88 results about "Contour analysis" patented technology

An image processing system detects a moving person or other object of interest. The system generates a thresholded difference image by processing a video signal or other type of image signal received from a camera. The difference image is then segmented into regions bounded by vertical lines passing through the image, and silhouette candidates are identified in one or more of the regions. Tensor voting is used to determine saliency values and corresponding tangents for each of the silhouette candidates, and the resulting values and tangents are used to detect the object of interest. In an embodiment in which the object of interest is a moving person, a neck position of the moving person may be detected by analyzing a sum of x-components of tangents along a corresponding silhouette. The detected neck position may then be utilized to determine a head position and a head size for the moving person.

Ventricular stroke volume variation (SVV) is estimated as a function of the standard deviation of arterial blood pressure value measured over each of at least two cardiac cycles, preferably over each of the cardiac cycles in a computation interval covering a full respiratory cycle. In one embodiment, maximum and minimum standard deviation values are determined over the computation interval. SVV is then estimated proportional to the ratio of the difference between the maximum and minimum standard deviation values and the mean of the standard deviation values. In another embodiment, SVV is then estimated proportional to the ratio of the standard deviation of the standard deviation values and the mean standard deviation over the entire computation interval. A pre-processing arrangement for improving reliability of estimates of more general cardiac or hemodynamic parameters is also disclosed and involves smoothing with an approximating function, and sampling and low-pass filtering at an adjustable rate.

The invention relates to a building point cloud model and base map aligned method based on an outline, which belongs to the technical field of computer vision. The method comprises the steps of: performing structural analysis on the building point cloud model obtained by three-dimensional reconstruction, automatically detecting the posture of the current model, ultimately realizing normalization of coordinates of the point cloud model, and obtaining a point cloud overlooking outline drawing of a current building on the basis; performing outline analysis on a remote sensing image of an area in which a reconstructed building is positioned, and automatically detecting to obtain a base map outline; estimating structure turning points of the current building according to the outline of the reconstructed building, and then realizing space alignment of the point cloud model and a satellite base map by virtue of outline matching information, so as to obtain a mapping matrix from the point cloud model to the satellite base map and automatically realize the placement of the building point cloud model on the satellite base map. Compared with the prior art, the method has the advantages that the point cloud model and the base map can be accurately and automatically aligned, manual workload required by alignment is effectively reduced, and the alignment cost is reduced.

An adaptive automobile meter detection method based on character segmentation cascade classifier which processes threshold segmentation, morphology and connected components analysis on the original image; extracts the pointer based on the contour analysis method, establishes a pointer information list; constructs a character segmentation cascade classifier by combining a HOG / SVM character segmentation classifier, a character filter and a CNN digit classifier. The character segmentation cascade classifiers is used to identify the digit character area of the automobile meter. Region analysis is performed to extract tick marks based on the center of the digit character area. The angular position corresponding to the tick mark is determined. The response value corresponding to the pointer is determined by establishing the Newton interpolation linear description relationship between the pointer angle and the response value and then the meter is classified as pass or fail. The invention is suitable for the field of visual inspection of automobile meter pointers.

A method for automated landmark extraction from three-dimensional whole body scanned data extracts feature points as landmarks from a plurality of point clouds obtained from a three-dimensional whole body scanner, without pre-making the landmarks on a human body before scanning. The system includes: (1) Silhouette analysis, which projects three-dimensional body scanned data onto a two-dimensional surface and observes the variations in curvature and depth; (2) Minimum circumference, which uses the variations of the circumference of body parts to define the locations of the landmarks and feature lines; (3) Gray-scale detection, which converts the color information of human body from RGB values into gray-scale values to locate the landmarks with greater variations in brightness; and (4) Human-body contour plot, which simulates the sense of touch to locate landmarks by finding the prominent and concave parts on the human body.

