81 results about "Convex hull algorithms" patented technology

The invention provides a three-dimensional object collision detection parcel grid generation method and system, a storage medium and a program product, and the method comprises the steps: extracting all vertex data of a first object according to a sub-Mesh data set after serialization of a three-dimensional model, and determining an initial center point; each vertex of each piece of sub Mesh data is converted from a local coordinate system to a global world coordinate system, and a point cloud data set is generated; constructing an initial convex hull by adopting a matrix hierarchical extremum retrieval algorithm, and generating a final convex hull by adopting a conflict graph and incremental convex hull algorithm; initializing a radius coverage model by taking the geometric center of the model as a sphere center; performing subdivision according to longitude and latitude to generate a hierarchical ball grid; emitting rays, correcting vertex positions layer by layer, approaching the surface of the model, and outputting a collider network. Through the overall point cloud set, the incremental convex hull algorithm and the spherical grid contraction technology, unified topological wrapping of the multiple three-dimensional objects is achieved, the visual fidelity and the collision precision are ensured, and resource consumption is reduced.

The invention discloses a natural growth apple three-dimensional attitude recognition method based on a real-time multi-task fusion algorithm, and the method comprises the steps: obtaining an aligned RGB-D image through a depth camera, constructing a data set in combination with an image enhancement technology, and synchronously training a classification, segmentation and key point detection model through a multi-task deep learning framework; judging the contour form of the apple through the classification model, and carrying out contour fitting by adopting a convex hull algorithm and an arc adjacency matrix ellipse detection algorithm; two-dimensional centroid positioning and three-dimensional point cloud analysis technologies are fused, and bimodal calculation in the three-dimensional fruit axis direction is realized in combination with a fruit stem and calyx key point detection result. According to the method, the precision problem of apple posture recognition in a complex natural environment is effectively solved, the anti-interference capability is remarkably improved through the three-dimensional point cloud denoising and feature value analysis technology, and reliable technical support is provided for automatic picking and growth monitoring.

The present invention relates to the field of nondestructive testing technology and provides a composite material damage imaging method based on the spatiotemporal cross-correlation of ultrasonic guided wave wavefields. The method comprises the following steps: S1, acquiring a wavefield signal of the plate being tested; S2, performing narrowband signal extraction on the wavefield signal; S3, filtering the wavefield signal after narrowband signal extraction; S4, shifting the wavefield signal to obtain wavefield signal cross-correlation features; and S5, using a convex hull algorithm to obtain damage imaging results. This method innovatively extracts the difference coefficient between wavefield signals as a basis for damage location, providing a new approach to feature extraction algorithms based on the time-frequency domain. This method enables automated identification of composite material delamination damage and is sensitive to both surface and internal damage.

The invention provides an automatic detection method and system based on a GNSS receiver hardware module, and relates to the technical field of electronic equipment testing, and the method comprises the steps: carrying out the analysis of a preprocessed signal data set, extracting a signal feature boundary based on a convex hull algorithm, and carrying out the detection of the signal feature boundary; calculating key performance indexes such as signal acquisition time, tracking precision, bit error rate and receiving sensitivity; comparing the key performance index with a preset performance threshold value, automatically judging whether the hardware module of the GNSS receiver is qualified or not according to a comparison result, and generating a performance judgment result; and automatically generating a structured detection report based on a performance judgment result, and outputting the report to a specified terminal or a storage position. According to the invention, by simulating GNSS signal excitation, response signal acquisition, a signal processing algorithm and an automatic performance judgment criterion, a scheme for efficiently and fully automatically detecting various key performance indexes of the hardware module of the GNSS receiver is realized.

