228 results about "Boundary contour" patented technology

A disk drive baseplate includes a raised major surface oriented essentially parallel to the disk, a recessed major surface, and a boundary contour between the raised major surface and the recessed major surface. In one embodiment, the baseplate further includes an air channel adjacent to and at least partially outboard of the raised major surface, wherein the air channel includes the recessed major surface.

A method for segmenting intravascular images includes acquiring a series of digitized images acquired from inside a vessel, each said image comprising a plurality of intensities corresponding to a 2-dimensional grid of pixels, providing a precomputed set of shapes for modeling contours of vessel wall boundaries, wherein a contour can be expressed as a sum of a mean shape and a inner product of shape modes and shape weights, initializing a boundary contour for one of said set of images, initializing said shape weights by projecting a contour into said shape modes, updating said shape weights from differential equations of said shape weights, and computing a contour by summing said mean shape and said inner product of shape modes and updated shape weights.

The invention discloses an image semantic segmentation method of guiding feature fusion based on attention mechanism. The image semantic segmentation method comprises the following steps: (10) constructing an encoder basic network: using an improved ResNet-101 to generate a series of features changing from high resolution low semantics to low resolution high semantics; (20) constructing a decoderfeature fusion module: extracting high-level semantics with strong consistency constraint by adopting a pyramid structure module based on three-layer convolution operation, and performing weighted fusion on low-level stage features layer by layer to obtain a preliminary segmentation heat map; and (30) constructing an auxiliary loss function: adding auxiliary supervision to each fusion output in the decoding stage, superposing the auxiliary supervision with main supervision loss after the up-sampling of the heat map, and strengthening the hierarchical training of the model to obtain a semanticsegmentation map. The image semantic segmentation method of guiding feature fusion based on attention mechanism is high in accuracy and clear in boundary contour.

The invention discloses an airborne laser radar data-based building three-dimensional reconstruction method and aims to propose a three-dimension reconstruction method combining a building roof boundary with a roof topologic graph so as to realize automatic detection and three-dimensional reconstruction of a building. The method comprises the steps of firstly, filtering airborne LiDAR (Airborne Light Detection And Ranging) data to obtain a ground point and a non ground point; secondly, extracting building point cloud from non ground point cloud in combination with point cloud characteristic information; thirdly, segmenting a roof surface and extracting a boundary contour line; and finally, obtaining key line segments of a roof in combination with the building boundary and the roof topologic graph, and constructing closed polygons of each root surface and combination of the closed polygons so as to obtain a building roof model. The wall surface can be obtained by elevation information of DTM or the ground point, so that the building 3D (three-dimensional) model reconstruction is realized. According to the method, the application bottleneck of the airborne LiDAR data at the present stage is broken through to a certain extent, the complexity of a reconstruction process is lowered, the flexibility of 3D reconstruction is improved, and a breakthrough is provided for city 3D reconstruction.

The invention belongs to the technical field of aircraft engines, and discloses a method for the composite machining of a blisk flow passage. The method is a numerical control machining method of selecting the adaptive machining strategy based on regional characteristics. The method comprises the following steps of: (1) dividing a machining region; and (2) selecting the adaptive numerical control machining strategy based on the regional characteristics. In the method, a rough slotting numerical control machining method of combining milling, drilling milling and slotting milling of a cycloid is adopted, so that a medial cambered surface is approximate to a curved surface of a blade profile, and the boundary contour and the optimal cutter-axis vector of a rough machining region in a channel are determined. The method has the advantages that the machining method of region division and cutting-in changing enhances the rigidity and machining stability of blades; a process method of combining the milling and slotting milling of the cycloid improves the durability of cutters and the removal rate of materials; and the vibration phenomenon in the rough machining process is prevented effectively, the cutting force is reduced, the movement space of the cutters is expanded farthest, and the machining efficiency is improved.

