107 results about "Edge tracking" patented technology

The invention provides a pattern detection method, a device, equipment and a system, which are applicable to the field of detection. The method comprises the following steps: collecting images comprising patterns to be detected; carrying out standardization treatment on the collected images to obtain standardized images; and comparing the standardized images with the pre-stored reference images to obtain image detection results. The standardization treatment comprises the gray scale binaryzation treatment on the collected images and the edge tracking treatment on the images after the binaryzation treatment. The embodiment of the invention avoids various defects caused by adopting a visual measurement method to detect the patterns, reduces the pattern detection cost and the incorrect judging rate, and improves the pattern detection efficiency and the consistency.

The invention discloses a cover defect detection method based on image processing. The method comprises steps that S1, an excircle contour radius and glue injection zone width of a standard image are acquired in a man-machine interaction mode; S2, edge detection on a to-be-detected cover image is carried out to determine an edge image; S3, edge tracking of the edge image is carried out to acquire edge points, edge fitting of the edge points is carried out to acquire a fitting circle center position and a fitting radius; in combination with the glue injection zone width of the standard image, the fitting circle center position and the fitting radius, the glue injection zone is determined; and S4, cover defect detection and identification are carried out. Through the method, properties of short time, high efficiency and good timely detection effect are realized, and the excellent detection effect is acquired in an actual production line.

The invention relates to the technical field of image processing, in particular to an unmanned aerial vehicle tracking method based on a twin neural network and an attention mechanism, which is applied to continuously tracking a visual single-target unmanned aerial vehicle. According to the method, weight redistribution of channel attention and space attention is realized by using two attention mechanisms, and the representation capability of the model on an unmanned aerial vehicle target appearance model is enhanced by using an attention model for template branches of a twin network; and search images are preprocessed in a multi-scale zooming mode, response graph calculations are separately carried out, inverse transformation of scale changes of the unmanned aerial vehicle in a picture issimulated in the mode, search factors capable of generating larger response values serve as scale inverse transformation of the unmanned aerial vehicle so as to correct the size of a frame used for marking a target, and the transverse-longitudinal proportion of the frame is not changed. According to the method, the tracking precision of 0.513 is obtained through testing (the average coincidence rate is used as a quantization precision standard), and compared with other leading-edge tracking methods, the method has the advantage that the performance is obviously improved.

An efficient method and system for multi-sample antialiasing in graphics processing is described. Geometric edges as well as implicit edges of primitives in a bin are identified by iteratively rendering bins of pixels. Selective multi-sample antialiasing is applied to pixels that are touched by either a geometric edge or an implicit edge; pixels that are fully covered are not antialiased.

The invention discloses a novel method for extracting building three-dimensional outline information by utilizing a stripe division method to quickly filter onboard LiDAR (light detection and ranging) point cloud data, and based on a random sampling consistency model. In the method, in allusion to high precision of the onboard LiDAR data per se, the mass onboard LiDAR point cloud data is directly processed; the method comprises the steps as follows: firstly performing classification filtering on the onboard LiDAR three-dimensional point cloud data through two-time stripe division in X direction and Y direction, so as to distinguish ground points from non-ground points; then performing triangulation on the non-ground three-dimensional point cloud data, and dividing building roof surface patches through the combination with the classification result, extracting LiDAR point sets contained in each building, and then, obtaining a edge point set of each building through edge tracking; and finally, using the random sampling consistency model to extract building edge straight lines from the edge point sets and form a building outline characteristic curve, so as to meet working requirements of digital city three-dimensional modeling and the like.

The invention discloses an unmanned aerial vehicle visual detection method for high-voltage line insulators. The method comprises the following steps of: obtaining a high-voltage line insulator image; converting the high-voltage line insulator image from an RGB model to an HSV model; detecting a brightness component of the HSV model by adoption of a cross gradient search method, and obtaining a direction gradient and a spatial brightness similarity so as to generate a foreground image; carrying out connected domain segmentation on the foreground image by adoption of a 8-neighborhood connected domain segmentation method so as to obtain an insulator connected domain image; carrying out edge extraction on the insulation connected domain image by adoption of an LOG operator so as to obtain edge information of the image; carrying out 8-neighborhood edge classification on an image edge point by adoption of a 8-neighborhood edge tracking method; for pixel class groups obtained through the 8-neighborhood edge classification, carrying out edge linear detection on the groups, and displaying a recognized insulator profile. According to the method, complete insulator connected domain images can be obtained, so that the insulation recognition correctness is effectively improved, the operation speed is effectively improved and the calculation complexity is reduced.

