82 results about "Convex hull algorithms" patented technology

The invention discloses a method for extracting parameters of tree branches based on ground laser radar three-dimensional scanning and belongs to the technical field of point cloud data processing of the ground laser radar three-dimensional scanning and the technical field of visualization model simulation of tree forms. The method comprises the following steps: utilizing a ground three-dimensional scanner as a data acquisition platform, scanning and obtaining point cloud data of the tree branches, and dividing the point cloud data of the tree branches into different parts; stratifying the point cloud data in the height direction of the tree, extracting point cloud isolines of different heights of the tree by means of a convex hull algorithm, and constructing tree branch triangular mesh between adjacent isolines by means of a triangulation network growth algorithm; combining a triangulation network of different parts of a tree, and constructing an overall triangulation network model of the tree branches. The parameters of the tree branches are extracted by means of ray collision detection technology, the error between the parameters and the actual measured value is kept to be within 5%, requirements of forest measurement can be met, and the efficiency of forest measurement can be greatly improved.

The invention provides a traffic sign recognition method for a driverless car, and belongs to the technical field of image processing. According to the traffic sign recognition method, on the basis of algorithms such as a convex hull algorithm and a Hu invariant moment and transverse and longitudinal histogram scaling fast matching algorithm. Compared with existing traffic sign recognition methods, the traffic sign recognition method for the driverless car has the advantages of being large in recognition range, capable of recognizing a ban sign and an indicative sign, good in real-time performance, high in recognition accuracy and low in mistaken recognition rate.

The invention discloses an ink-quantity limiting method for realizing maximization of ink-jet printing color gamut. According to the method, the problem of ink-quantity over-limit of ink-jet printing is solved while the maximization of printing output color gamut is realized. The method is suitable for the field of high-fidelity ink-jet copying. The method comprises the following steps: judging Neugebauer primary color scale of a pattern according to the condition of printing with specific ink-quantity over-limit and confirming the maximal color gamut of the color space of the control value of ink-jet printer equipment; calculating the volume of the color gamut through a convex hull algorithm; and taking the coverage rate of the maximal volume of the color gamut as a criterion, correcting a correlation coefficient in a nonlinear interpolation algorithm through circulating calculation, and finally confirming an ink-quantity limiting algorithm. According to the method, the problem of print waste caused by the ink-quantity over-limit is solved while the ink-jet copying quality is increased at the aspect of color reproduction, the implementation is convenient, and the method has higher applicability in the field of high-fidelity color copying.

Disclosed are a method and system for detecting point. The method of this invention comprises the steps of providing a set of points representing measured time delays between two clocks, providing a turning indicator to measure a magnitude of turns per measurement interval, and identifying a turn at a given point when the turn indicator at said given point is above a given value. With the preferred embodiment of the invention, the turn indicator is given by the equation: turn_indicator=(rtt_min−rtt_estimate) / number_of_points; where rtt_min is the minimum round trip delay in all the measurement points, and rtt_estimate is estimated round trip delay by using the symmetric convex hull algorithm.

A checkerboard angular point automatic detection method under a fish-eye lens comprises the steps of selecting a particular region of a checkerboard and shielding useless information to reduce angular point detection time and raise detection precision; carrying out Harris angular point detection on the particular region; adding an angular point manual extraction step in an algorithm and manually adding the hidden angular point information; adding angular point combining operation according to the total number of checkerboard angular points and angular point separation situations; as for the checkerboard which is small in checkerboard rotating angle, slight in distortion and few in angular corner, sorting the angular points in a binomial fitting manner; and as for the checkerboard distorted under the fish-eye lens, calculating the sequence of the checkerboard angular points by using a modified convex hull algorithm. Simulation software is utilized to automatically detect coordinates of the checkerboard angular points, and under the condition of ensuring detection reliability of the angular points, the detection speed and precision of the angular points are greatly raised, and the complexity of manually extracting the checkerboard angular points is removed.