The invention discloses a three-dimensional carotid artery ultrasonic image blood vessel wall segmentation method based on deep learning. The method comprises the following steps: (1) obtaining a three-dimensional ultrasonic image; (2) obtaining a two-dimensional ultrasonic image of a carotid artery cross section, and performing manual marking; (3) dynamically and finely adjusting the convolutional neural network model by utilizing the manually marked image block; (4) fitting vascular adventitia-tunica media boundary initial contour; (5) using the dynamically adjusted convolutional neural network model to carry out vascular adventitia-tunica media boundary contour segmentation; (6) obtaining a vascular cavity ROI region; (7)using U-Net network to divide the vascular cavity, and extractingthe vascular cavity-tunica media boundary contour through morphological processing. According to the method, the contours of vascular adventitia-tunica media boundary MAB and vascular cavity-tunica media boundary LIB can be accurately segmented out; the workload of doctors is greatly reduced, and the vascular wall volume (VWV), the vascular wall thickness (VWT) and the vascular wall thickness change (VWT-Change) can be calculated based on the segmentation result of the method.

The invention discloses a visual inspection method for chip electronic component position error based on edge fitting. The method is that a boundary contour point set of an entire image is obtained by adopting boundary following to generate the edge point chain structure of the image; the image center a chip electronic component is preliminarily calculated by adopting a contour gravity center method, a corner is roughly calculated by adopting a closest-to-center point method, two lines which pass through the image center and are perpendicular to each other are established, the two lines divides the image contour into four areas containing corner points, four corner points are obtained in each area respectively according to the distance from a point to the center and four line sides are formed, and the corresponding edge point set is extracted by using the four line sides to partition an edge; and the edge fitting is conducted to each edge point set by adopting an edge point set rectangle least square method to obtain the fitting rectangle of the edge, and the center coordinates of the fitting rectangle and the corner relative to a horizontal direction are calculated. The invention has the advantages that the center position and the corner of the chip surface mounting component can be accurately inspected and the calculation speed is fast.

The invention discloses an ROI image extraction method in finger vein identification. The method comprises that an original finger vein image is obtained from a collector; superhigh pixel segmentation is carries out on the original finger vein image to obtain a superhigh pixel image; Sobel operator edge detection is carried out on the original finger vein image to remove noise from the image and obtain a finger edge image; overlapped points of the superhigh pixel image and the finger edge image are obtained; tracking points are determined according to the overlapped points, edge tracking is carried out on the superhigh pixel image, and a complete finger edge image is obtained; the complete finger edge image is corrected to obtain a finger vein image with a fuzzy background; and angle correction, height cutting, width cutting and normalization are carried out on the finger vein image with the fuzzy background to obtain an ROI image. The extracted ROI image includes a complete finger boundary contour, and the segmentation effect is good.

The invention discloses a cooperative significance detection method based on superpixel clustering, and the method comprises the steps: constructing a superpixel pyramid, and replacing a common pixel through a superpixel block; speeding up the cooperative significance calculation; constructing the superpixel pyramid, so the feature information at different scale can be obtained, and the boundary accuracy of a cooperative significance target can be guaranteed. Based on the above, a clustering method can achieve the further classification of superpixel blocks, ad further speeds up the calculation of the cooperative significance. Finally, a method of fusion of a cooperative significance map and a saliency map is used for obtaining a final cooperative significance map, thereby guaranteeing the accuracy of the cooperative significance target. The boundary contour of the obtained cooperative significance target is more accurate in location, and the method has advantages in time and accuracy.

The invention discloses a power transmission line icing detecting and early-warning device. Filtering and edge detection are performed on image information transmitted through a network and acquired by a camera, coordinates of boundary contour points in an image are recorded through binarization processing and boundary tracking, finally, coordinates of boundary points before and behind a power transmission line are compared comprehensively, and the icing thickness of the power transmission line at the moment is obtained through calibration by the aid of the diameter of the power transmission line; data analysis is performed according to micro-meteorological parameters acquired by a data acquisition module, under the assistance of a lead tension change trend diagram and an icing thickness change trend diagram in a set time period, in combination with stored historical icing thicknesses, icing conditions and icing thicknesses in the next three to five days are predicted according to meteorological parameter conditions under which icing is formed easily, the line designed icing thickness and the actually measured icing thickness, and early-warning signals are sent out. With the adoption of the device, online detection is performed on icing of the power transmission line, meanwhile, dynamic early-warning can be realized, and analysis and early-warning can be performed on icing information timely.