The invention provides a palm print image acquisition device and a palm print image positioning and segmenting method. The method comprises the following steps that a trapezoidal marker serves as a template, palm print images are rapidly matched through the image processing algorithm, and a local image required for positioning is found; the optimal threshold value of the local image is worked out, thresholding is conducted on the local image, the local image is converted into a binary image, and the palm contour line of the local image is detected through the eight neighborhood edge tracking algorithm; two contour feature points on the palm contour line are extracted; the connecting line of the two contour feature points serves as the Y axis, the center point of the connecting line serves as O, the perpendicular line of the connecting line serves as the X axis, a reference coordinate system XOY is built, a square box with the fixed size is selected from the position which is parallel to the Y axis and has a fixed distance from the Y axis in the reference coordinate system XOY, the square box servers a template, and a central sub-image is separated from the corresponding position in the original palm print image; normalization is conducted on the central sub-image.

The invention particularly discloses a visual identification-based precise shaft size measurement method, apparatus and system. The method comprises the steps of establishing a mapping relationship between a pixel size and an actual spatial geometric size of a to-be-measured precise shaft; obtaining multiple to-be-spliced precise shaft images; building a conversion model between a reference imageand the to-be-spliced images to perform conversion; based on CS and NSST algorithm combination, performing splicing on the converted to-be-spliced images, and performing fusion to form a whole image of the to-be-measured precise shaft; and for the fused whole image, performing pixel level edge tracking and preliminary locating by using a multilevel filtered interested edge detection method firstly, then based on Sobel operator and least square curve fitting combination, obtaining sub-pixel-precision edges, and according to the detected edges and the mapping relationship, calculating the spatial geometric size of the to-be-measured precise shaft. According to the method, the apparatus and the system, the image splicing data processing amount is effectively reduced; the detection efficiencyis improved; and real-time online detection can be performed.

The invention discloses a laser production technology for a radio frequency identification antenna capable of easily discharging waste. In the laser production technology for the radio frequency identification antenna, changes of positions and dimensions of positioning marks at two sides of a radio frequency antenna substrate in X and Y directions are detected; real-time transformation is carried out on the shape, the dimension, the position and the angle of an engraved pattern through an adaptive algorithm according to the materials, the thicknesses, the composite temperatures, the environment temperatures and the humidities of a used metal film and the antenna substrate and the antenna design pattern and composition condition; meanwhile, laser engraving is compensated according to the monitored working temperature in a laser scanning system; and the influence of a temperature drift of the laser scanning system on the dimension and the position of the engraved pattern is reduced, so that the error of the system for overlapping positioning of the engraved pattern and a printed pattern is controlled within 20 microns or to be smaller than 1/2 laser spot size; the waste discharge technology is relatively easy to implement; the key problems of pattern positioning and edge tracking in the industrial process of the laser production technology for the radio frequency identification antenna are solved; and the industrialization process of the laser antenna production technology is pushed.

The invention provides a method for positioning a slight expanded target based on gesture compensation. The method comprises the following steps: firstly, pre-processing a to-be-processed image by adopting the Gauss smooth filtering to eliminate the influence of noise on a subsequent algorithm; secondly, obtaining a two-value edge of a single pixel by adopting an edge tracking method, and filtering peripheral points of the profile of a target to eliminate interfering points; thirdly, computing a momental ellipse of the target and parameters of the target by using the peripheral points of the profile of the target to confirm the gesture of the target; and finally, obtaining a translation parameter and a rotation parameter of the target by using the positioning method based on the gesture compensation. Therefore, through the adoption of the method, the fact that the target is positioned in the situations such as rotation, translation and scaling can be realized.

The invention relates to an automatic first break picking method based on edge detection, in particular to an automatic first break picking technology based on edge detection, which comprises the steps of: first, inputting original seismic data and conducting graying processing on the seismic data, and then, carrying out seismic data edge detection by virtue of airspace differential operators and finishing the edge detection by corresponding templates and image convolution, finally, correcting the edge detection results with an edge tracking technology, completing the automatic picking of seismic data of the first break and outputting picking results; a manual selection method is used for selecting an edge point as a starting point; and the edge tracking technology mainly utilizes tracking 'bug' for tracking. When being used for picking up the first break from three-dimensional seismic data, the automatic first break picking technology based on edge detection can reduce time by 7 days to 10 days. Compared with the existing automatic first break picking technology with the accuracy rate of 80 percent, the automatic first break picking technology based on edge detection can realize the accuracy rate of 99 percent.