The invention discloses a method for evaluating surface detects of small-sized workpieces. The method comprises the following steps of: acquiring surface images of small-sized workpieces under a specific light source with an industrial camera; detecting various defects such as unfilled corners, edge dropping, abnormity, pockmarks, cracks and the like of the small-sized workpieces based on image processing and profile analyzing methods; evaluating the severity and quantity of defects such as the unfilled corners, the edge dropping and the like of the small-sized workpieces by adopting a convexdefect evaluating parameter and a method for combining the convex defect evaluating parameter with a linearity evaluating parameter; and evaluating the degrees of the pockmark and crack defects with a threshold value segmentation method based on dimensional histogram statistic and the ratio of pockmark pixels. Based on the evaluation, the invention further provides a flow for detecting surface-unqualified workpieces. The flow has low false detecting rate, high detecting speed and wide application range, can meet the online detection requirement of the small-sized workpieces, and is suitable for detecting surface defects of various small-sized workpieces.

The present invention discloses a depth map based hand feature point detection method. The hand feature point detection method comprises the steps of: (1) acquiring a human body motion video sequence by utilizing Kinect for hand extraction, obtaining hand position information of a human body by utilizing OPENNI through the depth map, and preliminarily obtaining a palm point with a method of setting a search region and a depth threshold value; obtaining a hand contour by utilizing a find_contours function of OPENCV; accurately determining the palm point of the hand by finding the center of a maximum inscribed circle in the hand contour, and finding a maximum value M in the shortest distances by calculating the shortest distances m between all hand inner points and a contour point, wherein the hand inner point represented by M is the palm point, and the radius R of the inscribed circle is equal to M; (2)by continuously performing Gaussian smoothing on the hand contour, obtaining a CSS curvature graph in combination with a curvature threshold value, analyzing a limit value according to the CSS contour in the drawing to obtain coordinates of a finger tip point and a finger valley point of the hand, and completing the finger valley point unavailable according to the CSS curvature graph; and (3) completing a missing finger.

A system and method for analyzing the motions of an object based on the silhouettes of the object are provided. The system includes a foreground detector, a contour extractor, a model generator, a corner histogram generator, and a value of similarity measuring unit. The foreground detector detects a moving foreground object from an input image. The contour extractor extracts silhouette contour of the detected foreground object, and the model generator generates mean value histogram models as references to determine motions of the object. The corner histogram generator generates corner histograms of hierarchical multiband in the extracted contour signal, and the value of similarity measuring unit calculates a value of similarity between the generated corner histogram of a current frame and the average model histogram in a histogram unit, measures a value making a value of similarity with the calculated current frame histogram maximum, and determines the measured value as a posture of the object in the current frame.

Certain diseases of the retina are diagnosed by circular profile analysis of retinal parameters, such as thickness. Retinal thickness around a user-defined circle on the retina is measured by various ophthalmological techniques and ±mapped to a circular profile map. The circular profile map does not use segmentation of measurement data into arbitrary arcs, and thickness is mapped to a quasi-continuous range of display bands. The circular profile map is superimposed on a fundus image, or other two-dimensional image of the retina, allowing association of the circular profile map with the presence of blood vessels and other anatomical features. The simultaneous display of a series of circular profile maps generated from sets of measurement data taken at different times permits the ready visualization of the progression of retinal abnormalities.

An apparatus that continuously monitors the arterial pressure measured by a pressure sensor in an artery, which pressure is regarded as the reading Pao that approximately corresponds to the aortal pressure. In principle, the arterial pressure can be measured in the aorta, near the aorta, or in the arterial tree. To provide a second reading, the apparatus, via the input channel, continuously monitors the central venous pressure (CVP), which is regarded as the reading PIT that approximately corresponds to the intrathoracic pressure (ITP). The third reading is provided via the input channel as a reading Z which expresses the thoracic compliance. Via known algorithms of the pulse contour analysis, the apparatus calculates the stroke volume variation, using as the determining pressure the transmural pressure which is calculated according to the formulaPtransmural=Pao−f(C)*PITThe cardiac volume responsiveness indicator (CVRI) is calculated for mechanical positive respiration according to the formulaCVRI=k*(SVV / ΔCVP)or for spontaneous breathing according to the formulaCVRI=l−m*(ΔCVP / SVV).

