53 results about "Angular point" patented technology

The invention provides a three-dimensional calibration method using a speckle image as a calibration object and combining weight parameters. The method comprises the following steps: acquiring a standard grid image and the speckle image through a camera; determining angular point positions in the standard grid image through an angular point detection algorithm; calibrating the stereoscopic vision system by using the obtained angular points to obtain an initial calibration result; calculating 3D coordinates of the camera calibration board in the world by using the disparity map; determining a subset of the reference image corresponding to the checkerboard angular points in the target image through a digital image cross-correlation algorithm; calculating a weight parameter of each calibration point based on the re-projection error; solving internal and external parameters of the camera by using the weight parameters, radial alignment constraint and a least square method; optimizing the main point of the camera by using the weight parameters; and if the re-projection error or the number of iterations does not meet the requirement, repeatedly calculating the weight parameter of the calibration point and the subsequent steps, otherwise, obtaining a camera calibration result, and obtaining the distortion curved surface of the camera lens.

The invention relates to the technical field of object recognition based on image and point cloud fusion, and discloses a machine vision-based obstacle detection method and system in a container moving process, and the method comprises the steps: obtaining three-dimensional point cloud data through a laser radar installed on container carrying equipment, and obtaining a two-dimensional image through an industrial camera; performing terrain reference correction on the three-dimensional point cloud data, and establishing an inclined coordinate system matched with the ground gradient; performing perspective distortion correction on the two-dimensional image based on the inclined coordinate system to obtain a corrected two-dimensional image; extracting edge features, angular point features and semantic features from the corrected two-dimensional image; extracting geometric features and reflection intensity features of the three-dimensional point cloud data under the inclined coordinate system; performing cross-modal fusion on the edge features, the angular point features, the semantic features, the geometric features and the reflection intensity features to generate fusion features; and the obstacle type is identified based on the fusion feature, and the obstacle sensing reliability in the complex industrial environment is improved.

The invention relates to the technical field of SLAM, in particular to a dynamic environment robust positioning method based on improved ORB-SLAM2, and the method comprises the steps: obtaining RGB-D scene data containing an indoor moving object, and extracting ORB feature points described by FAST angular points and BRIEF descriptors; extracting feature points with unreliable heights according to geometric features of ORB feature points of adjacent frames; pixel analysis is carried out on the RGB-D scene data based on a Fast-SAM instance segmentation model, and a mask matrix containing environmental object space boundary information is generated; performing point-prompt extraction on the height unreliable feature points and the mask matrix to obtain a dynamic object mask; removing ORB feature points existing in the dynamic object mask to obtain a reliable feature point set; and calculating matching point pairs in the reliable feature point set to obtain an essential matrix and a corresponding rotation matrix and translation vector, and refining the pose based on a nonlinear optimization method.

The invention discloses an article grabbing and positioning method, system and device and a medium, and relates to the technical field of visual recognition. The method comprises the steps that a two-dimensional image, collected through a monocular camera, of an object to be grabbed is acquired; the types and the number of the to-be-grabbed articles in the two-dimensional image are recognized, and standard physical sizes corresponding to the types are determined; if the object is a single object, the object serves as a target to-be-grabbed object, the top face corner pixel coordinates of the object are detected, and the three-dimensional pose is solved through the PnP algorithm by combining the physical size and the internal reference of the monocular camera; if the number of the objects is multiple, all candidate corner points are detected, quadrilateral corner point sets of all the objects are screened out through clustering and geometric constraint analysis, then candidate three-dimensional poses of all candidate objects to be grabbed are calculated through a PnP algorithm, and then a target object to be grabbed is selected according to the candidate three-dimensional poses; and finally outputting the three-dimensional pose information of the target to-be-grabbed object. According to the invention, accurate positioning of a single article and three-dimensional pose recognition in a multi-article stacking scene are realized only by means of the monocular camera.

