31 results about "Chamfer distance" patented technology

The invention discloses a recognition and counting method for cells, comprising nine steps: preprocessing an image in a micron-order microscopic acquisition environment; extracting cell holes from the preprocessed image; performing closed hole filling of cells by using the knowledge of a connected domain; extracting a contour point sequence of cells from the image filled in step 3; filling non-closed holes of cells by adopting a non-closed hole filling method based on circularity determination; performing chamfer distance transformation on the filled image; performing extreme value uniqueness marking on cell hole positions; segmenting the image after extreme value uniqueness by using a marked watershed method; and quantifying and marking the segmented result. The method has the advantages that the influence of image noise can be greatly reduced, the phenomena of over-segmentation and discontinuous segment lines are eliminated, the segmentation effect is improved, and the cell recognition rate is improved.

This method makes it possible to plot, from a terrain elevation database, a map of the distances of the points accessible to a moving object subject to constraints (inaccessible reliefs, unnegotiable obstacles, weather disturbances, path with an imposed vertical profile, etc.), the distances being measured only along paths that are practical for the moving object. It employs a chamfer distance transform applied to the image consisting of the projection on the horizontal plane of a 3D representation of the flying space of the moving object, which is likened to a mesh of elementary cubes associated with specific negotiation danger levels. It lists the typical paths without exceeding an acceptable danger threshold, going from a target point, the distance of which is to be estimated, to a source point, the origin of the distance measurements, and likens the distance of the target point to the length of the shortest practicable path or paths.

An algorithm is disclosed that detects and analyzes possible lesions that are left in contact with a structural boundary. A morphological closing with a structuring element is performed along the boundary to detect lesions within the threshold of the structuring element. A deformable surface-based analysis is performed on distinctive surfaces of the structure for the identification of lesions. The integrated use of a deformable surface model and chamfer distance potential enables explicit representation of regularized, or smoothed, surfaces from which lesion candidates may be detected.

The invention provides a pedestrian detection method based on directional chamfering distance characteristics. The method comprises following steps of (1) inputting a to-be-detected image and preprocessing the to-be-detected image so as to obtain a to-be-detected edge image; (2) extracting the directional chamfering distance characteristics; (3) by taking a directional chamfering distance as a distance metric, comparing similarity of a profile of the to-be-detected edge image obtained in the step (1) with a profile of a template image and judging whether the to-be-detected edge image is a pedestrian image; and if the template image has been judged, judging that the template image is an image of a pedestrian profile. According to the invention, by expanding the original directional chamfering distance characteristics, adding characteristics in the profile direction, and further carrying out codebook and multi-scale describing on the characteristics, strong directional chamfering distance characteristics with scale invariance are formed; the anti-interference ability of the directional chamfering distance characteristics to noise is improved; and accuracy of the pedestrian detection system is sufficiently increased.

An object recognition method and an object recognition apparatus using the same are provided. In one or more embodiments, a real-time image including a first object is acquired, and a chamfer distance transform is performed on the first object of the real-time image to produce a chamfer image including a first modified object. Preset image templates each including a second object are acquired, and the chamfer distance transform is performed on the second object of each preset image template to produce a chamfer template including a second modified object. When the difference between the first modified object and the second modified object is less than a first preset error threshold, the object recognition apparatus may operate according to a control command corresponding to the preset image template.

The invention discloses a cloudscape rebuilding method combining satellite cloud pictures and natural images. According to the method, firstly, the natural images are subjected to pixel separation; the cumulus cloud pixel and external profile are extracted out; a cumulus cloud data set is generated on the basis of the multiplex forward scattering model modeling detail cumulus cloud; secondly, initial cumulus cloud is rebuilt from the satellite cloud picture; two-dimensional profile projection sampling is performed on different view angles of the initial cumulus cloud module and is used as a feature descriptor of the cumulus cloud model; then, the optimum matching in the cumulus cloud data set is retrieved by using the directed Chamfer distance as the measure; next, the corresponding cumulus cloud model surface detail features in the cumulus cloud data set are extracted; on the basis of the laplacian mesh deformation, the detail features are transferred into the initial cumulus cloud surface; finally, sampling is performed inside the cloud; a particle model is generated; drawing is performed. The method has the advantages that the features of the two kinds of data sources of satellite cloud pictures and natural images are combined; the cumulus cloud scene with the lifelike appearance and rich details can be rebuilt.