The invention provides a method for determining a complex geologic structure in a two-dimensional space according to seismic data and stratum geological data in petroleum and natural gas exploration. The method comprises the following steps: first, processing the seismic data to obtain overlaid seismic data, so as to determine horizontal distribution of an objective formation and the distribution of faults penetrating the objective formation; then determining fault shapes and contact positions between the faults and the stratum; performing fitting and interpolation on layer data, followed by connection point and undirected edge tracking so as to build a geological model undirected graph; allowing the undirected graph to be subjected to traversal processing according to a depth-first search algorithm to finally obtain a topological structure; and finally, performing traversal tracking for obtaining an inversion result. According to the invention, the method can correctly depict the characteristics of geologic structures including ordinary faults, reverse faults and the like, and depict special geologic bodies such as salt domes and lenses, and the practical geologic structures being subjected to geologic movements, such as extrusion, tension and cutting, and can accurately recognize complex structural hydrocarbon reservoir.

The invention provides a high-precision visual measurement method, device and system based on a bionic algorithm. The method comprises the following steps: establishing a mapping relation between a pixel size and an actual spatial geometric size of a to-be-measured object; Obtaining low-resolution images of the plurality of to-be-measured objects; Carrying out super-resolution reconstruction through a super-resolution algorithm based on a residual network; For the reconstructed image, extracting edge points by using a Canny edge detection operator, extracting corner points by using Hilbert transform, and carrying out edge tracking by using the edge points and the corner points as heuristic information through a fruit fly algorithm; And finally, obtaining a single-pixel edge by utilizing arelated mechanism, and calculating the spatial geometric dimension of the to-be-measured object. The device comprises a mapping module, an image acquisition module, a reconstruction module, an edge coarse detection module, a fruit fly detection module and a calculation module. The system comprises an objective table, a CCD camera, a two-dimensional workbench and the like. According to the invention, the field of view of single imaging is effectively expanded, the measurement cost is reduced, and the detection efficiency is improved.

The invention discloses a skipping rope counting method based on image information and belongs to the technical field of smart fitness sports. The method comprises the steps of 1, extracting image data from original video data and performing single frame processing on the image data to obtain a group of sequentially arranged single frame image; 2, determining a common reference area of all the single frame images and intercepting the reference area to obtain a group of reference images, wherein a skipping rope passes through the reference area each time; 3, performing binarization processing on each reference image and filtering out interference data; 4, performing edge tracking on each processed reference image to separate out a target; 5, extracting judgment data of the target and judging whether or not the target is the skipping rope according to the judgment data; 6, judging whether or not the time interval between every two adjacent reference images is shorter than the time of 10adjacent frames when the target is the skipping rope, determining the time interval as a repeat count without increasing a count if yes, and otherwise, increasing a count; 7, outputting and displayinga counting result.

A low-power full-rate semidigital DLL architecture using an analog-based FSM (AFSM). The AFSM is a mixed-mode FSM in which analog integration is substituted for digital filtering, thus enabling a lower power implementation of the clock and data recovery function. An integrated voltage is converted to a digital code by an analog-to-digital converter (ADC), and the digital code is used either directly or after (low frequency) digital signal processing to control a a controllable delay element, such as, a phase rotator, for data edge tracking.

The invention discloses a method for coating thickness measurement based on optical coherence tomography imaging technique and application thereof. The method comprises the steps of: constructing an optical coherence tomography imaging (OCT) test system; scanning a depth direction and a two-dimensional cross section of the surface of a sample material by using the OCT system and acquiring a three-dimensional scanning image of a measured object; and after denoising processing on the obtained three-dimensional scanning image, carrying out conversion to obtain binary image, carrying out segmentation to obtain an upper boundary and a lower boundary of a coating based on an edge tracking method, and calculating the thickness of the coating. With the disclosed method, the non-destructive automatic detection of the coating is realized; and the rapid, accurate, and wide-range measurement of the thickness of the coating can be realized. The method can be applied to measure the thicknesses of film coatings including tri-proof paint on the surface of a printed circuit board, a pigmented polymer coating, a porcelain protective coating, a plasma sprayed ceramic body, a biodegradable composite coating, a varnish coating, and a transparent or translucent tablet film coating.

The invention discloses an augmented reality technology Implementation method based on physical detection and tracking algorithm. The method comprises the following steps that a local marker file is generated, an augmented reality system is initialized, an image of a real scene is acquired, the edge of the image is extracted and optimized, template match is conducted on the edge of the image, and if the matching succeeds, a decision window in the region is generated, a position matrix corresponding to the template is imported, the position matrix is conducted to fine adjust and amend, the special effect of the augmented reality is rendered according to the amended position matrix, corners are extracted in the decision window, appropriate tracking algorithms are chosen according to the number of the corners, If the number of the corners is larger than the threshold value, the system enters the feature point tracking algorithm. If the number of the corners is smaller than the threshold value, the system enters the edge tracking algorithm, animation effects of the augmented reality are updated according to the tracking algorithms, different tracking algorithms are adopted according to the number of the corners, then the detection and tracking for three-dimensional physicals are achieved when rich feature points do not exist or without feature points.