The invention discloses a building seismic damage analysis method, and relates to the field of seismic damage assessment. The method comprises the following steps of: setting a contour sampling interval, extracting a point cloud set of different height of a building, then, obtaining a contour section profile line through a convex hull algorithm, and converting three-dimensional point cloud of the building into a two-dimensional contour polygon sequence; extracting shape discrete parameters of the building; and calculating standard deviation of building shape parameters, and through a Fisher discriminant analysis, performing statistics on the standard deviation obtained in a step S22 or a step S23, thereby obtaining a discrimination function corresponding to three types of degree of damage of the building. According to the invention, edge extraction, shape clustering and discriminant analysis are fused, and problems in building contour polygon sequence extraction, shape discrete parameter calculation, irregular building block segmentation and damage analysis are solved effectively.

The invention relates to a single-camera-based gesture convex hull detection and palm positioning method. The method comprises the steps that a single camera acquires a gesture image; denoising preprocessing is carried out on the acquired image; gesture motion, color detection and segmentation are carried out; gesture contour tracking detection is carried out; gesture contour convex hull detection is carried out; and gesture convex hull convex defect detection and gesture palm positioning are carried out. The method uses a convex hull algorithm and a smallest enclosing circle algorithm to realize gesture convex hull detection and palm positioned. The problems of gesture variability and diversity and light factor interference vulnerability are solved. The accuracy, the stability and the practicality of gesture palm positioning are improved. The method has high research significance and broad application prospects in a gesture identification technology based on vision.

The invention discloses a calibration method and device for a hand-eye relation of a polishing operation arm. The calibration method comprises the steps that collected three-dimensional point cloud data of the surface of a calibration block is acquired; preprocessing is carried out on the three-dimensional point cloud data to remove corresponding noise points; a consistency algorithm is adopted tocarry out plane segmentation on the three-dimensional point cloud data with the noise points being removed, and an estimation planar model is obtained; according to a preset convex hull algorithm, the estimation planar model is solved, and a corresponding convex polygon is obtained; and by fitting vertexes of the convex polygon, an outer parameter rotation translation matrix RW(Cam) of a camera is calculated so that a transformational relation MCam(Tool) of a three-dimensional scanning camera coordinate system relative to the robot actuator tail end in robot hand-eye calibration can be determined. The technical problems that traditional visual sensor 3D calibration is not suitable for rapid automatic identification required by a polishing system is solved.

The invention provides a grain transportation vehicle grain loading state detection method and a device based on a visual image. The method comprises steps: image acquisition is carried out on the grain transportation vehicle through a color digital camera, and a planar color image for a grain box port and the grain is acquired; smoothing, edge, line detection and other filters are used for processing the planar color image, and a complete edge contour of the grain box port is acquired; an image processing method such as color space conversion and threshold segmentation is used for acquiring the area of the grain in the grain box in the image; a two-dimensional convex hull algorithm is used for acquiring convex points representing edges of the grain area, and the linear distances between the convex points and the four edges of the grain box port are calculated respectively; according to the distances measured in real time, the loading state of the grain at each distribution position of the grain box port is judged through a preset threshold, and detection on the grain loading state when a reaper unloads the grain to the grain transportation vehicle can be realized; and finally, the grain state and the corresponding operation instruction are displayed on a detection terminal in real time. The method and the device have the advantages of convenient mounting and maintenance, easy implementation, quick detection speed, high stability and high precision.

The invention aims to provide an SVD (Singular Value Decomposition)-based method for extracting joint features of multi-source face images so as to increase the recognition rate of face recognition. The method comprises the following steps of: A, extracting the grayscale image, the binary image and the intuitive feature image of a human face sample to serve as data sources and merging into a jointfeature; B, extracting the attribute value of the joint feature, calculating a reverse integral image, carrying out singular value decomposition on the reverse integral image to obtain the singular value of the reverse integral image and calculating a singular value matrix of the reverse integral image by utilizing the singular value of the reverse integral image; C: accelerating the feature calculation by utilizing the singular value matrix of the reverse integral image and trilinear interpolation to obtain a high-dimensional directional gradient histogram; and D: carrying out feature dimension reducing calculation on the directional gradient histogram by utilizing a local mean-based kernel nearest neighbor convex hull algorithm to obtain a low-dimensional face image directional gradienthistogram feature.