The invention relates to a kind of method to matching the shape of persons' faces based on Shape Context. At first, preprocessing with the image pyramid and diffusion-filtering technology; then abstracting the boundary-outline information; carrying on the logarithmic coordinate transformation of the abstracted boundary-outline information to get bar chart of the logarithmic coordinate; calculating the value of Cs to get all the phase match points; calculating the value of the degree of approximation to judge the shape matching. The invention is characterized in simplicity, accuracy, being economical, expandability and so on, so the person-face shape matching can be used for entrance-exit control, safe verification, safe-guard monitoring, searching criminals and so on.

The invention relates to a traffic light lane level control relationship extraction method and device. The method comprises the steps that laser point cloud data and high definition real image data ofa road intersection in the forward direction are collected, and on the basis of a data fusion technology, two kinds of data are matched; by means of an image semantic segmentation technology, boundary contour coordinates of the traffic lights are extracted from a laser point cloud; characteristic information of traffic light images is extracted from the high definition real image data, and the types of the traffic lights are judged; the traffic lights and the corresponding lanes are matched according to the types, and the traffic light lane level control relationship is obtained. A self-driving car runs according to the lane level path, when the self-driving car passes through the intersection, position and attribute information of controlled signal lights is obtained from high definitionmap data, environmental perception and vehicle body positioning are conducted together with other types of sensors in a matched mode, and then traffic signals are interpreted in a targeted mode; theprocessing efficiency and accuracy degree of self-driving cars on the traffic signals are greatly improved.

A method for segmenting intravascular images includes acquiring a series of digitized images acquired from inside a vessel, each said image comprising a plurality of intensities corresponding to a 2-dimensional grid of pixels, providing a precomputed set of shapes for modeling contours of vessel wall boundaries, wherein a contour can be expressed as a sum of a mean shape and a inner product of shape modes and shape weights, initializing a boundary contour for one of said set of images, initializing said shape weights by projecting a contour into said shape modes, updating said shape weights from differential equations of said shape weights, and computing a contour by summing said mean shape and said inner product of shape modes and updated shape weights.

The invention relates to a 3D shape measuring method by identifying a boundary on the basis of abrupt change of the light stripe width, and belongs to the technical field of computer vision measurement. Light stripe images collected by a binocular visual measurement system are denoised in a difference image method, the light stripe width is changed in the boundary contour according to that reflection characteristics of laser stripes in the surfaces of objects of different materials are different, the boundary position is identified accurately according to change of the light stripe width, light strip information within the boundary is matched and reconstructed, and the 3D shape is measured rapidly in high precision. The measuring method comprises the steps of collection of images, denoising and synthesis of the light stripe images, extraction of the center of the light stripes, identification of the boundary contour, matching of the central points of the light stripes and reconstruction of the 3D shape. The method is high in measuring precision, reduces the redundancy computational complexity, improves the processing efficiency, and can meet requirements of rapid high-precision 3D shape measurement for general large composite components.

The invention provides an additive manufacturing method based on a vector sintering technology, and the method is realized through a vector sintering system. The system comprises a powder paving module, a vector heating module, a laser processing module and a temperature control module. When a part is processed, a coated powder body is preheated to a specific temperature not exceeding a sintering temperature through the temperature control module; a working cavity is descended by the thickness of a layer; while the powder paving module is used for finishing a powder paving operation in a moving process, the vector heating module at the rear end of the powder paving module is used for performing region selection sintering on powder in the part to obtain a plane sintering region covering a cross section slightly larger than a single-layer part; the laser processing module is used for scanning the powder body at the contour line part and invalidating the adhesion to obtain an accurate part boundary contour of the layer; after the operations of layer-by-layer descending, powder paving, vector sintering and contour invalidation scanning are performed, the stacking operation is performed to obtain a sintered part slightly larger than a part CAD (computer-aided design) model, the overall body is baked and reinforced, and few peripheral wastes are removed along a separation surface obtained by contour invalidation to obtain a final part.

The invention discloses an in-situ measurement method of adherent mushroom vision based on RGB-D camera; The RGB-D camera continuously collects the depth video stream; dynamic threshold segmentation is performed on each acquired depth image to remove the soil background, extract the mushroom connected domain and smooth the edge; and use the eight-neighbor tracking method for the extracted mushroomconnected domain Sequence traversal of its boundary contour, preliminary detection of its center and radius based on circular fitting, sequential extraction of boundary points within 1.3 times radiusaround the center of the circle and conversion to polar coordinates, denoising between the found adhesion points, Interpolation, in order to obtain the two-dimensional coordinates of the boundary contour of the single mushroom; calibrate the camera coordinate system, verify the accuracy of the in-situ measurement method based on the ceramic circular plate, and calculate the position, diameter, and deviation angle of the single mushroom in the camera world coordinate system . The method of the invention can accurately and quickly identify adhered round-like mushrooms, and the picking robot hashigh coincidence degree, short running time and high real-time performance.