A structure preserving texture synthesis method based on the Chamfer distance comprises the steps that a texture sample drawing S and a texture characteristic pattern Fs are synthesized into a four-channel image S', and the format of each pixel is RGBA; the RGB channels of the S' store color values of the sample drawing, and the alpha channel stores a label value in a structure characteristic pattern corresponding to the sample drawing; an output image R' is initialized, an area is selected randomly from the S' and copied to the upper left corner of the R', the R' also has four color channels, the RGB channels store color values of an output texture image, and the alpha channel stores a label value in a structure characteristic pattern corresponding to the output texture image; blocks of the output image R' are synthesized sequentially one by one according to a scanning line; for a block to be synthesized currently, the distance between the block to be synthesized and the edge area of each block in the sample drawing is calculated, and the blocks, with the distances meeting a threshold value, in the sample drawing serve as candidate blocks; a block is selected randomly from the candidate blocks to be placed at the position for synthesis, and the block and a synthesized area form a certain overlapping area; a sewing path with the smallest error is found in the overlapping area, and a newly selected block is sewn into an output texture along the sewing path.

The invention relates to monocular localization in urban environments using road markings, in particular to methods and systems for monocular localization in urban environments. The method may generate an image from a camera at a pose. The method may receive a pre-generated map, and determine features from the generated image based on edge detection. The method may predict a pose of the camera based on at least the pre-generated map, and determine features from the predicted camera pose. Further, the method may determine a Chamfer distance based upon the determined features from the image andthe predicted camera pose, optimize the determined Chamfer distance based upon odometry information and epipolar geometry. Upon optimization, the method may determine an estimated camera pose.

The invention relates to an image-based track foreign body detection method, which comprises the following steps: acquiring a track image to be detected, and setting corresponding track matching templates for tracks with different forms; calculating an edge feature in the to-be-measured track image by using a Canny operator, and calculating a chamfering distance of the image according to the edgefeature to obtain a distance feature map of the image; carrying out convolution matching operation on different track templates in the distance feature map so as to determine the chamfering distance between each template and the target on the image, the smaller the chamfering distance is, the higher the matching value of the template at the position is; finding out a template with the smallest chamfering distance from the to-be-detected track image in the templates so as to determine a track foreign matter detection area; establishing a sample library, wherein the sample library is composed ofan image set containing contour information calibrated by foreign matters on a track; establishing a DenseUNet model, wherein the DenseUNet model mainly comprises a Dense module, a transition moduleand a deconvolution module; and training the DenseUNet model based on the data in the sample library, and identifying a foreign body position and a contour area in the track area by adopting the trained DenseUNet model.

The invention discloses a depth point cloud compression coding method based on a full-self-attention network, and the method comprises the steps: constructing a point cloud full-self-attention network which comprises an encoder and a decoder; obtaining training data, and constructing a chamfering distance objective function to train the point cloud full-self-attention network; inputting point cloud data into the trained point cloud full-self-attention network, performing feature sampling processing on the point cloud data by using an encoder to obtain a point cloud code, and completing point cloud compression; and reconstructing point cloud data by using a decoder according to the point cloud code to complete point cloud decompression. According to the method, learning of local and global correlation among points of the point cloud is enhanced based on network training of a chamfering distance objective function, point cloud codes capable of accurately representing semantic information of the point cloud are obtained through feature sampling of the point cloud by a coding machine, safety and stability of storage and transportation of the point cloud information are guaranteed, and the method is suitable for large-scale popularization and application. The method can be widely applied to the technical field of point cloud compression coding.