The invention discloses a tunnel back break analysis method based on a dynamic center line and an incremental convex hull, and belongs to the technical field of tunnel engineering detection. According to the system, the problems that a traditional detection method is low in efficiency and large in error are solved by fusing dynamic slice projection, robust registration and intelligent contour generation technologies. The method specifically comprises the following steps: carrying out configurable downsampling (the density is reduced to 0.2%-0.5% at an interval of 50-1000 points) on an original point cloud, constructing a tunnel three-dimensional model (the side length of a grid is less than or equal to 0.1 m) by adopting a dynamic center line extraction algorithm, and carrying out ICP (Inductively Coupled Plasma) registration (error lt; the method comprises the following steps of: synchronously extracting a design / actual measurement section contour by using a combined slicing technology, generating an actual measurement polygon based on an incremental convex hull algorithm (a filtering threshold value is 20 points), and calculating an over-break volume (an error is less than or equal to 1.5%) by using a spatial Boolean operation. A dynamic center line extraction and incremental convex hull algorithm is innovatively introduced, automatic generation of a color gradation analysis atlas (over-excavation RGB (255, 69, 0) and under-excavation RGB (30, 144, 255)) is supported, and full-process closed-loop processing is realized.

The invention discloses a method, device and equipment for reconstructing an assembled grain three-dimensional model, and relates to the technical field of three-dimensional model construction.According to the method, end face measurement data is used for constructing a grain digital model, and aiming at the problems that an existing point cloud reconstruction technology depends on large-space detection (shooting) equipment and is lack of reconstruction model precision, an end face shooting mode is adopted, and the reconstruction precision is improved. And integrating optimized data into new point cloud data by using a radius outlier removal (ROR) algorithm, a convex hull algorithm and a least square method according to the obtained point cloud initial data, and finally reconstructing a high-precision three-dimensional geometric model. According to the method, the precision of the point cloud reconstruction model is improved, and the applicable scene is expanded.

The invention provides an IGBT (Insulated Gate Bipolar Translator) turn-off time measuring method, system, equipment and medium, and belongs to the field of power semiconductor device testing technology and state monitoring, and the method comprises the following steps: firstly, selecting 20% Vdc and 80% Vdc as turn-off initial measuring points, secondly, carrying out first error correction by adopting a linear estimation method, and finally, carrying out second error correction by adopting a linear estimation method; performing linear fitting on the voltage difference Vce data through a least square method to obtain initial turn-off time; and finally, aiming at systematic deviation caused by inherent concavity and convexity characteristics of the voltage curve, performing secondary error correction by adopting a convex hull algorithm, acquiring a lower envelope line of the ic-tOFF relation curve through an Andriw algorithm, and identifying and eliminating the problem of Overestimation of the turn-off time based on a convex hull accurate point set. According to the invention, accurate measurement of the turn-off time of the IGBT under the condition of low sampling rate is realized, the hardware cost is significantly reduced, and a reliable technical basis is provided for IGBT state monitoring and life prediction.

The invention relates to the technical field of fingertip touch, and discloses a fingertip extraction touch method and system based on a convex hull algorithm, and the method comprises the steps: collecting an original hand image, and extracting a hand layered sparse point set; performing feature space mapping on the hand layered sparse point set to obtain a hand feature enhanced point set; generating a hand contour convex hull according to the hand feature enhancement point set; performing point angle analysis on the hand contour convex hull to obtain a hand fingertip candidate point set; performing inter-finger relation constraint analysis on the hand fingertip candidate point set to obtain a hand fingertip target point set; and executing critical angular velocity touch control judgment on the hand fingertip target point set to generate a touch control trigger signal. According to the method, the integrity and continuity of a convex hull result are ensured, the problem of incoherence caused by region segmentation is avoided, and the stability in a rapid gesture conversion process is enhanced.

The invention relates to the technical field of froth flotation, in particular to a concentrate froth image recognition and concentrate grade qualification rate prediction method, which is characterized in that a block adaptive threshold segmentation method is adopted to distinguish image foreground and background information, and an image expansion algorithm is adopted to obtain froth contour information; adopting a convex hull algorithm to process vector information of all concentrate foam contours; and according to a set particle threshold value, outputting and calculating the current concentrate grade and an image recognition result to predict the concentrate grade qualification rate. The method has the beneficial effects that in the aspect of feature value extraction of the foam image, the problem of uneven illumination caused by factors such as time and angle is solved by calculating the gradient magnitude and direction of the image, the edge can be accurately identified, the system performs block processing on the foam image, and each sub-module automatically sets an optimal threshold value according to the pixel information of the image, so that the accuracy of feature value extraction is improved. Accurate threshold segmentation of the foreground and background images of the image is realized, and the accuracy and robustness of foam recognition are improved.

