9 results about "Point density" patented technology

The invention relates to a 3D point cloud-based concrete crack early recognition method. The method comprises the following steps of modeling a fine curvature change generated by a concrete surface at the initial stage of a crack into a high-order local nonlinear offset function; curvature derivative tensor fitting is carried out by using the third-order neighborhood of the point so as to distinguish a real crack sign and noise disturbance; a space stress projection field of the candidate area is formed in the crack germination stage, so that early deformation cracks are identified; based on a stress field guided point cloud reconstruction mechanism, enhancing the point density of the potential crack region identified by the perturbation model by adopting a local reversible interpolation algorithm; according to the interpolation, a crack space consistency constraint graph is constructed according to the continuity and curvature gradient of a surrounding normal, and the structural integrity is improved while the original crack form is reserved; and constructing a crack region into a vector flow field form, and fitting a growth path prediction trajectory of the crack region through the minimum principal curvature direction of crack end points in the point cloud.

The invention discloses a topology enhancement prototype and geometric self-adaption few-sample point cloud semantic segmentation method, which belongs to the technical field of three-dimensional vision, and comprises the following steps of: extracting geometric features of point cloud from a support set and a query set by using a feature encoder; a topology guiding prototype optimization module is introduced, persistent coherence analysis is carried out on point clouds of each category in a support set, multi-scale topology features are extracted, topology descriptors are generated, and the topology descriptors are used for guiding structural adjustment of prototype representation; a difficult sample correction module is designed, a difficult sample area in a query set is identified based on a topology instability index, and fine correction is performed on geometric features in combination with local point density and curvature information; and constructing a topology perception optimization module. According to the method, topology priori is systematically introduced into a few-sample point cloud segmentation task for the first time, and a more stable and fine semantic segmentation result is realized under the condition of not depending on a large amount of annotations.

The present application relates to the technical field of seismic data processing, and particularly relates to a seismic data sampling method, device, equipment, medium and product, wherein the method comprises: substituting shot point data of an initial bubble point set into a point density function to obtain a point density function value of a shot point; if the point density function value of the shot point is greater than a point density threshold value, deleting the shot point to obtain an intermediate shot point set; judging whether the point density function values of all shot points of the intermediate shot point set are all less than or equal to the point density threshold value; if the point density function values of all shot points of the intermediate shot point set are all less than or equal to the point density threshold value, taking the intermediate shot point set as a uniform shot point set, and randomly sampling shot points of the uniform shot point set; the uniformity of random sampling can be improved.

The present invention discloses a method for estimating the magnetotelluric vector space, a computer device, and a computer-readable storage medium. The method realizes the rapid indexing of 4D space points by constructing a 4D space index tree structure; then, according to the obtained magnetotelluric data block, it traverses the frequency points of the magnetotelluric apparent resistivity data of the magnetotelluric data block, obtains each frequency point, and adds the 4D vector data of each obtained frequency point to the vector space in the order of four dimensions A, B, C, and D; then, it performs point density space filtering on the scattered points in the 4D space, and searches for the maximum value of the point density after filtering as the central estimated value. The algorithm is simple and efficient, and has a small deviation after 4D space estimation.

It is provided a method for encoding a point cloud within a volume, the method being performed in an encoder. The method comprises: recursively splitting the volume into sub-volumes until each sub-volume satisfies a completion criterion, in which case the sub-volume is an end sub-volume; estimating, for each end sub-volume comprising points, its density of points, by selecting a density value, from a set of pre-selected density values, that is closest to the density of points of the end sub-volume; encoding the density of the sub-volumes by encoding a general mode density value, being the most common density value of the sub-volumes, and encoding density values of any sub-volumes comprising points, where the density value of the sub-volume differs from the general mode density value; and providing indications of the sub-volumes and the encoded density values. Corresponding embodiments for the decoder side are also provided.

The invention provides a high-risk operation real-time identification and millisecond-level alarm method for an operation and maintenance session, and the method comprises the steps: calculating the angle deviation between vectors in a hidden malicious behavior segment through employing cosine similarity, adjusting the center coordinate and radius expansion of a hypersphere according to an angle deviation sequence, fusing the deviation trend and point density distribution, and obtaining a high-risk operation real-time identification and millisecond-level alarm. Obtaining an expanded hyper-spherical boundary parameter; reversely adjusting the center coordinate adjustment amount and the boundary adaptive expansion amplitude of the high-dimensional curved surface space through the expanded hyper-spherical boundary parameters, and updating curvature local deformation parameters to obtain a corrected curved surface boundary model; and updating the state of the operation and maintenance session through an alarm response mechanism, obtaining an alarm confirmation mark and a response state, and adjusting the session access permission level according to the alarm confirmation mark to obtain updated session context data.

The present disclosure is a kind of based on depth datum unified variable resolution seabed terrain grid generation method and device, the method comprises: data preprocessing;Point density and slope calculation;According to the point density, the function relationship between the point density and the slope correction amount is established, and the slope is corrected;Based on the inverse distance weighted method, the corrected slope is smoothed;According to the slope threshold, the point cloud is filtered;Variable resolution triangular irregular grid is constructed.This embodiment effectively eliminates redundancy and noise by filtering repeated points of multi-source point cloud, ensures the accuracy and consistency of input data;Combined with point density and slope information, the terrain features are corrected and optimized, so that the terrain model can adaptively reflect the detail difference of different regions;Adaptive threshold screening improves the retention rate of key terrain features;Variable resolution TIN construction effectively reduces the overall data volume while maintaining important terrain details, balances modeling efficiency and storage cost.

The invention provides a meshless device structure optimization method and equipment based on random features, which can be applied to the technical field of computational mathematics and topological optimization. The method comprises the following steps: performing iterative structure optimization on a device by utilizing a displacement field network and a density field network to obtain a device structure optimization result; wherein the step of executing the primary structure optimization comprises the following steps of: adjusting parameters of a previous displacement field network by utilizing previous point density parameters of a plurality of device points of the device to obtain a current displacement field network; inputting respective point coordinates of a plurality of device points of the device into the current displacement field network to obtain respective current point displacement of the plurality of device points; performing parameter adjustment on the previous density field network according to the respective point coordinates of the plurality of device points and the current point displacement to obtain a current density field network; and inputting the respective point coordinates of the plurality of device points into the current density field network to obtain respective current point density parameters of the plurality of device points.

The invention discloses a point cloud registration method adaptive to point density change. The method comprises the following steps: step 1, collecting point cloud data; step 2, point feature extraction; 3, removing point pairs based on the relative height of the points; 4, carrying out point filtering and fine registration based on probability distribution; according to the invention, large-scale point cloud registration adapting to point density change can be realized in a large-range outdoor environment with communication bandwidth change.