The invention discloses a lead ice monitoring method and system based on three-dimensional reconstruction. The lead ice monitoring method comprises the steps: 1, installing two cameras which are different in position and have a certain angle for imaging a lead to be monitored, obtaining a boundary contour of the lead to be monitored; 2, acquiring ice images of the lead to be monitored, shot under the ice condition; 3, processing the ice images of the two cameras, which are obtained from the same section of line from different angles, to ensure that image points of the same point on an ice line in two images show different vision positions, obtaining a geometrical relationship between the two image points according to a camera imaging principle, and further figuring out three-dimensional coordinates of this point; 4, reconstructing a three-dimensional mode of line icing through acquiring the data of the three-dimensional coordinates of a large quantity of characteristic points; and 5, comparing the three-dimensional mode of the line icing with the boundary contour of the lead to be monitored, and figuring out an ice thickness. The invention provides a lead ice monitoring method and system based on three-dimensional reconstruction, which are capable of carrying out remote real-time monitoring on ice.

The invention discloses a shape matching and target recognition method based on a PCA-SC algorithm. The method comprises the steps of carrying out preprocessing on a target image, filtering part of noises in the target image, extracting the edge of the target image, extracting information of boundary contour points, working out the rectangular coordinate parameters of the contour points, converting the contour points from rectangular coordinates into polar coordinates, obtaining a corresponding logarithmic polar histogram of each point to forming a local feature descriptor, forming a covariance matrix, extracting a corresponding feature vector of a larger characteristic value of the matrix, adopting a linear transformation method to drop the matrix from high dimension to low dimension, forming a new characteristic matrix, wherein the new characteristic matrix is used for the shape matching and the target recognition, calculating matching degree, and obtaining a matching degree value between the target image and each template image. According to the shape matching and target recognition method based on the PCA-SC algorithm, characteristic extracting and effective representation for the image can be achieved, scale invariance, rotation invariance and translation invariance are achieved, accuracy rate and efficiency are improved, and interference of the noise is effectively restrained.

Provided is a method for effectively segmenting a hyperspectral oil-spill image. The method comprises steps of: defining an initial level set function and other related functions; acquiring a new fitting item in combination with a Fisher criterion; constructing an edge stop function to obtain a new length item; performing improvement in combination with an end member extraction algorithm; introducing a level set regular item to prevent reinitialization of the level set function; minimizing an energy function to obtain an Euler-Lagrange equation; setting parameters; selecting a display band and an initial contour; displaying a segmentation result graph; calculating various segmentation precision evaluation indexes; comparing and evaluating the accuracy of the segmentation results. The method can classify a target area in a simulated hyperspectral image and a real hyperspectral image, and effectively segments the hyperspectral oil-spill image with boundary blur and noise, improves the segmentation accuracy of the hyperspectral image, obtain a more accurate classification effect, makes the parameter change more stable, makes the contour curve more accurate, obtains the continuous and closed boundary contour, and has higher precision of segmentation.

The invention relates to an SWI (Susceptibility Weighted Imaging) image based cerebral vein vessel segmentation method, which comprises the steps of reading each two-dimensional SWI cerebral vein vessel image through magnetic resonance equipment, adjusting the image resolution, and removing the skull and skin covering the periphery of the brain in the images; performing anisotropic filtering enhancement processing on the processed images; then performing improved 2D Hessian matrix filtering enhancement processing; removing a large area of noise caused by artifacts of the brain, segmenting out most small vein vessels, and removing the boundary contour of a brain image; reserving the small vein vessels mistakenly removed; and calculating a DSC value, a PPV, the sensitivity and a Kappa value of each segmented image, and performing optimization on a region with the segmentation effect being poor. The cerebral vein vessel segmentation method can acquire a stable result in different SWI brain images with great difference and under noise and interference conditions, the false positive rate is low, segmentation of a structure which does not belong to the vein is avoided, and requirements of medical images for the safety are met.