For the mapping in a non-linear color space of source colors belonging to a same constant-hue leaf, a lightness mapping function can be used which is based on a cusp lightness condition stating that a source cusp color of this leaf should be mapped into a corresponding target cusp color of this leaf. Source and target cusp colors are determined by the intersection of a source and target cusp lines with this leaf. The invention describes a method providing simplified source and target cusp lines based on a convex hull algorithm which are used instead to determine modified source and target cusp colors for the definition of the lightness mapping function.

The invention discloses a sea surface small target detection method based on multi-domain multi-dimensional feature fusion. The method comprises the following steps: carrying out the combined processing of echo data under four polarizations: HH, HV, VH and VV, and extracting the multi-dimensional feature information of a time domain, a frequency domain, a time-frequency domain and a polarization domain; through linear fusion of a polarization domain and a feature domain, multi-dimensional features are compressed into a 3D feature space, and the calculation cost is reduced while the high-dimensional information amount is ensured; and in the 3D feature space, a quick convex hull algorithm is adopted to obtain a false alarm controllable judgment area, and target detection is realized. The detection method provided by the invention has steady performance in different marine environments, improves the detection performance of the radar on the sea surface small target, and provides a new thought for the detection of the sea surface small target.

The invention provides a text image processing method which comprises: preprocessing a text image to obtain a binary image and a plurality of connected domains included in the binary image; obtaining convex hulls corresponding to the plurality of connected domains and a character zone circumscribing the convex hulls by using a convex hull algorithm; and subjecting the obtained character zone to character segmentation in the horizontal direction to obtain a plurality of character blocks distributed in the binary image; and merging the character blocks according to the heights of the character blocks in the binary image to obtain word blocks included in the text image. In addition, there is provided a text image processing device matching the method. The text image processing method and device can improve the universality and the accuracy of character segmentation.

The invention belongs to the technical field of geological modeling, and particularly relates to a three-dimensional geological model modeling method. The three-dimensional geologic model modeling method comprises the following steps: acquiring exploration data of a plurality of exploration holes, extracting drilling data, performing data interpolation according to the drilling data to obtain interpolation data, and integrating the interpolation data and the drilling data into modeling data; generating point cloud data through the modeling data, and generating a stratum curved surface throughthe point cloud data; searching a convex envelope curve by adopting a convex hull algorithm, and cutting the stratum curved surface according to the convex envelope curve to obtain a cut stratum curved surface; and generating a stratum entity through the stratum curved surface, extracting soil layer data from the exploration data, taking the soil layer data as project parameters of the stratum entity, and generating a three-dimensional geological model through the stratum entity. The stratum curved surface generated by the method is smoother and is completely controlled by parameters, and themethod has a good effect on stratum simulation.

An apparatus and method for determining a driving direction of a navigation system are provided. The apparatus includes: an input unit for receiving longitude and latitude coordinates, a road type, and a number of lanes; and a traveling direction processor for determining a current road on which the own vehicle is located using point information received through the input unit. When the road ahead is of the same type and the same lane, the driving direction is determined as straight forward, and when the current road and the road ahead are determined to have different types and lanes from each other, the driving direction is determined as a left turn / right turn using a convex hull algorithm and straight ahead; and an output unit for displaying on a screen which driving direction of the road is a left turn / right turn and straight ahead.