The invention provides a vSLAM and display screen coordinate system alignment method and system based on a two-dimensional code, and the method comprises the steps: controlling a display screen to synchronously display a two-dimensional code with a known physical size and a known position parameter under a display screen coordinate system, collecting an image frame containing the two-dimensional code, and obtaining a camera pose synchronized with the image frame in time; identifying the two-dimensional code, extracting the coordinate of the angular point image, and calculating the spatial pose of the two-dimensional code in a vSLAM world coordinate system in combination with the physical size of the two-dimensional code participated in the camera; establishing a coplanar consistency constraint based on a plurality of two-dimensional code space poses, jointly estimating a display screen world pose of the display screen in a vSLAM world coordinate system, and calculating a deformation parameter used for representing a non-ideal plane state of the display screen; and further constructing a coordinate system mapping relation between a display screen coordinate system and a vSLAM world coordinate system, realizing accurate mapping from a screen coordinate to a world coordinate, and performing verification and exception processing on an alignment result.

The application relates to an object positioning method, device, system, readable storage medium and program product. The method comprises the following steps: acquiring an object angle range and an object edge direction vector of a target object; extracting at least one candidate corner point and at least one candidate edge line segment combination from a shooting image corresponding to the target object; extracting at least one candidate geometric feature combination from all candidate corner points and all candidate edge line segment combinations according to the object angle range and the object edge direction vector; screening a target object shape from a candidate object shape corresponding to each candidate geometric feature combination, and positioning the target object according to the target object shape. The method can accurately position the object.

The invention is suitable for the technical field of lasers, and provides a distortion correction method and device for an ultra-wide-angle camera and electronic equipment, and the method comprises the steps: obtaining a to-be-corrected image corresponding to a to-be-engraved material collected by the ultra-wide-angle camera in a laser engraving machine; distortion correction parameters corresponding to the ultra-wide-angle camera are obtained, the distortion correction parameters comprise a distortion coordinate matrix and a standard coordinate matrix corresponding to each angular point in a reference image, and the reference image is a calibration plate image collected by the ultra-wide-angle camera; and correcting the to-be-corrected image according to the distortion coordinate matrix and the standard coordinate matrix to generate a corrected image corresponding to the to-be-engraved material. Therefore, each pixel point in the to-be-corrected image is corrected to the correct position according to the pre-stored distortion correction matrix and the standard coordinate matrix, and image distortion generated during shooting of the ultra-wide-angle camera is eliminated, so that the image quality of the engraving breadth is improved, and the engraving precision of the laser engraving machine is improved.

The invention discloses a camera calibration method, electronic equipment and a computer readable storage medium, and the method comprises the steps: obtaining a plurality of target poses corresponding to a calibration plate, and determining a target pixel coordinate corresponding to an angular point under each target pose based on the angular points on the calibration plate and the image size of a camera; acquiring a current image of the calibration plate under a current pose acquired by the camera, determining a current pixel coordinate corresponding to an angular point in the current image, adjusting the calibration plate from the current pose to a corresponding target pose based on the current pixel coordinate corresponding to the angular point and the target pixel coordinate, and acquiring a target image corresponding to the calibration plate by using the camera; calibrating the camera by using the target images under the plurality of target poses to obtain a plurality of groups of candidate camera internal parameters; and acquiring a verification image corresponding to the verification plate by using the camera, and obtaining a target camera internal reference from the multiple groups of candidate camera internal references based on a distortion removal effect of each group of candidate camera internal references on the verification image. According to the scheme, the convenience and precision of camera calibration can be improved.

Systems and methods are described relating to pose and shape estimation of objects. In some examples, a camera generates images depicting an object from different viewpoints. Within the images, a system identifies corners of the object, and the system uses these corners to generate lines that are projected from different viewpoints through the corners. Points may be identified by intersecting lines such that the points may be used to generate estimations of the object. The estimation with the highest score may be used to place the object in a location different from where the object was located.

The invention relates to a marker (400), in particular a two-dimensional marker, which is optically detectable and has a polygonal, in particular rectangular, marker base shape (402) with corners (404), wherein the marker has at one or more of the corners (404) a teardrop-shaped extension (406) located outside the marker base shape (402), which creates a saddle point at the respective corner. The invention also relates to an object with such a marker and to a method for determining the position of the saddle points.