The invention provides an image boundary scan method, which makes use of an image boundary scan system to scan the image boundary of a work piece. The system includes a computer and an image measurement machine station. The image measurement machine station includes a charge coupled device sensor, a working platform for carrying the work piece and a motor. The method includes steps mentioned below. The image of the work piece is intercepted through the charge coupled device sensor. The computer is equipped with parameters of the image boundary scanning. The image process is implemented on the image and a starting point is chosen from the processed image. A boundary point which is nearest to the starting point is found along the starting point. The boundary contour near the boundary point is extracted. A recursive method is used to search for other boundary points on the boundary contour. The boundary points are searched precisely on the processed image according to the parameters of the image boundary scanning and are regarded as scanning points for boundary scanning.

The invention discloses a dynamic three-dimensional tunnel section deformation detection and analysis system, method and device. The method comprises the steps that a scanner is controlled to acquiretunnel section information and rail section information to acquire a point cloud data set; an inclinometer is controlled to acquire two-dimensional altitude data of a detection device; a total stationis controlled to acquire trajectory data of the detection device; the two-dimensional altitude data and the trajectory data are subjected to data synchronization integration on the basis of a synchronization signal to generate positioning information; a tunnel three-dimensional space information data set is generated according to the point cloud data set and the positioning information, and thendeformation analysis is conducted on a tunnel; and whether or not a boundary invasion phenomenon is generated is determined according to the point cloud data set and a preset boundary contour. According to the method, the three-dimensional space information data set can be quickly acquired in a short maintenance period to conduct deformation analysis and boundary invasion analysis on the tunnel, and the data acquisition efficiency is improved; and the data acquisition precision can be improved through joint movement of the scanner and a vehicle-borne platform, and then accurate evaluation on the safety state of the tunnel is achieved.

The invention relates to a method for modifying the contour after changing the contour interpolation point in the contour picture, by obtaining the coordinates of n points in the plane point in the contour picture, and treating the interpolation point p(x, y) Perform hyperboloid interpolation. The method of the invention can effectively improve the interpolation error caused by the obvious trend of the data and the inconsistency of the spatial distribution of the data because the weight is quoted in the polynomial interpolation algorithm. The invention also relates to the establishment of a non-degenerate straight line method, and on this basis, a new algorithm for the relationship between points and polygons (points within polygons) is developed, and the algorithm is used in isoline boundary clipping to reduce computing time.

The invention discloses a notch defect detection method for a flexible IC package substrate line. The method comprises the following steps: obtaining target prior knowledge; extracting a boundary contour of a line of a flexible IC package substrate image to be detected to obtain a contour map; using the target prior knowledge as the constraint condition of Radon transform to carry out Radon transform on the contour map, extracting the boundary line in the contour map, and calculating a standard line width or a standard line space between the contours through the boundary line; determining a notch defect on the contour and a curve portion on the contour by aiming at each contour in the contour map; comparing the distance between each pixel point in the curve portion of each contour and thecurved portion of the next contour with the standard line space or the line width between the contour and the next contour, and determining whether the pixel point in the curve portion of the contourhas the notch defect or not according to the comparison result. The method can accurately and quickly detect the notch defects of the linear portion and the curve portion.

The invention discloses a pattern spot merging method for maintaining the contour characteristics of a structured ground object, which comprises the following steps of: A, acquiring original pattern spot data of a full overlay vector, and extracting the structured ground object; b, identifying a typical spatial distribution mode in the structured ground object; c, identifying aggregated structuredground objects forming a typical spatial distribution mode, and extracting peripheral boundary contours of the structured ground objects; d, merging the aggregated structured ground objects in the peripheral boundary contour to obtain a structured ground object merging processing result; carrying out merging operation on the original image spot data to obtain an original image spot data merging result; e, performing dissimilar operation on the peripheral boundary contour of the aggregated structured ground object and the original image spot data merging result, and determining each backfill position needing to be backfilled in the original image spot data merging result; and F, inlaying the structured ground object merging processing result into each backfill position in the original pattern spot data merging result to obtain a backfilled pattern spot data full-coverage merging result.