The invention discloses a three-dimensional modeling method. The method comprises the following steps: generating a topological surface morphology of medicinal three-dimensional marking data through a convex hull algorithm; initially generating a deformable sealed cambered surface which envelops all the acquired three-dimensional marking data points; iteratively approaching the deformable sealed cambered surface to the topological surface morphology by using a level set reconstruction method; and partially adjusting the deformable sealed cambered surface near each medicinal three-dimensional marking data, so that part of the deformable sealed cambered surface can be superposed with each three-dimensional marking data, thereby establishing a sealed smooth superpose model of which the accuracy approaches to the topological surface morphology of the medicinal three-dimensional marking data. By acquiring the medicinal three-dimensional marking data with medicinal ducts in the cardiac chamber, the medicinal three-dimensional mark data is performed three-dimensional surface reconstruction ao as to establish a three-dimensional surface model of which the accuracy approaches to the surface morphology of the endomembrane in the cardiac chamber; and the three-dimensional surface model can be used for supervising and assisting doctors to do researches of cardiac electrophysiology or arrhythmia.

A novel dynamic spatial clustering construction and visualization system is configured to intelligently and autonomously determine optimal office space divisions for improved space efficiency and / or productivity of an organization in a particular office floorplan. Instead of requiring a human operator intervention to input step-by-step instructions or floor plan map manipulations into the system, the dynamic spatial clustering construction and visualization system can autonomously and intelligently determine optimal spatial clustering from a given set of space segments, grouping properties, and geographic features of a computerized office floor plan, by machine-determined isolation and sub-grouping of space marker icons based on known common traits per icon, by determining separate clusters within each subgroup using a density-based clustering algorithm, by creating space marker boundary indicators to approximate required spaces without spatial waste, and by executing a convex hull algorithm on each defined cluster to define the new boundary per cluster without spatial waste.

A multicolor printing system splitting modeling method for maximizing a color gamut comprises steps of manufacturing and printing single color step-wedges of all primary colors, determining the ultralimit threshold value of each single color ink quantity, measuring the spectrum value of each single-color step-wedge subjected to equal ratio compression and establishing a corresponding relation; splitting an original multicolor printing system into printing system submodels; performing ink color space overall sampling to obtain the sampling samples under the sprinting system submodels, and obtaining the prediction spectrum reflectivity information values by simulation; targeting at all submodels, using a convex hull algorithm to calculate to obtain the ink color space overall sampling color gamut volumes; combining K submodels and using the convex hull algorithm to obtain the color gamut volume of each combination, and locking the combination corresponding to the maximum color gamut coverage rate when the maximum color gamut coverage rate is larger than the preset threshold value, thus realizing the splitting modeling of the original printing system. The method has the technical advantages of obviously reducing the splitting modeling complexity of the system, avoiding the ink quantity ultralimit problem, improving the stability of output colors and the like.

The invention provides a system and a method for wireless network signal scope sensing and displaying. A vehicle-mounted terminal device periodically obtains present wireless network signals and a latitude and longitude coordinate, and periodically sends the wireless network signals and the latitude and longitude coordinate back to a rear-end cloud server. The cloud server can draw a transmission scope of each region according to the wireless network signals and the latitude and longitude coordinate set back by the vehicle-mounted device and by means of a convex hull algorithm. Thus signal conditions in road segments which no vehicle passes can be compensated. Conditions that a same area is covered by a plurality of regions are analyzed according to the transmission scope. Which area is not covered by any region or is only covered by one region and located at the edge of the region is determined according to the region covering conditions, thus determination of poor mobile communication receiving or communication corner areas is realized, and reference is provided for operating supporting.

The invention provides a multi-color halftone equipment split modeling method and system oriented to a copy object. The method includes the steps that multi-color halftone equipment is split into multiple four-color models containing black ink in an enumeration mode; a Nigboer primary color step wedge prepared in the ink quantity limited modeling process is used for acquiring color gamut of the four-color models through a convex hull algorithm; a color gamut judgment method is used for acquiring color gamut coverage rates of different four-color model combinations on a to-be-copied image, based on the principle of maximization of the color gamut coverage rates, the four-color model combinations are determined; aiming at the probable condition that multiple sets of combinations meet condition requirements at the same time, a center aberration minimization method is used for determining a final optimal four-color model combination, and equipment split modeling is completed. Aberration samples are prepared in the ink quantity limitation process, under the condition without an additional sample preparation step, the overall modeling process of the multi-color halftone equipment is simplified, and implementation is convenient.