Provided are a method of extracting ridge data in a method of tracking motion of an object and an apparatus and method for tracking a joint motion of an object. The method of extracting ridge data includes selecting a point in the contour of an object from an image of the object, calculating a chamfer distance of the point in the contour, and determining whether the point in the contour is a ridge point based on relationships between the chamfer distance and chamfer distances of points in the contour adjacent to the point in the contour. The method of tracking a joint motion of an object includes acquiring initial information of the object including information on at least one ridge region representing a distribution of ridge points of the object, and tracking the joint motion of the object from temporally consecutive images of the object based on the initial information.

The invention provides a three-dimensional face model edge optimization method and system based on micro rendering, and relates to the technical field of computers. The method comprises the steps that: edge feature information is extracted from depth information in a derivable step; then, barycentric coordinate interpolation is carried out according to the edge feature information to obtain edge feature points of a three-dimensional face model; a total loss function Iall composed of a distribution loss function ldis, a chamfering distance loss function lchamfer and a regularization loss function lreg is constructed, and iteration is carried out on the total loss function Iall to obtain an optimal expression weight; and finally, an optimal three-dimensional face model is generated according to the optimal expression weight. The optimal three-dimensional face model has the advantages of high matching degree with the target face image and good authenticity.

The invention relates to the technical field of ship models and provides a ship model and a manufacturing method of the ship model. The ship model comprises multiple mutually connected ship model parts. Chamfers are formed in the splicing places of the ship model parts. Each chamfer comprises a first chamfer distance and a second chamfer distance. The specific value of the first chamfer distance and the second chamfer distance is 1.1-1.3. The first chamfer distance, the length of each ship model part, the smallest thickness and the cutting intensity meet the preset condition. By setting the splicing places between the ship model parts to be chamfer splicing, and setting the specific value of the first chamfer distance and the second chamfer distance in each chamfer, the size of each ship model part can be well controlled and part zeros are reduced. In addition, according to the length of the ship model part, the smallest thickness and the cutting intensity, the first chamfer distancesare set, so raw materials of the ship model parts can be used to the greatest degrees, resource utilization rate is improved and cost for manufacturing the ship model is reduced.

The invention discloses a God bead identity identification method and device based on image pyramid gradient histogram features. The method includes: adopting a portable image acquisition device with a network function to acquire a God bead image; extracting pyramid gradient histogram features of a rectangular region of the celestial bead image; based on the God bead template library established by the pyramid gradient histogram features of the God bead image, calculating the Chamfer distance between the pyramid gradient histogram features of the rectangular area of the extracted God bead image and the real God bead features collected and stored by the template library, and carrying out God bead identity authentication. A God bead identification challenge of shooting at different resolutions and any angle under different illumination environments is overcome; under the condition that the local features of the God bead image are lost, a robustness result can still be obtained, universality and robustness are achieved, the God beads can be automatically, efficiently, objectively and accurately recognized, and a simple, convenient, accurate and intelligent God bead identity recognition platform is provided for celestial bead fans and collection.

This application provides an AGV initial positioning method and positioning system based on a chamfer distance shape matching degree. The AGV initial positioning method comprises the following steps:generating a distance map according to an established environment grid map; preprocessing a 2D LiDAR point cloud; and performing matching search on the preprocessed 2D LiDAR point cloud and the generated distance map, to obtain an AVG initial position and a direction angle. According to the initial positioning method, a chamfer distance between an outline currently obtained by the LiDAR and the environment grid map is used as a matching measurement evaluation, and an AGV initial position and an azimuth angle are determined by finding a maximum matching degree without manual reconnaissance. TheAGV initial positioning method has a very high automation degree, and can significantly improve calculation efficiency.