The invention discloses a lightweight real-time mapping system based on cloud edge collaboration and a VLA-HDMap high-frequency interaction method, and aims to solve the problems of slow updating, high cost and low safety of HDMap. The model has the capabilities of real-time environment understanding, semantic analysis and end-to-end decision instruction output. The vehicle end can download the cloud HMap in real time, and also participates in crowdsourcing data acquisition and uploading. According to the local lightweight mapping method provided by the invention, a real-time circulation mechanism of'acquisition-application-correction 'is used, so that the HDMap is highly consistent with the actual environment. A cloud edge collaborative hierarchical parallel incremental update global HDMap is also provided, the essence is a time-space intersection multi-scale analysis model of minute-level traffic flow prediction and day / week-level map differentiation detection, a high-risk region of traffic flow peak + map change can be efficiently identified, and a convex hull algorithm is used to obtain a region, which needs to be locally updated, of the whole HDMap. According to the invention, a new-generation intelligent traffic system can be constructed, and the safety application of group intelligence, symbiotic economy and large-scale automatic driving can be promoted.

The invention relates to an optimal scheduling method for port system electricity-hydrogen energy resource allocation, and belongs to the field of resource optimal scheduling. The method comprises the following steps: establishing a mapping relation between a logistics operation organization mode and an energy resource demand, and forming a unified collaborative information model; based on the collaborative information model, a configuration decision evaluation model considering electrochemical energy storage and hydrogen energy resources is constructed and used for generating evaluation indexes corresponding to a resource configuration scheme under the condition that the resource configuration scheme and a logistics operation organization scheme are given; constructing a three-dimensional decision-making space taking the operation cost, the carbon emission and the average ship in-port time as dimensions, constructing an observation grid on a preset discrete value of a corresponding index, solving a boundary state point set point by point, and constructing an envelope surface of the three-dimensional decision-making space by using a three-dimensional convex hull algorithm; and screening a configuration scheme based on decision boundary characteristics of the three-dimensional decision space. According to the invention, the applicability of an electricity-hydrogen energy resource configuration result in a complex operation scene can be improved.

The invention provides a live broadcast automobile marketing flow fluctuation prediction method and system, and relates to the technical field of cross-domain fusion, and the method comprises the steps: synchronously collecting data in a live broadcast platform and cross-domain extension data, and generating a platform data set and a cross-domain data set; based on the cross-domain data set, constructing an association space of the social hot events and the vehicle type attributes through a mirror image algorithm, and when the event feature intensity of the association space exceeds a preset threshold value, generating an event influence feature value; based on the event influence characteristic value and the platform data set, generating a basic flow prediction curve according to a fixed period and selecting a key point set; and processing the key point set through a convex hull algorithm to generate a convex polygon region. According to the invention, through constructing a full-link closed-loop design of data acquisition-event association-traffic prediction-parameter optimization-promotion triggering-verbal skill adjustment, accurate control of live broadcast automobile marketing traffic fluctuation and improvement of the marketing effect are realized.

The invention relates to a diffracted guided wave-based metal cylinder crack detection method, which comprises the following steps of: axially arranging a sensor array on a metal cylinder in an array manner, and collecting signals of all sensing paths; the sensing path is a connecting line between each excitation sensor and each receiving sensor; preprocessing each received path signal, judging a direct wave and a diffracted wave, and constructing a damage index of each sensing path on the basis of the amplitude difference between the diffracted wave and the direct wave; according to the signal and the damage index of each sensing path, drawing a total damage path diagram in the whole detection range, counting intersection points of each damage path and calculating the damage index of the corresponding intersection point; an improved clustering algorithm is utilized to obtain effective damage intersection points and a weighted centroid, and a damage area of the metal cylinder is delimited based on the damage intersection points around the weighted centroid through a convex hull algorithm. Compared with the prior art, the method has the advantages that dependence on health reference data can be eliminated, the calculation complexity is effectively reduced, and the detection efficiency is improved.