An imaging apparatus 6 has means for illuminating a selected surface with a radar beam footprint, and means for profiling / processing the resultant radar returns in an efficient and logical fashion such as to derive radar attitude information in real time about a number of predefined axes associated with the radar which depends upon the relative dispositions of the radar and the selected surface and upon the radar beam footprint characteristics. Advantageously, a plurality of transmit beams 10,11,12 are used to image the surface and the processing arrangement has the capability of determining roll, pitch and / or yaw pointing data associated with the radar, such pointing data being determined by derivation of the attitude information and by selective input of terrain elevation data so as to take account of variations in the radar viewing geometry with terrain elevation. The inventive radar bears a definite low cost advantage over known radar designs, and retains utility for many applications, for example spaceborne and airborne applications.

A method for manufacturing a microstructure, which includes at least one fine feature on an existing feature, using an NSOM laser micromachining system. A microstructure device preform is provided. A portion of its top surface is profiled with the NSOM to produce a topographical image. This profiled portion is selected to include the existing feature. An image coordinate system is defined for the profiled portion of top surface based on the topographical image. Coordinates of a reference point and the orientation of the existing feature in the image coordinate system are determined using the topographical image. The probe tip of the NSOM is aligned over a portion of the existing feature using the determined coordinates of the reference point and the orientation of the existing feature. The top surface of the microstructure device preform is machined with the micro-machining laser to form the fine feature(s) on the existing feature.

The invention relates to a power station arrangement method, in particular to a mountain land photovoltaic power station arrangement method. According to the mountain land photovoltaic power station arrangement method, Google globe software and ArcGIS software are used for completing sloping field modeling, a TIN topographic map is generated, and therefore manpower consumption and material resource consumption are reduced; analogy computation is carried out on the TIN topographic map of a mountain land area through sunlight analysis software so that the generating-capacity computing accuracy can be improved; the mountain land trend is sufficiently depended, photovoltaic arrays are arranged along a mountain, the amount of steel used in supports and the occupied area are reduced, the construction difficulty is reduced, a large amount of capital is saved, and influences on the generating capacity are not serious; shadow contour analysis is carried out on the TIN topographic map of the mountain land area through the sunlight analysis software, locating points of the arrays are determined, and the effective method which is rapid and convenient to use is provided for northern hemisphere sloping field photovoltaic arrays.

The invention discloses a virtualized platform performance evaluating method based on program contour analysis, comprising the following steps of: acquiring the resource requests of a macro load by using a contour analysis technology, wherein the resource requests of the macro load includes various operation numbers of CPU (Central Processing Unit), virtual machine sensitive operation number causing context switching of a virtual machine, memory read-write number and cache hit ratio, disk read-write data volume and network read-write data volume; using micro-reference test to acquire the resource supply capability of a virtualized platform to be tested, wherein the resource supply capability includies various operation speed of a CPU, virtual machine sensitive operation delay, memory read-write speed and cache read-write speed, disk read-write speed and network read-write speed; and calculating the response time, CPU utilization ratio, disk utilization ratio and network utilization ratio of the macro load. The method uses the program contour analysis technology to acquire the resource requests of the macro load and uses the micro-reference test to acquire the resource supply of the virtualized platform to be tested; moreover, the macro-performance is analyzed and calculated so that the complexity and the cost of the test are decreased.