The invention discloses a three-dimensional point cloud reconstruction method based on deep learning; the method comprises the steps that: the coordinate information of a three-dimensional point cloud is predicted through a point cloud predictor according to the inputted potential feature representation, wherein each branch takes potential feature representation output by the corresponding feature encoder as input, and learns complementary features combined with other branch information; by applying a cross-view interaction unit, each sparse point cloud reconstruction subnet captures cross-view complementary information and feeds back the information to the point cloud predictor to generate sparse point clouds; a global guidance dense point cloud reconstruction module composed of a plurality of point cloud feature extraction subnets, a global guidance feature learning subnet and a generation layer is constructed, each point cloud feature extraction subnet is composed of a series of multi-layer perceptron sharing weights, and the multi-layer perceptron extracts point cloud features from generated sparse point clouds; finally, chamfer distance loss is adopted as geometric consistency constraint, and semantic consistency constraint is constructed to optimize generation of dense point clouds.

The invention provides a three-dimensional reconstruction method and device, and the method comprises the steps: calculating the first direction information of each first data point in a generated point cloud, and calculating the second direction information of each second data point in a real point cloud; calculating a direction difference between each first data point and each second data point; calculating the distance difference between each first data point and each second data point; carrying out fusion processing to obtain a directional chamfer distance loss function; and training the reconstruction network through the chamfering distance loss function with the direction, and performing three-dimensional reconstruction on the to-be-constructed object through the trained reconstruction network. Therefore, when the difference between the generated point cloud and the real point cloud is measured, not only the distance difference is considered, but also the direction difference is added, so that the difference between the generated point cloud and the real point cloud can be measured more accurately, the generated point cloud is more helpful to approach the real point cloud in a network training process, and the quality of three-dimensional point cloud reconstruction is improved.

The invention provides a gesture observation likelihood modeling method for three-dimensional tracking. Firstly, a gesture state model and ormation of the scene where the gesture is located are defined as a similarity measurement likelihood model; the gesture state model adopts a classic three-dimensional gesture modeling method; the information of the scene where the gesture is located mainly relates to depth information similarity measurement extraction of a three-dimensional scene; and then, a gesture observation likelihood modeling method is created according to the foreground informationof the classic three-dimensional gesture modeling method and the high-dimensional well depth information matched with the Chamfer distance, and a three-dimensional tracked gesture observation likelihood model is obtained according to the gesture contour and the classic three-dimensional gesture modeling result.

The invention discloses a pipe plug of a fine grinding mill. The pipe plug comprises a metal block and a thread. The metal block is cylindrical and is arranged in the front-rear direction and provided with a round front surface and a round rear surface, the diameter of the outer circumferential surface of the metal block is 70 mm, the front-rear length of the metal block is 40 mm, and the thread is turned on the outer circumferential surface of the metal block from the front surface to the rear surface of the metal block. A groove is longitudinally formed in the middle of the round front surface of the metal block, and the longitudinal central line of the groove and the longitudinal central line of the metal block are located on the same perpendicular plane. The upper end and the lower end of the groove are in a circular arc surface shape and are matched with the outer circumferential surfaces of the upper part and the lower part of the metal block respectively. The front-rear depth of the groove is 5 mm, and the left-right width of the groove is 4 mm. The circumferential edge of the round front surface of the metal block is turned into a circle of chamfer at 45 degrees, and the chamfer distance is 1 mm. The circumferential edge of the round rear surface of the metal block is turned into a circle of chamfer at 45 degrees, and the chamfer distance is 1 mm.

The invention discloses a portable port chamfering device capable of being automatically adjusted along with the diameter of a bar. The device comprises a chamfering device body, a cutting mechanism is arranged in the chamfering device body, the cutting mechanism is used for adapting to the size of the diameter of the bar so as to conduct corresponding chamfering treatment, the cutting mechanism comprises a cutting motor, and a driving shaft is rotationally arranged on the cutting motor; and a fixed block is fixedly arranged on the driving shaft, a sliding block is arranged in the fixed block in a sliding mode, a connecting rod is fixedly arranged on the sliding block, the connecting rod penetrates through the fixed block, a rotating block is fixedly arranged at the other end of the connecting rod, and a cutting knife is fixedly arranged on the rotating block. According to the device, through cooperation of built-in clamping blocks and an air bag, the position of a blade can be adjusted to adapt to the end opening of the bar while the bar is clamped, then the chamfering distance is controlled through a built-in distance sensor, operation is easy and convenient, and the device can further adapt to bars of various diameters.