The invention relates to an electronic equipment parameter calibration method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: in response to a triggered calibration starting operation, obtaining a corner image containing a target calibration pattern and distance information; performing angular point extraction on the angular point image to obtain image coordinates of each angular point in the angular point image; on the basis of the image coordinates of all the angular points, the internal parameters and distortion parameters of the rearview mirror lens and the world coordinates of all the angular points, the calibration external parameters of the rearview mirror lens are determined; wherein the world coordinate of each angular point is determined based on the distance information; and determining a calibration result of the calibration external parameter based on the precision value of the calibration external parameter. By adopting the method, the parameter calibration accuracy of the electronic equipment can be ensured, and the parameter calibration efficiency of the electronic equipment can be effectively improved.

The invention relates to a camera pose determination method and device, a medium, electronic equipment and a program product, and the method comprises the steps: obtaining an RGB image and a depth image which are shot by a camera and are used for describing an inverse color calibration board, and the inverse color calibration board is obtained through the inversion of the color of the calibration board; according to the target image, the depth image and the camera internal reference, determining a first three-dimensional coordinate of a target feature point in the target image in a camera coordinate system, the target image being an image obtained by performing color inversion processing on the RGB image, and the target feature point being an angular point of a black square block or a central point of the black square block in the target image; performing point cloud matching on the first three-dimensional coordinate and a second three-dimensional coordinate of the target feature point in a calibration plate coordinate system to obtain an initial camera pose; according to the camera internal reference, the second three-dimensional coordinate, the pixel coordinate of the target feature point and the initial camera pose, the target camera pose is determined, and the estimation precision of the target camera pose determined by depending on the initial camera pose is improved.

This information processing device (100) comprises: an acquisition unit (120) that acquires a panoramic image including an object that is a cube or a rectangular parallelepiped; a detection unit (130) that detects the object from the panoramic image; and a determination unit (140) that, when four corners of the object are detected, determines whether a circle passing through four points corresponding to the four corners can be calculated, and when the circle passing through the four points can be calculated, determines that the accuracy of the four corners is high.

The invention provides an instance segmentation method and device in the field of computer vision, which are used for aligning primitives segmented in a three-dimensional model based on a two-dimensional instance in an open scene so as to output three-dimensional instance segmentation with consistency, and the same three-dimensional instance can be segmented even if a visual angle is switched. The method comprises the following steps: acquiring at least one image of a target scene, a camera view angle of the at least one image and a three-dimensional model of the target scene; instance segmentation is carried out on the at least one image to obtain a first two-dimensional instance segmentation result, and the first two-dimensional instance segmentation result comprises at least one two-dimensional instance; obtaining at least one primitive in the three-dimensional model, wherein the at least one primitive comprises primitives formed by angular points and surface patches in the three-dimensional model; then projecting at least one two-dimensional instance and at least one primitive based on a camera visual angle, and determining at least one three-dimensional instance that each three-dimensional instance comprises one or more primitives; and outputting the instance identifier of each three-dimensional instance.

The invention provides an unstacking and stacking method, device and equipment, and relates to the field of visual algorithm recognition. The method comprises the following steps: acquiring actual three-dimensional point cloud data of a target material; intercepting target point cloud data in the actual three-dimensional point cloud data according to the attribute information of the target material; converting the target point cloud data into a two-dimensional image; wherein the attribute information is an actual accumulation parameter. According to the two-dimensional image, determining a target angular point of the target material, and determining an angular point position and a rotation angle of the target angular point; and according to the angular point position and the rotating angle, a grabbing posture is determined, and the grabbing posture is sent to a preset mechanical arm controller, so that the mechanical arm controller executes stacking and unstacking operation on the materials according to the grabbing posture. The method is used for achieving the effect of improving the stacking and unstacking positioning precision.

This application provides a five-axis measurement and evaluation method for wheel-type tenon grooves, relating to the field of wheel-type tenon groove measurement technology. After the wheel is clamped, a wheel coordinate system is established. Based on the nominal position and orientation of the tenon groove, a nominal coordinate system for the tenon groove is established, and correction measurement points are collected on the working surface to be measured. The initial coordinate system of the tenon groove is obtained by correcting it according to the least squares criterion. A high-density point cloud is obtained by five-axis oscillation scanning and fitted to the actual working surface. The nominal corner points are projected onto the actual working surface to obtain the measured corner points. The coordinate system is corrected using the least squares method for corner point deviation. The oscillation-fitting-correction process is iterated until the change in coordinate system between two adjacent iterations is less than 10% of the tolerance bandwidth, thus obtaining the measured coordinate system of the tenon groove. Geometric tolerances such as center point position, axial / radial tilt, and angle are calculated in the wheel coordinate system to complete the machining quality evaluation. This method reduces the accumulated error of the reference structure and improves the accessibility and measurement repeatability of oblique tenons.