A heart parameter measuring method comprises following steps: synchronously detecting electrocardiosignals and photoplethysmography signals and carrying out preprocessing; carrying out waveform contour analysis for the electrocardiosignals and the photoplethysmography signals which are preprocessed to obtain pulse wave arrival time and pulse wave inflection point area ratios; calculating a pulse wave reflection coefficient ratio according to the pulse wave inflection point area ratios; and estimating the pulse wave arrival time and the pulse wave inflection point area ratios to obtain blood pressure and heart output. By means of calculating the pulse wave inflection point area ratios obtained from the electrocardiosignals and the photoplethysmography signals, the pulse wave reflection coefficient ratio is obtained, then the pulse wave reflection coefficient ratio and the pulse wave arrival time are estimated so as to obtain the blood pressure and the heart output, the blood pressure and the heart output are synchronously measured, accuracy is guaranteed, invasive-method measurement is omitted, and operation convenience under a movement state is greatly improved.

Certain diseases of the retina are diagnosed by circular profile analysis of retinal parameters, such as thickness. Retinal thickness around a user-defined circle on the retina is measured by three-dimensional optical coherence tomography or other ophthalmological techniques. Abnormally thin regions are identified by comparing a measured function of thickness vs. polar angle to a reference function of thickness vs. polar angle. A degree of abnormality is characterized by the ratio of the integral of the measured thickness function to the integral of the reference thickness function over the abnormally thin region, as specified by a range of polar angles.

The invention discloses an illumination invariant extracting method for complex illumination face recognition. The method is characterized in that the method comprises the following steps: (1) reading a face image; (2) carrying out multiple-dimensioned and multidirectional contour analysis on the image; (3) solving the multiple-dimensioned and multidirectional contour information of the image; (4) solving the contour information of each dimension of the image; (5) solving the contour direction information of each dimension of the image; (6) solving the main direction of the contour of the image. The invention solves the problem of poor face recognition effect under the complex illumination in the prior art, provides an effective illumination invariant extracting method on the basis of multiple-dimensioned and multidirectional contour feature to enhance the accuracy of the face recognition under the complex illumination; in addition, the method also can be applied to the field of imagesearch.

An object recognition method analyzes an imaged object based on its contour. Extracted contours are characterized by wavelets and slope sequence, and compared to sets of stored contours to recognize a known feature. If a match of sufficiently high confidence is not found, the image is distorted to simulate an incrementally different perspective of the imaged object, and the process of contour identification, characterization and comparison is repeated until a match of sufficiently high confidence is found. The cycle of image distortions allow two-dimensional images obtained from a monocular vision system to be analyzed for three-dimensional motion for optimal recognition performance.

The invention discloses a pedestrian space-time outline presenting method based on ellipse Fourier decomposition. The method comprises the following three steps of: firstly, updating a background model by using a multi-distribution model, and then extracting time-space outline of original pedestrian to be processed by a binaryzation, morphology operator and 8-communication outline tracking method; confirming the point correspondence based on MDL (Master Data Library), increasing CDI (Comprehensive Dissertation Index) feature representing the grayscale distribution feature for the target function to obtain accurate point correspondence; and finally analyzing the closure of pedestrian outline, considering the periodicity characteristic that any point returns to the original position finally by moving along the outline, using the ellipse Fourier decomposition, neglecting the high-frequency shape detain section and keeping the low-frequency overall situation shape information so as to represent the high-dimensional 2D+time outline vector to be the low-dimensional EFCShape vector. The invention maintains the effective information of the outline analysis, realizes effective compression of outline data and provides an efficient outline representing method for application such as the subsequent gait identification.

The invention discloses a grabbing point positioning method for naturally placed grape bunches on the basis of machine vision. According to the space grabbing range, grabbing precision and speed and other requirements of a bunch-type fruit sorting robot, a hardware platform for acquiring grape bunch images is constructed first; the colored images of the naturally placed grape bunches on the basisof machine vision are obtained and preprocessed to obtain binary images of the grape bunches containing main fruit stem regions, secondary fruit stem regions, fruit stalk regions and other regions; multi-dimensional feature vectors of the grape bunches and a BP neural network are constructed to extract the main fruit stem regions, and gradient direction contour analysis is adopted for locating grabbing points; finally, on the basis of binocular vision calibration parameters, image coordinates of the grabbing points are converted into spatial three-dimensional coordinate points to obtain spatial three-dimensional coordinates of the grabbing points of the naturally placed grape bunches, so that the robot automatically grasps the naturally placed grape bunches.

