely, acamera takes a picture to obtain a product picture, and the software loads template parameters; S2, according to the template parameters, the product picture is preprocessed to obtain a preprocessedpicture; S3, traversal of the picture is carried out, and through the template parameters, area coordinates the same with a template are found out and that is a product center; and S4, a template picture is used again, an area is subjected to characteristic point and gray matching and outline fitting, an angle is obtained, and visual treatment is finished. Different clamps do not need to be manufactured according to products, red light preview is not needed, product image template parameters are edited in advance, subsequent any placement is achieved, a visual system automatically finds a target position and direction, deviation is corrected automatically, and precise marking is achieved.

An apparatus receives a first image captured by an image capture device of a mobile apparatus. The first image corresponds to surroundings of the mobile apparatus. Each artificial image of a first plurality of artificial images respectively is a two-dimensional projection of a three-dimensional model from a perspective of an image position and image pose. The apparatus determines one or more first image attributes respectively for one or more sections of the first image; identifies at least one artificial image of the first plurality of artificial images that has one or more artificial image attributes that substantially match the one or more first image attributes for corresponding sections of the at least one artificial image and the first image; and determines a location and / or pose of the mobile apparatus based at least in part on the image position and / or the image pose associated with the at least one artificial image.

The present invention discloses a visual localization method based on a color block and a topological relationship thereof. An artificial road mark image consisting of color blocks is acquired by means of a visual sensor and identified, and corresponding location coordinates are acquired using a correspondence between a preset image and position. A given artificial road mark has uniqueness; the attribute is easy to be extracted; a constructed artificial road mark topological mode has certain robustness for rotation and twisting and the like of images; and a matching method of the artificial road mark topological mode is rapid and effective and is capable of meeting localization requirements.

The invention discloses a hot-stamping process for a tank body of a metal circular tank. The hot-stamping process specifically comprises the following steps: loading the tank, performing surface treatment, performing visual localization, performing base coat silk-printing, performing UV roasting and curing, performing hot-stamping, and automatically discharging from the tank. According to the hot-stamping process for the tank body of the metal circular tank provided by the invention, hot-stamping processing is performed on the tank body of the metal circular tank on the basis of a conventionalhot-stamping process, so that hot-stamping processing can be performed on the tank body, and therefore, the aesthetic feeling of the tank body is strengthened, and tank body recognition degree is improved.

A system and method for visual anomaly localization in a test image includes generating, in plurality of scaled iterations, attention maps for a test image using a trained classifier network, using image-level. A current attention map is generated using an inversion of the classifier network on a condition that a forward pass of the test image in the classifier network detects a first class. One or more attention regions of the current attention map may be extracted and resized as a sub-image. For each scaled iteration, extraction of one or more regions of a current attention map is performedon a condition that the current attention map is significantly different than the preceding attention map. Visual localization of a region for the class in the test image is based on one or more of the attention maps

The invention discloses a visual localization guidance system and method for sorting, and relates to the field of warehouse logistics. The visual localization guidance system for sorting comprises a package transfer system, a smart camera vision system, a guidance bin localization guidance system and a controller, wherein the package transfer system comprises a conveyor line for transferring packages and a speedometer for measuring the speed of the conveyor line, the smart camera vision system comprises one or more smart cameras arranged above the conveyor line and is used for scanning and localizing the packages on the conveyor line and transmitting information of the packages to the controller, the guidance bin localization guidance system comprises a plurality of visual guidance bins arranged above the conveyor line at a certain distance from the downstream of the smart cameras and used for projecting color light onto target packages in a guidance area according to an instruction ofthe controller so as to provide visual localization guidance, and the controller is used for assigning guidance tasks to the visual guidance bins based on the information of the packages so as to instruct that visual localization guidance is conducted on the target packages. The visual localization guidance system and method for sorting reduce the error rate of manual sorting and improve the sorting efficiency.

Embodiments herein provide a method for stereo-visual localization of an object by a stereo-visual localization apparatus. The method includes generating, by a stereo-visual localization apparatus, a stereo-visual interface displaying the first stereo image of the object and the first stereo image of the subject in a first portion and the second stereo image of the object and the second stereo image of the subject in a second portion. Further, the method includes detecting, by the stereo-visual localization apparatus, a movement of the subject to align the subject in the field of view with the object. Furthermore, the method includes visually aligning, by the stereo-visual localization apparatus, the subject with the object based on the movement by simultaneously changing apparent position of the first and the second stereo images of the subject in each of the first portion and the second portion in the stereo-visual interface.

The invention provides a MEMS automatic wire bonding apparatus based on zoom microscope positioning, comprising a base. The base is equipped with an X-axis workbench, a Y-axis workbench and a microscope, a bonding head is arranged above the workbench, a liftable support is arranged on the base, the microscope is mounted on the liftable support and provided with a Z-axis stepping motor, and an automatic magnification regulating mechanism is arranged on the microscope. The introduction of the continuous zooming microscope solves the vision positioning problem of the large visual field and high precision microscope, and the use of the image processing algorithm realizes the whole bonding operation automation and positively promotes the realization of batched MEMS wire bonding operation.

The invention discloses a positioning method and device, electronic equipment and a storage medium. The method comprises the following steps: determining a first position of target equipment; determining a second mapping position set from the first mapping position set according to the first position; wherein the first mapping position set comprises mapping positions corresponding to a plurality of mapping areas, and each mapping position in the first mapping position set has a corresponding relationship with one mapping area; based on the second mapping position set, determining a mapping area list; each mapping area in the mapping area list has a corresponding relationship with a high-precision map; and based on the high-precision map corresponding to at least one mapping area in the mapping area list, performing visual localization on the image acquired by the target device.

The invention discloses an Apriltag visual localization processing method and a system thereof, a wall-climbing robot and a storage medium. The method comprises the steps of collecting a label image in a field of view of an industrial camera, performing equalization processing on the label image, denoising the label image after equalization processing, determining the pose of the wall-climbing robot based on an Apriltag positioning algorithm according to the denoised label image, and planning a moving path of the robot according to the pose, when the robot moves along the planned path, machining the to-be-machined workpiece in the normal direction of the curved surface, receiving working state parameters of the robot in real time, and sending out a working instruction. The Apriltag visual positioning is applied to the specific curved surface mobile adsorption processing robot, so that real-time autonomous positioning and accurate positioning of the mobile processing wall-climbing robot on the target curved surface are realized, meanwhile, the cost is saved, the processing performance is improved, and the influence of the environment on positioning is reduced.

In an embodiment of the invention there is provided a method of visual localization, comprising: generating a plurality of virtual views, wherein each of the virtual views is associated with a location; obtaining a query image; determining the location where the query image was obtained on the basis of a comparison of the query image with said virtual views.

The invention relates to a welding machine based on visual localization. The welding machine comprises a base, a movement mechanism mounted on the base, a welding gun module connected with the movement mechanism and a visual localization module, wherein the visual localization module comprises an image acquisition module which is fixedly connected with the welding gun module and is used for acquiring image information of a workpiece to be welded, an illumination module which is matched with the image acquisition module and is used for illuminating the workpiece to be welded, a reference objectwhich has a determined coordinate value, is arranged between the workpiece to be welded and the image acquisition module and is located in a photographing range of the image acquisition module, and an image processing device which is connected with the image acquisition module and is configured to identify and acquire a characteristic pattern in an image, compare the characteristic pattern with the coordinate value of the reference object, calculate an effective coordinate value of the characteristic pattern and transmit the effective coordinate value to the movement mechanism. The inventionfurther relates to an automatic welding method by utilizing the welding machine based on the visual localization, and automatic localization and welding can be realized.