The invention provides a magnetic material chamfering system and method, and the system comprises front-section chamfering equipment which is disposed along a first direction as a feeding direction, and rear-section chamfering equipment which is disposed at the rear section of the front-section chamfering equipment and is disposed along a second direction which is used as the feeding direction and is perpendicular to the first direction. The output end of the front-section chamfering equipment is connected with the input end of the rear-section chamfering equipment, the front-section chamfering equipment and the rear-section chamfering equipment are respectively provided with the chamfering mechanisms, each chamfering mechanism is used for chamfering four edges parallel to the feeding direction, and after a workpiece is chamfered by the front-section chamfering equipment, the chamfering mechanisms are used for chamfering the four edges parallel to the feeding direction. And the direction of the workpiece is not changed, and the workpiece is conveyed to the rear-section chamfering equipment from the input end of the rear-section chamfering equipment to be chamfered. The chamfering precision is high, the chamfering distance is shortened, arrangement is more compact, the chamfering efficiency is high, chamfering adjustment is more convenient, the whole machine adjustment time is short, tools can be conveniently and rapidly replaced, and consumed time is less.

The invention provides a passenger flow identification and tracking method, and the method comprises the following steps: equally dividing an image into m parts of first equally divided images, performing foreground extraction to obtain a corresponding chamfer distance transformation image, selecting any part of first equally divided image as a template tree extraction image, equally dividing the template tree extraction image into n parts, respectively establishing template trees, matching a chamfering distance transformation image corresponding to the template tree extraction image with the template tree, rotating the chamfering distance transformation images corresponding to the remaining first equally divided images to the positions of the chamfering distance transformation images corresponding to the template tree extraction image, and then performing matching with the template tree, so as to determine whether there is a pedestrian or not; after the existence of the pedestrian is determined, tracking the position of the pedestrian; therefore, the recognition precision is improved through the matching of the template tree, the number of matched templates is further reduced in combination with the chamfering distance transformation image, and the efficiency is improved. The invention also provides a passenger flow identification and tracking system.

A method of extracting ridge data includes selecting a point in the contour of an object from an image of the object, calculating a chamfer distance of the point in the contour, and determining whether the point in the contour is a ridge point based on relationships between the chamfer distance and chamfer distances of points in the contour adjacent to the point in the contour. A method of tracking a joint motion of an object includes acquiring initial information of the object including information on at least one ridge region representing a distribution of ridge points of the object, and tracking the joint motion of the object from temporally consecutive images of the object based on the initial information.

The invention belongs to the technical field of visual detection, and discloses a method for measuring the seat posture of a hob cutter of a shield machine based on monocular vision. S1 knows the three-dimensional model of the tool holder and the calibration results of the camera. Through the camera projection model, the world coordinates of the corner points of the tool holder in different poses are converted into image coordinates, and the image coordinates of the corner points under each pose are stored to form the tool. Seat posture template library; S2 collects the actual image of the tool holder to be measured and performs image processing to obtain the coordinates of the corner points in the actual image and the chamfering distance transformation map of the binary image containing only the corner points; S3 through the chamfering distance transformation map , calculate the distance feature value between each template in the pose library and the actual tool seat, and use the template pose corresponding to the minimum distance feature value as the estimated pose of the tool seat to achieve accurate positioning of the tool seat; through the present invention, the tool seat pose estimation accuracy high and computationally efficient.

This application provides an AGV initial positioning method and positioning system based on chamfering distance and shape matching degree. The AGV initial positioning method includes the following steps: generating a distance map based on the built environmental grid map; preprocessing the 2D LiDAR point cloud ; Match the preprocessed 2D LiDAR point cloud with the generated distance map to obtain the initial position and orientation angle of the AVG. This application uses the chamfering distance between the contour currently acquired by LiDAR and the environmental grid map as the matching measurement measure, and determines the initial position and azimuth of the AGV by finding the maximum matching degree, without manual point selection, and has high automation degree; and can significantly improve computational efficiency.