A 3D scanning apparatus configured to image a physical entity comprises a radiation projector for projecting a plurality of radiation stripes onto the physical entity, a detector for detecting striped radiation received from the physical entity; and a data storage device for storing the received radiation as a pixellated bitmap image, the apparatus characterised in that the physical entity, the radiation projector and the detector are geometrically arranged to constrain the number of times that each received stripe is permitted to occur in a given pixel row of the bitmap, the geometrica) constraint determining a stored occlusion classification comprising a plurality of types of occlusions and the apparatus additionally comprising an occlusion type processing means configured to utilise the classification for detecting the various types of occlusions in said received image data.

The invention relates to the technical field of image processing, in particular to a document correction method based on deep learning semantic segmentation, which comprises the following steps: classifying document pixels of an image to be corrected by using a deep neural network model to obtain a document semantic segmentation feature map; performing contour analysis on the document, and determining document deformation information in the to-be-corrected image; constructing an assistance correction plane, obtaining a transformation relationship of document deformation information in the to-be-corrected image on an auxiliary correction plane. According to the method, image correction is carried out through the transformation relation, so that additional hardware equipment is canceled to obtain enough shooting freedom degree, limitations caused by factors such as an operation threshold and an application scene are broken through, and image correction can still be accurately and reliably carried out when complex documents or documents with complex backgrounds are faced.

A method for automated landmark extraction from three-dimensional whole body scanned data extracts feature points as landmarks from a plurality of point clouds obtained from a three-dimensional whole body scanner, without pre-making the landmarks on a human body before scanning. The system includes: (1) Silhouette analysis, which projects three-dimensional body scanned data onto a two-dimensional surface and observes the variations in curvature and depth; (2) Minimum circumference, which uses the variations of the circumference of body parts to define the locations of the landmarks and feature lines; (3) Gray-scale detection, which converts the color information of human body from RGB values into gray-scale values to locate the landmarks with greater variations in brightness; and (4) Human-body contour plot, which simulates the sense of touch to locate landmarks by finding the prominent and concave parts on the human body.

An electronic template delineating distinct selected patterns corresponding to predetermined regions of interest in a sample part is used to limit analysis to those regions. The surface of the sample is first measured using conventional techniques. The data so acquired are used to identify boundaries between distinct regions, which are then compared to a predetermined pattern boundary in the template to find a best-fit match. The position of the pattern is then shifted to overlay the match, thereby automatically aligning the template's selected patterns with the regions of interest in the sample surface. As a result, profilometric analysis can be limited to the regions of interest. Correction factors are also assigned to each selected pattern in the template to account for physical differences in the corresponding regions of interest of the sample part that affect the profilometric measurement.

The invention provides a spinneret plate special-shaped hole size detection method. The spinneret plate special-shaped hole size detection method comprises the following steps: based on a line segmentand arc combination mechanism during spinneret plate hole silk outlet surface outer contour machining design, adopting universal non-rigid parametric matching model based on line segment and arc combination to carry out modeling and detection on any spinneret plate special-shaped hole, achieving the universality of special-shaped hole detection. The method comprises the steps: carrying out imageextraction, contour analysis and feature extraction through a standard special-shaped hole image, and automatically generating a universal matching model; correcting the rotation angle of the image contour of the to-be-detected hole; carrying out global and local connected region pre-registration and parameterized model non-linear registration based on least squares, and optimizing the positions of the feature control points, optimally registering the model contour reconstructed by the non-rigid matching model to the contour position of the hole to be measured, and solving the hole quality monitoring parameters according to a preset hole size feature parameter calculation method.