In implementations of techniques and systems for chamfer tools in volumetric sculpting software, a processing device receives an indication of a brushed region of a three-dimensional (3D) object to be chamfered and a chamfer size. A machine-learning model determines surface voxels in the brushed region representing an edge or corner voxel. The machine-learning model also determines parameters for the edge or corner voxel that define a chamfer plane and a chamfer coefficient. The chamfer plane is then moved in a direction normal to the chamfer plane by a distance based on the chamfer size and the chamfer coefficient. To generate a chamfer, the processing device performs an addition or subtraction operation on the 3D object to the moved chamfer plane for the edge or corner voxel.

The invention provides a corner position identification-based processing starting point determination method, device and chip, and the method comprises the following steps: determining a corner position identification corresponding to each corner point of a target primitive according to the relative position relation between each corner point of the target primitive and all rectangular edges of a target processing frame; the angular point position identifiers comprise flag bits in one-to-one correspondence with the rectangular edges; and sequentially processing each triangular edge according to the processing sequence, and if the angular point position identifiers of the two end points of the current triangular edge have the condition that the value of the flag bit corresponding to the same rectangular edge is a second preset value, ignoring the current triangular edge. The number of times of calculating whether the triangular edge and the rectangular edge have the intersection point can be reduced, so that the calculation power consumption is reduced.

The present disclosure relates to a depth calculation method and device, a computer device and a medium, comprising: calculating an initial relative displacement between a main image and a sub-image in a to-be-processed image; moving the sub-image according to the initial relative displacement, so that the relative displacement between the main image and the moved sub-image is less than the initial relative displacement; determining a main camera corner point in the main image, and determining a corresponding sub-camera corner point of each main camera corner point in the sub-image to obtain a corner point pair set; determining a mis-matched corner point pair in the corner point pair set, and deleting the mis-matched corner point pair from the main image and the sub-image to obtain an effective main image and an effective sub-image; moving the effective sub-image back to the initial position in reverse, and performing affine transformation on the sub-camera corner points in the reversed and moved effective sub-image to the effective main image; and determining the depth of a target object according to the sub-camera corner points in the effective main image after affine transformation and the main camera corner points in the effective main image. The accuracy of determining the depth of the target object can be improved.

Provided in the present application are an instance segmentation method and apparatus in the field of computer vision, which are used for aligning, on the basis of a two-dimensional instance in an open scene, primitives segmented from a three-dimensional model, and then outputting consistent three-dimensional instance segmentation, such that even if the view is switched, the same three-dimensional instance can be segmented. The method comprises: acquiring at least one image of a target scene, a camera view of the at least one image, and a three-dimensional model of the target scene; then performing instance segmentation on the at least one image, so as to obtain a first two-dimensional instance segmentation result, wherein the first two-dimensional instance segmentation result comprises at least one two-dimensional instance; acquiring at least one primitive in the three-dimensional model, wherein the at least one primitive comprises a primitive formed by a corner point and a surface patch in the three-dimensional model; next, on the basis of the camera view, projecting the at least one two-dimensional instance and the at least one primitive, so as to determine at least one three-dimensional instance, wherein each three-dimensional instance comprises one or more primitives; and outputting an instance identifier of each three-dimensional instance.