The invention discloses an orbit foreign matter detection method and system, and relates to the technical field of orbit detection.The method comprises the steps that an unknown foreign matter model is built through Gaussian distribution and a convex hull algorithm, and a known foreign matter model is built in combination with prior data; extracting a flat area of the ballast bed, projecting the foreign body model into a depth map and generating a shadow, and superposing to obtain a synthetic image with a mark; training a segmentation network by using the synthetic image and a real image, and obtaining a potential foreign matter mask through thermodynamic diagram binaryzation, connected domain analysis and size screening during reasoning; and inherent facility false alarms are eliminated through a secondary classification network, a real foreign matter detection result is finally output, alarm is triggered according to a detected threshold value, and efficient and accurate detection is realized. According to the method, on the premise that real foreign matter samples are scarce and ballast bed scenes are complex, efficient, accurate and real-time ballastless track foreign matter detection is achieved, meanwhile, the false alarm rate is reduced, and the generalization ability of the model to unknown foreign matter is improved.

The invention discloses a power network dynamic security protection system and method based on multi-source data fusion, and relates to the technical field of power systems. According to the method, the original heterogeneous data set is generated by comprehensively collecting the multi-dimensional data of the power network, and the problems of data format difference and time asynchronization are solved through protocol conversion and timestamp calibration; the line bearing capacity is dynamically calculated by fusing multiple types of data, the equipment risk level is generated, and the abnormal current feature extraction precision is enhanced by combining a mirror image algorithm, so that the security situation assessment is more accurate and comprehensive; abnormal current fluctuation is used as a trigger point, network malicious features and environment abnormity records in the same area are associated through a convex hull algorithm, threat types are comprehensively judged, and a target is clarified for protection execution. Threatened areas are rapidly isolated, alarms are pushed, a standby power supply is started, important user load distribution is adjusted in combination with a convex hull optimization algorithm, the fault influence range is reduced, key power supply is guaranteed, and the emergency processing capacity and the continuous power supply reliability are improved.

The invention belongs to the technical field of tunnel engineering digital construction, and particularly relates to an unsupervised learning drilling and blasting method tunnel point cloud feature extraction and back break detection method. The method comprises the following steps: firstly, acquiring initial point cloud data of tunnel excavation by adopting three-dimensional laser scanning, and preprocessing by adopting point cloud thinning and registration; then, a convex hull algorithm is used for calculating an actual excavation axis, excavation sections with the thickness of 0.1 m are cut out according to the fixed interval of 10 m, a K-means clustering algorithm is used for conducting noise reduction processing on point cloud slices, and convex hull vertexes are screened to accurately extract actual excavation contour lines; and finally, the back break amount of each convex hull is calculated, and real-time quantitative analysis of the back break condition of the tunnel is achieved. According to the method, by fusing three-dimensional laser scanning, unsupervised learning and a convex hull algorithm, crossing of tunnel back break detection from discrete, lagging and manual detection to continuous, real-time and intelligent detection is achieved, and efficient, accurate and reliable digital technical support is provided for drilling and blasting method tunnel construction.

The invention provides a method and a system for determining operation nodes in an organoid culture process, and relates to the technical field of data processing, and the method comprises the following steps: carrying out interval division on a state track curved surface to obtain a plurality of physiological state intervals; performing convex hull algorithm calculation on the data point set in each physiological state interval, obtaining distribution boundary characteristics of the data points by constructing a minimum convex polygon, and obtaining morphological dynamics indexes based on the distribution boundary characteristics and change gradient calculation; performing normalization calibration processing on the monitoring data according to the morphodynamic indexes to obtain a calibrated growth state evaluation value; comparing the calibrated growth state evaluation value with a preset operation threshold value to obtain a comparison result; and determining an operation node of the organoid growth state based on a comparison result. According to the method, the operation node of the organoid culture can be accurately determined, and the controllability of the culture process and the stability of the growth state of the organoid are effectively improved.