The invention relates to a method for accurately dividing a blade cross section character point cloud. According to the method, firstly, initial dividing is achieved by extracting a midpoint of every two adjacent points in the point cloud; secondly, the parameters of a front edge and the parameters of a back edge are accurately extracted; thirdly, the intersections with the front edge and the intersections with the back edge are deleted from a blade pot point set and a blade back point set respectively so that the character point cloud can be accurately divided, wherein the blade pot point set and the blade point set are formed through initial dividing. The method overcomes the defects of a convex hull algorithm and an algorithm improved on the basis of the convex hull algorithm without loss of generality, and is suitable for dividing cross section character point clouds of blades within a wider range, and accuracy of extracting of the subsequent blade molded face feature parameters is guaranteed.

The invention discloses a human body posture visual identification method of a transfer carrying nursing robot. According to the method, a two-stage network (the first-stage neural network is PAF andthe second-stage neural network is improved ResNet) based on RGBD (color image RGB + depth image Depth) is utilized to realize close-range higher-precision identification; a ResNet network commonly used for classification is improved, and input and output structures of the ResNet network are modified, so that the ResNet network can solve the problem of human joint recognition, serves as a second-level neural network, and obtains higher precision compared with the existing method. According to the nearest neighbor insect following method, under the condition that only two joint coordinates aredepended on, the armpit contour of the human body is automatically tracked, the armpit contour is further repaired through the convex hull algorithm, the armpit center, namely the armpit point, is finally obtained, and by means of the armpit point recognition method, the dependence of the armpit point on the joint position is reduced, and meanwhile accuracy is improved.

The embodiment of the invention provides an outer contour point labeling method and device, electronic equipment and a readable storage medium, and relates to the technical field of image processing. A to-be-processed image is obtained, the to-be-processed image comprises a target area, the target area represents the area where a to-be-labeled object is located, and three-dimensional model parameters of the to-be-labeled object in the target area are obtained; and a three-dimensional model is constructed based on the three-dimensional model parameters and comprises a three-dimensional outer contour point candidate area and a three-dimensional outer contour vertex. A three-dimensional outer contour point candidate region is projected to the to-be-labeled object to obtain a two-dimensional outer contour point candidate region of the to-be-labeled object. Based on a convex hull algorithm, screening is performed from the two-dimensional outer contour point candidate region to obtain a target outer contour point. Finally, the target outer contour point on the to-be-marked object is marked to obtain an outer contour point marking result. Thus, labeling of the two-dimensional human body outer contour can be rapidly and accurately achieved so that a large amount of training data can be obtained, and therefore the accuracy of the training result is improved.

The invention discloses a method for constructing an external rectangular frame of an obstacle point cloud, and the method comprises the steps: a step of constructing an external convex polygon: obtaining the obstacle point cloud, projecting the point cloud to a two-dimensional plane in a top view projection mode, and drawing the external convex polygon for an obstacle point convex hull algorithm,thereby obtaining a convex polygon corner set; a polygonal corner point selection step: reducing the convex polygonal corner point set; traversing candidate boxes, namely drawing an external rectangular box for the convex polygon corner set by taking a connecting line of every two points as the direction of one side of the rectangular box, counting the sum of distance values from each corner to the nearest side of the rectangular box, recording the box and the distance values, traversing each point pair in the central point set, and drawing a candidate box; and selecting the frame with the minimum distance value as a final obstacle external rectangular frame. According to the method, on the basis of an obstacle point cloud external convex polygon, every two points are used as a projectionedge picture frame, the point-to-edge distance is used as a judgment standard to select a final frame, and a better effect is achieved on a sparse point cloud picture frame.

A method and apparatus for determining a boundary of a work area are provided in an embodiment of that present application. The method include: acquiring a plurality of track points of a movable object when performing a work on the work area; generating an initial trajectory range according to the plurality of trajectory points; dividing the initial trajectory range into a plurality of grids according to a preset interval; determining a boundary of the work area according to a positional relationship between the plurality of track points and the plurality of grids. The embodiment directly determines the boundary through the collected information of the trajectory points, effectively improves the accuracy of the boundary, and solves the problem that the boundary error of extracting the workarea by the convex hull algorithm in the prior art is large.