The invention discloses a manipulator visual grabbing method and device based on pixel coordinate system direct mapping, and the method comprises the steps: firstly setting a black and white comparison checkerboard angular point image at the wrist part of a robot arm, obtaining the pixel coordinates of the tail end of a dexterous hand through the pixel coordinates of the wrist part, and calibrating the coordinates of the working space of the robot arm through a depth camera; extracting pixel coordinates of the target object; then, according to the current pixel coordinates of the tail end of the dexterous hand, the target displacement amount is calculated, and then the robotic arm is driven to conduct corresponding displacement; according to the pixel error amount of the displacement, the joint speed is updated; and the mechanical arm is continuously driven to move until the pixel error meets the requirement or the number of times of displacement tasks reaches the maximum value. According to the method, the trajectory is directly planned in the pixel space, multi-coordinate system conversion calculation is omitted, only two-dimensional pixel deviation needs to be processed, and the system operation load is reduced. Angular point mark design and a self-adaptive scale factor and force feedback mechanism support detection at any visual angle, and mechanical arm parameter errors and environment disturbance are effectively compensated.

Approaches presented herein provide for the automatic inspection of material element having an expected size, shape, and location. Such automatic analysis can be useful for inspecting the installation of an element, such as a patch of thermal interface material (TIM) attached to a heat sink of a graphics card. A digital image can be captured and cropped to a region of interest including an element to be inspected. A contour of the element can be identified and used to determine the location of a first edge of the element. Test lines can be swept across the area of the contour until at least one edge criterion is satisfied for additional edges of the element. The intersections of these edges can be identified and used as approximations of the corners or vertexes of the element. The coordinates of these elements, or values calculated therefrom, can be compared to expected coordinates from a reference standard to determine whether the element satisfies one or more inspection criteria.

The invention provides a dual-camera weak texture multi-target space matching and positioning method and an application system, and belongs to the technical field of computer vision and three-dimensional perception. According to the method, based on a dual-camera epipolar geometric model, on the basis of obtaining a dual-camera synchronous image and performing multi-target detection, four corner points and a central point of a target outer surrounding frame are introduced to construct a multi-key-point combined epipolar constraint, and target space structure information is fully utilized. The multi-key-point combined epipolar constraint is constructed by selecting the four corner points and the center point of the outer surrounding frame of the target, the overall space structure information of the target is introduced into the dual-camera matching process, a matching constraint mechanism of multi-geometric clue fusion is formed, and compared with an epipolar constraint mode only based on a single image point, the matching constraint mechanism is more accurate. The matching ambiguity caused by appearance similarity or deformation is effectively reduced, so that the matching accuracy in a multi-target scene and the anti-interference capability of the system are improved, and the method is particularly suitable for a weak texture environment.

The invention provides a method and a device for determining a processing starting point of a target primitive and a chip, and the method comprises the following steps: determining an angular point position identifier corresponding to each angular point of the target primitive according to a relative position relation between each angular point of the target primitive and all rectangular edges of a target processing frame; determining a processing sequence of each triangular edge according to a preset rule; obtaining a current processing sequence, and performing first processing on the triangular edge corresponding to the current processing sequence; the first processing comprises the steps of determining at least one target rectangular edge according to a target rectangular edge determination method corresponding to the target angular point position identifier; and if at least one intersection point exists between the target triangular edge and one or two target rectangular edges, determining one of the at least one intersection point as a processing starting point, and ending the processing starting point determination method of the target primitive. According to the method, the intersection points of each triangular edge of the triangular primitive and all rectangular edges of the target processing frame do not need to be calculated, so that the required calculation power is reduced.

The invention provides a four-corner-point-based visualization processing method and system for Web end SAR view field projection, and relates to the field of satellite task visualization, and the method comprises the steps: determining a starting angle and a crossing angle of a fan-shaped angular domain; performing equal-interval discretization on azimuth angles, performing spherical surface positive solution on each discrete azimuth angle, and generating a polygonal entity; the spatial position of a platform where the SAR load is located is used as a vertex, and each vertex of the earth surface projection polygon is used as a bottom edge vertex to construct a beam cone. According to the method, the spherical fan-shaped construction and the earth surface projection generation of the SAR view field are realized at the browser end in a data driving mode that the central point is matched with the four angular points, and the view field boundary is accurately restored with fewer input parameters; the continuity and the closeness of projection topology are ensured through azimuth angle sorting and angular domain normalization; through the construction of the wave beam cone, the direction and the opening angle of the field of view are visually presented, and the visual and interactive display of the SAR field of view space range and the orientation direction is realized.