The invention discloses a fresh tea leaf recognition method based on contour analysis. An existing fresh tea leaf recognition method does not consider the problems of fresh tea leaf damage and shielding. The method comprises the following steps of firstly, acquiring fresh tea leaf images and extracting fresh tea leaf outer contours; carrying out polygon fitting on the contour, extracting contour feature points and judging the concave-convex property of the feature points; secondly, introducing directional outline width characteristics of the fresh tea leaves, and searching petiole points in convex vertexes; cutting out sub-contours from the outer contours of the fresh tea leaves; after fitting the sub-contours by an ellipse fitting method, identifying whether the sub-contours are leaves orbuds by using a trained support vector machine model with a linear kernel function; and finally, grading the fresh tea leaves according to the number of leaves of the fresh tea leaves and the presence or absence of buds, wherein the fresh tea leaves are divided into single buds, one bud and one leaf, one bud and two leaves, one bud and three leaves and unqualified raw material grades. The methodis high in efficiency, high in accuracy and is free of requirements for blade placement; and the identification accuracy of the partially shielded image can also reach a relatively higher level.

The invention discloses an intelligent video smoke vehicle detection method based on contour analysis, which comprises the following steps: (1) extracting a moving object from a road monitoring videoby using a foreground detection algorithm; (2) removing a non-vehicle target and tracking the vehicle target; (3) extracting the contour of the rear of the vehicle object, calculating a series of static and dynamic features based on the contour, and fusing them to form a feature vector; (4) using a SVM classifier to classify the feature vectors of each frame, and identifying the smoky vehicles bymulti-frame analysis, and automatically keeping the license plate, passing place, passing time and other evidence of smoky vehicles. The method can make up for the inefficiency of the traditional manual monitoring of smoky vehicles and reduce the false alarm rate. The advantage of the algorithm is more obvious in windy weather.

Systems and methods for detecting copies of images are described. In one example, the system utilizes a threshold based contour analysis technique that uses known contour ratios to determine if a scan of an image or document being tested is a first generation scan of an original or of a higher generation copy. In another example, a method for calibrating the copy detection system for use with a particular printer and scanner in described.

The invention discloses a ship hitting effect evaluation method based on jet flow edge analysis and hitting deviation feedback. The ship hitting effect evaluation method comprises the following steps of S1, collecting the video image information through the photoelectric holder equipment; S2, processing the video image information through a jet segmentation model, and extracting a jet contour and a ship contour; S3, judging whether the jet flow striking exists or not based on an extraction result, and if no jet flow exists, returning to the step S2; if the jet flow exists, entering S4; S4, judging whether the jet flow hits the ship or not based on the contour analysis, and entering S5 if the jet flow does not hit the ship; if the jet flow hits the ship, taking the position of the ship as a hitting coordinate, and entering S7; S5, calculating the deviation of the water cannon shooting by combining the coordinate position of the target ship and the position of the actual jet striking point; S6, converting the position of the actual jet striking point and the calculated deviation into a striking coordinate of the water cannon through a coordinate conversion method; and S7, feeding the striking coordinate back to a water cannon controller so as to correct the control parameters of the water cannon.

The invention discloses a pointer-type instrument reading method suitable for multi-view images based on machine vision. The method comprises the steps: inputting an instrument photograph, inputting instrument parameters, recognizing a dial circle, carrying out the cutting based on the dial circle, carrying out the image preprocessing, obtaining a binary dial, carrying out the contour analysis of a dial image, and extracting a scale region based on scale hypothesis; calculating the average radius from the scale center to the center of a circle, recording the average radius as the radius of a scale area, performing normal calibration based on dial scale feature extraction by using the radius of the scale area and the binarized dial image in combination with instrument parameters, identifying a dial circle again for the binarized dial image after normal calibration, and performing contour analysis; extracting the scale and pointer areas, cleaning the extracted pointer areas, acquiring a pointer deflection angle through calculation by using a slope fitting method based on multiple times of sampling and averaging, and substituting the pointer deflection angle, the range included angle and the start-stop scale readings into a reading calculation formula to obtain instrument readings through calculation.