The invention relates to the field of industrial automation virtual simulation, in particular to a mixed collision detection method based on a convex hull and pre-judgment, which comprises the following steps of: 1, setting a similar-to-sphere threshold value and an inclined included angle threshold value; 2, calculating a convex hull of a model by using a rapid convex hull algorithm, and calculating a central position of a bounding box and a model similar-to-sphere rate according to the convex hull of the model; 3, when the similar-to-sphere rate is smaller than the similar-to-sphere threshold value, adopting an SBB; otherwise, calculating the inclined included angle of the model, and if the inclined included angle is smaller than the inclined included angle threshold value, selecting an AABB; when the inclined included angle is greater than the inclined included angle threshold value, selecting an OBB; and 4, carrying out intersection test on the bounding boxes. According to the method, the appropriate bounding box type can be selected based on the convex hull characteristics of models, the intersection test efficiency is improved, and the tightness of constructed bounding boxes is better.

The invention discloses a method for analyzing the collection range of a gravity pipeline pumping station, which includes such steps as modeling the data of drainage pipeline, connecting drainage pipeline to each other to obtain drainage pipeline network, carrying out topological inspection by suspension inspection and connection analysis after connection is completed, and carrying out topologicalinspection by connecting the drainage pipeline network to the gravity pipeline pumping station to obtain drainage pipeline network; analyzing the flow direction of the established drainage pipe network and marking the flow direction of the drainage pipe line with an arrow; starting from the pumping station, each branch of incoming water being tracked upstream in the direction of water flow according to the depth-first algorithm until all pipelines and pipe points at opposite junctions of water flow direction are found; carrying out spatial position analysis on the pipeline and the pipe point,respectively judging the relationship between the pipeline and the road surface, the district or the factory area; all the coordinates of the points in the result set being traversed in turn, and theouter contour being obtained by a Graham convex hull algorithm, which is the water collection range of the pumping station. Through the method, the water collection area of the backbone network can be known, and the corresponding drainage pump station of the pipeline can be quickly found.

The invention relates to a large-scale point cloud data-based large-scale high-speed rotary equipment error hybrid evaluation method. The method is a flatness error evaluation method according to a minimum area criterion. Updating the speed and the position of the planeness error solving particle swarm algorithm; simulating an annealing algorithm mechanism, and determining a transition probabilityfunction according to Gibbs criterion regular distribution; optimizing calculation is carried out, falling into a local optimal solution is avoided, and the solving precision is improved; removing invalid data from an optimization calculation result by adopting a convex hull algorithm to obtain effective measurement data points, and solving the planeness by utilizing a particle swarm intelligentoptimization algorithm. The flatness shape error evaluation value of large-scale point cloud data can be accurately obtained, and the method is suitable for measurement and evaluation of geometrical shape errors of large-scale rotary products. The method is simple and convenient in calculation and high in solving precision, is used for evaluating the shape error of the rotor connecting surface ofthe aero-engine, and can improve the final measurement and evaluation precision, thereby improving the assembly precision.

In view of how to analyze the operating state of a transformer and perform fault diagnosis based on a vibration method, the embodiment of the invention discloses an online abnormality detection methodfor transformer vibration. The method comprises: obtaining new data samples about a vibration signal of the transformer operating state within a preset time period; extracting the characteristic parameters of the new data samples based on wavelet packet analysis; completing the learning of the new data samples based on a fast convex hull algorithm, training and updating a single-class abnormalitydetector model; and calling the updated single-class abnormality detector model to perform online abnormality detection of the current transformer vibration. According to the online abnormality detection method for transformer vibration, since the incremental learning algorithm is used to learn only the new data samples within the preset time period, the new data can be processed in time and effectively, the new training samples can be learned online in real time, and rapid upgrade of detection models can be achieved. At the same time, time and space requirements of the model update can be reduced.