The invention discloses a 3D target detection method and system. The method comprises the following steps: S1, obtaining a two-dimensional image of a target scene by using a monocular camera; s2, performing target detection on the two-dimensional image through a lightweight single-stage detection network to obtain a first two-dimensional detection frame; s3, calculating a perspective coefficient based on camera calibration parameters and known sizes in a scene; s4, performing proportional transformation on the first two-dimensional detection frame according to the perspective coefficient to generate a second two-dimensional detection frame; s5, projecting corresponding vertexes of the first two-dimensional detection frame and the second two-dimensional detection frame to a camera coordinate system, and constructing a three-dimensional target frame containing eight space angular points; and S6, transmitting the three-dimensional target frame information to a rear-end control unit through a high-speed SRIO interface. The invention belongs to the technical field of computer vision and target detection, and particularly relates to a 3D target detection method and system.

The invention relates to the technical field of visual positioning, in particular to a visual positioning method and system for carton stacking, and the method comprises the steps: uniformly dividing a carton image into a plurality of windows, calculating entropy values of angle histograms in the windows, and taking the windows with the entropy values smaller than a first threshold value as candidate windows; calculating an angular point existence index for the angle histogram of the candidate window; the candidate windows with the corner existence indexes larger than a second threshold value serve as corner windows; for any pixel point in the corner window, constructing a structure tensor matrix by a sub-window taking the pixel point as a center, and calculating a harmonic average of a plurality of feature values; and clustering harmonic averages of all pixel points in each corner window to obtain a plurality of clusters, and executing different carton positioning strategies according to different clusters. According to the technical scheme, the carton positioning efficiency in the carton stacking process can be improved.

The invention relates to a flaw recognition system for a special-shaped sealing bolt forming process based on image processing, and belongs to the technical field of image recognition. The system considers that a complex structure on the surface of the special-shaped sealing bolt enables the gray gradient magnitude difference of different angular points to be large, abandons a traditional angular point judgment method for comparing the gradient magnitude with a threshold value, and determines the angular point degree according to the multi-dimensional features of a target pixel point and a neighborhood pixel point about the distance, the gradient vector difference and the gradient magnitude of the target pixel point and the neighborhood pixel point. All the pixel points are clustered in combination with the gray values and the angular point degrees of the pixel points, and the error degree caused by local too dark influence of the collected image when the angular point degrees of the pixel points in the class clusters are determined is determined according to the gray values, the sizes and the angular point degree mean values of all the classes of clusters; and on the basis of the error degree, combining the outlier degree of the target pixel point in the class cluster to determine an angular point degree correction coefficient, and completing angular point judgment. According to the method, the angular point judgment accuracy of the special-shaped sealing bolt can be remarkably improved, so that the defect recognition accuracy is improved.

The application provides a fixed viewpoint hand-eye calibration method and system based on a two-dimensional code, according to obtained two-dimensional code image data of a mechanical arm end, an edge line of a two-dimensional code positioning area is obtained by using a Canny edge detection operator, and a corner point position of the two-dimensional code is obtained through coordinate transformation; according to obtained two-dimensional code image data of the mechanical arm end, a corner point position of the two-dimensional code is obtained by solving through EPnP; whether errors of four corner points of the two-dimensional code are all less than a first set threshold value is judged, if yes, calibration of positive and inverse matrix parameters of the mechanical arm is carried out according to the corner point position of the two-dimensional code obtained through coordinate transformation; the application combines the recognition result of the two-dimensional code at the mechanical arm end, ensures positioning accuracy of the mechanical arm, is simple to operate and has good timeliness.

The invention relates to the technical field of mode recognition, and discloses a plane measurement method and device, computer equipment and a storage medium, and the plane measurement method comprises the steps: obtaining color image data and depth image data of an environment; determining a target plane based on the color image data; determining a target three-dimensional point cloud corresponding to the target plane based on the depth image data; according to the target plane and the target three-dimensional point cloud, a target normal vector and a target angular point of the target plane are determined, and the target normal vector and the target angular point are used for measuring the target plane. Through the technical scheme of the invention, the problem of low accuracy of plane measurement is solved, and the accuracy of plane measurement is improved.