The invention discloses a convex hull algorithm optimization and applicability analysis method based on rotation contraction. The method comprises the following steps: step 1, processing a point set by a rotation contraction preprocessing strategy; and step 2, convex hull calculation. According to the method, the filtering operation is executed, a large number of non-convex hull points in the point set are deleted, the convex hull is calculated based on the classical convex hull algorithm, the convex hull algorithm optimization and applicability analysis method based on rotation contraction is constructed, the computer memory can be saved when the convex hull is calculated based on the method, meanwhile, the convex hull calculation time can be shortened, and the calculation efficiency is improved. And technical support is provided for the fields of computational geometry, computer graphics, computer vision and the like.

This invention provides an automated testing method and system for GNSS receiver hardware modules, relating to the field of electronic equipment testing technology. The method includes: parsing a preprocessed signal dataset and extracting signal feature boundaries based on a convex hull algorithm, calculating key performance indicators such as signal acquisition time, tracking accuracy, bit error rate, and receiver sensitivity; comparing these key performance indicators with preset performance thresholds, automatically determining whether the GNSS receiver hardware module is qualified based on the comparison results, and generating a performance judgment result; and automatically generating a structured test report based on the performance judgment result, and outputting the report to a designated terminal or storage location. This invention achieves an efficient and fully automated solution for testing various key performance indicators of GNSS receiver hardware modules by simulating GNSS signal excitation, response signal acquisition, signal processing algorithms, and automated performance judgment criteria.

The present invention provides a method, system, terminal, and medium for aggregating multi-target REDD project areas, specifically relating to the field of forest carbon sequestration technology. This solution constructs a multidimensional weighted feature set for each coordinate point based on the collected coordinate points within several target REDD project areas and the geographic feature data of the target layer at each coordinate point, as well as preset spatial coordinate aggregation weights and preset geographic feature aggregation weights. A clustering algorithm is used to cluster all the multidimensional weighted feature sets to obtain grid point sets corresponding to the multidimensional weighted feature sets. A convex hull algorithm is used to calculate the convex hull of the coordinate points in each grid point set to obtain the convex hull set of the grid point sets. Based on the target REDD project areas corresponding to all convex hull sets, an aggregated region for all target REDD project areas is obtained. This solution can effectively improve the accuracy of REDD project area boundary identification, thereby facilitating the comprehensive management and assessment of dispersed areas.

The invention discloses a new energy high-permeability power distribution network power balance method, and the method comprises the steps: firstly constructing a seasonal and moment-based wind and light confidence output probability model based on historical data; secondly, by utilizing the model, selecting new energy output values under corresponding confidence coefficients for different typical day scenes to carry out power balance analysis so as to determine a power grid construction scale; then, a fast convex hull algorithm is adopted, decoupling and enumeration regrouping are carried out on the time sequence data of the load and the new energy output, and an extreme operation mode curve representing the power shortage risk and the consumption challenge is systematically extracted; and finally, by taking the extreme curve as a security boundary, comprehensively considering the coordinated regulation capability of the source, the network, the load and the storage, and performing security verification on the power grid in the planning year. According to the method, confidence modeling and extreme scene extraction are combined, the defect that a traditional method is too conservative or insufficient in scene coverage is overcome, safe and stable operation of a power grid can be guaranteed, and meanwhile the economical efficiency of power grid investment is remarkably improved.

The present application relates to photovoltaic planning technical field, especially to a kind of solar radiation resource evaluation method for photovoltaic power station.The method is first according to installed capacity target to determine ideal state laying area, and combined with the building coverage of target area and shading loss is modified, and delimitation redundant target area.Subsequently, candidate point is screened in area, using the energy efficiency evaluation index of expected radiation resource and three-dimensional topological grid line cost, point is divided into different spatial clustering cluster by clustering algorithm.Aiming at each cluster, the minimum circumscribed polygon area is fitted using the convex hull algorithm, and the site selection area set is formed.Further, the present application simulates dynamic shadow by digital elevation model and ray tracing algorithm, and calculates the actual radiation resource amount of each area;if the total resource amount does not meet the target, a multi-objective function is introduced for Pareto optimization iteration, and a matching solution is selected from the optimal solution set and the polygon boundary is adaptively adjusted until the demand is met.