A dzi bead identification method and device based on image pyramid gradient histogram features. Adopt the portable image acquisition equipment with network function to collect dzi bead image, extract the pyramid gradient histogram feature of dzi bead image rectangular area, the dzi template library based on the pyramid gradient histogram feature establishment of dzi bead image, calculate described extracted dzi bead The Chamfer distance between the pyramid gradient histogram feature in the image rectangular area and the real dzi feature collected and stored in the template library is used for dzi identity authentication. Overcome the challenges of dzi bead identification under different lighting environments, different resolutions and shooting at any angle. In the case of loss of local features of dzi bead images, robust results can still be obtained, which is universal and robust, and can Automatically, efficiently, objectively and accurately identify dzi bead, providing a simple, accurate and intelligent dzi bead identification platform for dzi bead lovers and collectors.

The invention discloses a handle bar of a fine grinding mill. The handle bar comprises a handle, a connecting ring, a connecting rod and a threaded rod, wherein the handle is located at the right of the handle bar, is cylindrical and is transversely placed; the external peripheral side of the right side of the handle is turned to form a circle of 45-degree chamfer, and the chamfering distance is 1 mm; the left side of the handle is connected to the right side of the connecting ring; the connecting ring is cylindrical and is transversely placed; the connecting ring and the handle are concentric circles; the left side of the connecting ring is connected to the right side of the connecting rod; the connecting rod is cylindrical and is transversely placed; the connecting rod and the connecting ring are concentric circles; the left side of the connecting rod is connected to the right side of the threaded rod; the threaded rod is cylindrical and is transversely placed, and the thread pitch is 1mm; the external peripheral side of the left side of the threaded rod is turned to form a circle of 45-degree chamfer, and the chamfering distance is 1 mm; the threaded rod and the connecting rod are concentric circles; and the handle bar is easy to operate and swift in action and has a good effect.

The invention discloses a three-dimensional point cloud reconstruction method based on deep learning, which includes: a point cloud predictor predicts coordinate information of a three-dimensional point cloud according to an input potential feature representation; Representation is used as input to learn complementary features that combine information from other branches; by applying cross-view interaction units, each sparse point cloud reconstruction subnet captures cross-view complementary information, which is fed back to the point cloud predictor to generate sparse point clouds; A global-guided dense point cloud reconstruction module composed of a point cloud feature extraction subnet, a globally guided feature learning subnet and a generation layer, each point cloud feature extraction subnet is composed of a series of multi-layer perceptrons with shared weights. The described multilayer perceptron extracts point cloud features from the generated sparse point cloud; takes the chamfering distance loss as the geometric consistency constraint, and constructs the semantic consistency constraint to optimize the generation of dense point clouds.

The invention discloses a flat grinding joint of a fine grinding mill. The flat grinding joint comprises a metal block and a threaded hole and is characterized in that the metal block (1) is cylindrical, and an external round face (2) and an internal threaded hole (3) are formed in the metal block (1); the diameter of the external round face (2) is 54+0.011mm to 54+0.030mm; the threaded hole (3) is downwards drilled in the center of the upper surface of the metal block (1), is a through hole and is drilled through from the upper surface of the metal block (1) to the lower surface of the metal block (1); the radius of the threaded hole (3) is 16mm, and the thread pitch is 1mm; the external round face (2) and the internal threaded hole (3) are concentric circles; the height of the metal block (1) from top to bottom is 28mm; the external peripheral side of the upper surface of the metal block (1) is turned to form a circle of 45-degree chamfers (5), and the chamfering distance is 1 mm; and the external peripheral side of the lower surface of the metal block (1) is turned to form a circle of 45-degree chamfers (4), and the chamfering distance is 1 mm.