An embodiment of the invention discloses a glass boundary defect detection method and device, a terminal and a storage medium, and the method comprises the following steps of acquiring an image corresponding to the glass to be detected; generating an edge contour map corresponding to the image through an edge detection algorithm and an edge tracking algorithm; using a k-cosine curvature algorithmto calculate the curvature of the edge contour map, comparing the curvature with a preset curvature threshold, determining that there are defects in the boundary of the glass to be detected corresponding to the edge contour map if the curvature is less than the curvature threshold, performing feature screening on the defects to determine whether the defects are actually defects on the edges of theglass to be detected; and if the curvature is greater than or equal to the curvature threshold, determining that there is no defect in the boundary of the glass to be detected. In addition, in the embodiment of the invention, a glass boundary defect detection device, a terminal and a storage medium are also disclosed. By adopting the invention, the speed and accuracy of glass boundary detection can be improved.

The invention provides a method for monitoring a running speed of a steel pipe. The method comprises the steps of taking a photo of the steel pipe with a CCD (Charge Coupled Device) camera at a predetermined time interval and consequently, obtaining an image sequence of equal time interval, performing median filtering on images in the image sequence, performing Gamma transformation on the median-filtered images to remove the background, calibrating the images, determining edges by virtue of an edge tracking algorithm, calculating an actual distance by which the steel pipe moves according to image calibration and edge detection, and performing curve fitting on the distance data by the least square method to obtain the slope of a straight line, namely, the moving speed of the steel pipe. The method is better than the traditional contact and non-contact speed measurement methods, and also well applicable to the severe environment in an industrial field, and furthermore, the measurement accuracy is improved and lots of manpower and material resources are saved.

The present invention discloses a local three-dimensional maximum inter-class variance segmentation method of a three-dimensional CT image, and belongs to the field of image segmentation, in order tosolve difficult segmentation and mis-segmentation problems caused by a non-uniform gray phenomenon of a CT image. The method mainly comprises the following steps: (1) initializing the size of the statistics calculation window, the weight template and the size of the three-dimensional local space, and carrying out continuation processing on the original image; (2) using the one-dimensional maximuminter-class variance segmentation algorithm and the edge tracking algorithm to extract the target contour of each layer of the image, and then using the morphological dilation method to obtain the marker image of the edge transition region; and (3) according to the pixel gray level, the weighted average of the neighborhood, and the weighted median of the neighborhood, calculating the three-dimensional histogram of each three-dimensional local space, and constructing a lookup table by using a recursive method to obtain the optimal segmentation threshold. Compared with the prior art, the methodprovided by the present invention has advantages that a three-dimensional CT image with uneven grayscale can be segmented and the time complexity is O(L3).

A storm curtain assembly (10) includes a header (11) and side rails (13) extending downwardly from each end of the header (11). A roll (15) is positioned in the header (11) and carries a fabric curtain (14) which has one end attached to the roll (15). A bar (16) is attached to the other end of the fabric curtain (14). A track (20) is formed on each side rail (13), and the edges of the fabric curtain (14) are normally received in the tracks (20). One of the walls (18) which forms the track (20) terminates at an upper edge (21) which is spaced from the header (11) so that if an edge of the fabric curtain (14) is displaced from the track (20), the displaced edge will be allowed to be received in the track (20) as the fabric curtain (14) is being received on the roll (15).

The invention discloses an online diameter detection method of varnished wires. The online diameter detection method includes steps of acquiring diameter images of the varnished wires; filtering noise in the diameter images; subjecting the diameter images which are treated by noise filtering to edge detection to obtain edge images; calculating the diameter of the varnished wires according to the edge images. By adopting the edge detection method combining the iteration method to segment threshold of the images and the edge tracking method based on image scanning, background and target of the edge images obtained are well differentiated, edge continuity, no fake edges and burrs are generated, edge width is single pixel, and edge details are well reserved, so that the online diameter detection method has higher detection accuracy. Meanwhile, compared with a conventional edge detecting method, the edge detection method costs less time, so that the online diameter detection method has higher detection speed. In conclusion, the online diameter detection method has high theoretical value and engineering application population value.