The invention discloses a meta-grid generation method and device based on geographic information. The meta-grid generation method comprises the following steps: inputting electronic map data in a target area range; preprocessing the electronic map data according to a set processing object to obtain a preprocessing result; determining a region range of a generated meta-grid framework according to the preprocessing result; traversing the meta-grid framework, extracting boundary feature points of a specific object, and executing a Thiessen polygon algorithm to obtain a plurality of irregular closed polygons; and aggregating all irregular closed polygons of the specific object to obtain a boundary contour, namely a meta-grid corresponding to the specific object. According to the meta-grid generation method, the meta-grids can serve as the basis of all the grids, and all the service types can be constructed in a unified mode based on the existing meta-grids in an aggregation mode when the grids are drawn, and the problems that the sizes, shapes and positions of the grids of the same service type are inconsistent after and after two times of drawing due to lack of a unified drawing benchmark are fundamentally solved.

The invention relates to a selective-failure-based 3D (three-dimensional) printing method for preparing a mold, and belongs to the technical field of 3D printing. The method comprises the following steps of: drawing three-dimensional CAD solid models of a casting and the mold; carrying out hierarchical slicing along the Z direction to form a boundary contour line of the casting-mold of each layer of slice; carrying out full region failure on powder of the contour line and the inner side of the contour line, or carrying out the full region failure on the corresponding part of the casting or forming a series of fragments which are intersected with one another and easily stripped by using a different-angle crosshatch selective failure method; after completion of hierarchical processing, carrying out the post-treatment enhancement on the powder, and the overall consolidation on the corresponding part of the mold; carrying out the full region failure on the corresponding part of the casting or forming easily stripped fragments, and removing by exerting the appropriate external force and other methods to form the mold. The method provided by the invention has the effects and benefits that the molding speed and efficiency are greatly improved when the mold is prepared, a failure region is conveniently removed, the details of the mold are effectively protected, and the molds with high dimensional accuracy requirements and complex shapes can be prepared.

The invention provides an image-based stripe cloth detection locating cutting method and system, and a storage medium. The method comprises the steps of obtaining detection parameters; detecting a main line position; according to the detection parameters, obtaining a boundary contour of a target region on stripe cloth; according to the boundary contour of the target region, obtaining an upper boundary line and a lower boundary line of the target region; and according to the upper boundary line and the lower boundary line of the target region, determining a cutting line. According to the image-based stripe cloth detection locating cutting method, in combination with a camera image collection system, a mechanical motion control system and a laser cutting system, automatic production of stripe cloth cutting is realized; and the position needed to be cut can be accurately located, the cutting is finished once, the sizes of the hems of cut pieces are ensured to be unified, and the line cutting does not need to be re-performed manually, so that the cutting precision is ensured, the labor force is liberated, and the production efficiency is greatly improved.

The invention discloses a method for automatically obtaining the geometric parameters of a stone based on machine vision, and the method comprises the following steps: (1), building a stone image collection platform; (2), collecting a calibration board image and a stone image, and selecting effective areas of the images; (3), carrying out the tilt correction, distortion correction and size calibration; (4), generating gray scale images of the effective areas according to different backgrounds and corresponding conversion formulas; (5), generating binary images of the effective areas; (6), detecting the boundary contour of the stone, and calculating the actual area and perimeter of the stone; (7), calculating parameters corresponding to different filling and cutting schemes according to the boundary contour of the stone. The method can obtain the geometric parameter information of the stone quickly and accurately, is simple and quick in operation, is high in precision of the measurement result, is low in implementation cost, and is strong in anti-interference capability.

The invention discloses a test method and system for testing an internal eye diagram of a chip. The method comprises the following steps: S1, erecting a test platform and inputting a reference voltagesignal and a clock signal at a to-be-tested chip receiving end to obtain a sampling point; S2, adjusting a clock signal phase and a reference voltage signal to change the location of the sampling point, and testing the obtained sampling point to obtain a boundary sampling point with an error code; S3, repeated the previous step to obtain a plurality of boundary sampling points, connecting the boundary sampling points to form a boundary contour of an internal signal of the chip so as to obtain eye height and eye width data and a signal simulation eye diagram; and S4, compensating the referencevoltage signal and the clock signal and determining the compensated eye height and eye width data to obtain a final internal signal eye diagram of the chip. According to the invention, with internalsimulation testing, parameter adjustment becomes convenient; the consistency of the testing result and the internal actual signal of the chip is ensured; and the precision of the chip signal eye diagram result is improved.