The present application belongs to the technical field of tunnel engineering digitized construction, and particularly relates to a method for extracting point cloud features and detecting overbreak and underbreak of a tunnel by drilling and blasting method without supervised learning. Firstly, three-dimensional laser scanning is used to obtain initial point cloud data of tunnel excavation, and point cloud thinning and registration are used for pretreatment. Then, a convex hull algorithm is used to calculate the actual excavation axis, and an excavation section with a thickness of 0.1m is intercepted at a fixed interval of 10m. The K-means clustering algorithm is used for noise reduction processing of the point cloud slices, and the convex hull vertices are screened to accurately extract the actual excavation contour line. Finally, the overbreak and underbreak of each convex hull is calculated to realize real-time quantitative analysis of the overbreak and underbreak of the tunnel. By fusing three-dimensional laser scanning, unsupervised learning and the convex hull algorithm, the present application realizes the leap of overbreak and underbreak detection of a tunnel from "discrete, lagging and manual" to "continuous, real-time and intelligent", and provides efficient, accurate and reliable digital technical support for the drilling and blasting method tunnel construction.

The application discloses a point-based shape contour recognition method, which comprises the following steps: constructing a standard pattern library, wherein the standard pattern library comprises regular patterns or custom patterns, and the picture size in the standard pattern library is recorded as m*n; obtaining point data of each shape in a to-be-recognized drawing; adopting a clustering algorithm to divide a group of points into different clusters in space, applying a convex hull algorithm to each cluster to obtain a convex polygon wrapping all points in the cluster; compressing the convex polygon to the size of m*n to form a standard convex polygon, and recording a compression rate; comparing the standard convex polygon with each pattern in the standard pattern library by IoU to obtain a pattern in the standard pattern library with the highest score as a predicted contour of the convex polygon; and then reflecting the original size according to the compression rate based on the predicted shape to obtain a result, so that the shape contour recognition in the drawing is realized. The method can be applied to recognition and analysis of various shapes, and has the characteristics of high efficiency, stability and wide applicability.

The application discloses a three-dimensional aggregate modeling method based on surface subdivision and step-by-step reshaping, and belongs to the field of granular material modeling simulation, and specifically comprises the following steps: a basic convex aggregate is established in a unit cube by using a convex hull algorithm; new vertices are generated at the midpoints of each edge of each triangular facet of the aggregate, so that the aggregate surface is subdivided; principal component analysis is used to determine the aggregate particle size and establish a local coordinate system; a displacement coefficient of the aggregate vertex is determined, which is specifically composed of an upper limit value of the displacement coefficient, a vertex scope factor and a vertex displacement floating coefficient, and is used for respectively controlling the overall intensity of aggregate shape reshaping, the moving range of densely distributed vertices and the difference in the moving range of the vertices; the vertices are moved according to the displacement coefficient in the local coordinate system, so that the shape of the reshaped aggregate is parameterized; and through repeated execution of the above shape reshaping process based on surface subdivision, an aggregate model with increasing fidelity is obtained.

The invention provides a single wood volume nondestructive measurement method based on a foundation laser radar, belongs to the field of single wood volume, and focuses on feature extraction of trunk structures in different forms and calculation of the trunk volume. According to the method, masson pine is taken as a research object, 52 individual trees are taken as samples, true values of different height diameters are acquired, the volume of the individual trees is calculated, and meanwhile, a high-density three-dimensional laser point cloud is scanned by using a foundation laser radar system. The method comprises the following steps: firstly, sampling original point cloud data, constraining growth of seed points by using normal vectors and angles, and accurately extracting trunk point clouds; simulating interception of a real disc through point cloud slices, and calculating a single-layer point cloud area by using an improved convex hull algorithm; in the shielded region, constructing a fitting circle by using the local point cloud to calculate the area of the layer; and finally, the single wood volume is obtained through the thought of an infinitesimal method. The correlation R2 between the calculated wood volume and the real calculation reaches 0.995, the RMSE is 0.017 m < 3 >, and the nRMSE is 2.38%, which indicates that the tree trunk can be accurately extracted and the single wood volume can be calculated under the non-contact condition.

The present invention relates to a method for detecting the outline of an explosion fireball, and specifically to a method for detecting the outline of an explosion fireball based on a convex hull algorithm, which solves the technical problem that existing methods for extracting the outline of an explosion fireball cannot obtain the complete outline of a target area of an explosion fireball. The present invention segments the original image of the explosion fireball directly or after grayscale conversion to obtain a binary image of the explosion fireball image, uses morphological operations to denoise the binary image, extracts the edges of the explosion fireball sub-regions, obtains a set of boundary coordinates of the explosion fireball sub-regions, and uses a convex hull algorithm to solve the minimum circumscribed polygon in the set of boundary coordinates of the explosion fireball sub-regions to obtain the complete outline of the explosion fireball. The explosion fireball outline detection method provided by the present invention has accurate detection results and high image processing efficiency, and can be applied to the dynamic analysis of explosion fireballs in videos.

The invention belongs to the technical field of data identification, and discloses a building classification method and device, equipment and a storage medium. The method comprises the following steps: preprocessing original building point cloud data to obtain a reference building point cloud; performing hierarchical clustering on the reference building point clouds to obtain a plurality of initial clusters; convex hulls are constructed for the initial clustering clusters, convex hull features are extracted, clustering cluster optimization is carried out according to the convex hull features of the initial clustering clusters, and reference clustering clusters are obtained; performing point cloud recognition according to the point cloud features and the convex hull features of the reference cluster to obtain a building category; according to the method, the building parts are preliminarily segmented through hierarchical clustering, the geometric features of each cluster are constructed in combination with a convex hull algorithm to optimize the initial clustering result of hierarchical clustering, the accuracy of clustering boundaries is improved, and based on joint recognition of the point cloud distribution features and the convex hull features of the optimized clusters, the recognition efficiency is improved. The building category identification result has overall shape reliability and local detail authenticity, and the accuracy of building point cloud classification is effectively improved.

The present invention discloses a three-dimensional thin-walled parts layout algorithm for unilateral infinite extension raw material plates based on reinforcement learning. The present invention comprises: using a three-dimensional projection algorithm, a minimum convex hull algorithm and a Minkowski vector sum algorithm to generate a two-dimensional expansion polygon; the polygon is encoded as a generalized graph, the selected graph is an action, the current layout is encoded as a generalized graph as a state, the placement area and compactness are used as action rewards, and deep reinforcement learning is used to iteratively obtain an action sequence, and finally the placement order, angle and position of each polygon are obtained. The present invention can solve the layout problem of thin-walled three-dimensional feature parts in mechanical processing; reinforcement learning fully explores the unknown states of the placement process, fully considers the influence of the parts arrangement order on the final result, and can quickly give a better layout method; in addition, the use of deep learning to fit the reinforcement learning value function has a strong generalization ability, and for a completely new layout problem, better layout results can be obtained without retraining the network parameters.

The present application relates to the field of image recognition technology, and specifically to a face recognition method, device, and equipment applied to an enterprise management system. The method comprises: obtaining a face image to be detected; obtaining non-skin pixels based on the color features of pixels; dividing the image into blocks and obtaining the degree of face occlusion by combining a convex hull algorithm; extracting connected domains of non-skin pixels within all convex hulls in the grayscale image of the face to be detected; obtaining the degree of suspicion of facial features based on the gradient features and color features of the edge pixels of each connected domain within each convex hull; obtaining the degree of information carrying based on the distribution of the degree of suspicion of facial features in the connected domains where the pixels in the pixel block are located; and obtaining the result of face recognition by combining the LBP coded histograms of the pixel blocks of the grayscale image of the face to be detected and the standard grayscale image of the face. The present application aims to extract the amount of feature information carried by different regions of the face and improve the accuracy of face recognition based on the LBPH algorithm.