25 results about "Point cloud" patented technology

The invention provides a full-automatic three-dimension characteristic extracting method. The method comprises the steps of firstly converting the three-dimension point clouds of all single view fields obtained by a stereoscopic vision measuring system into depth images of a single view field; splicing the point clouds of the single view fields; carrying out automatic searching and extracting on the corresponding depth images according to the known positioning characteristic of a workpiece to obtain the positioning characteristic of the workpiece, thus automatically converting the spliced three-dimension point clouds into the workpiece coordinate system; subsequently inputting the ideal coordinates of all characteristics to be detected of the workpiece in the workpiece design model; dividing the characteristics into a plurality of types such as aperture, tendon, plane and the like; automatically extracting the depth images respectively and obtaining the two-dimension coordinates of depth images; searching the three-dimension points of two-dimension depth image coordinates under the corresponding workpiece coordinate system; and obtaining the three-dimension information of each characteristic by calculation. The method has the characteristics of full automation, good flexibility and good expandability, and can be used for automatically extracting three-dimension characteristicsof various workpieces with known positioning characteristics.

The invention discloses an obstacle feature extraction method based on a millimeter wave radar and a laser radar. The method includes the following steps that: step 1, radar raw data are acquired; step 2, the radar original data are solved according to a radar protocol, so that obstacle position information can be obtained; step 3, the solved millimeter wave radar data are screened; step 4, time and space synchronization is performed on the millimeter wave radar and the laser radar; step 5, regions of interest (ROI) are established in a laser radar coordinate system according to millimeter wave radar information and are denoted as Ri, wherein i is equal to 1, 2, ... n, wherein n is the number of the effective targets of the millimeter wave radar; step 6, the point cloud data of the i-th ROI are processed according to a scattered point cloud data simplification processing method; and step 7, a two-dimensional image is generated on the basis of the point cloud data, edge detection is performed on the image, and obstacle features are extracted. With the method adopted, obstacle edge information can be obtained, and therefore, more references can be provided for a driverless decision-making layer.

The invention discloses a dangerous rock body landslide monitoring method based on characteristics and relates to the dangerous rock body landslide technology field. The method comprises steps that a local curved surface fitting method is employed, and a surface normal of reference point cloud data is determined; a point of the reference point cloud data is taken as a circle center of a circular column bottom circle, the surface normal direction of the reference point cloud data is taken as an axis of a circular column, and the circular column is determined according to a radius and axis length; comparison point cloud data of the circular column is projected to the axis of the circular column, through calculating the average position of projection points, a gravity center of the comparison point cloud data in the circular column is determined; through calculating a distance between the point of the reference point cloud data and the gravity point of the comparison point cloud data in the circular column, deformation quantity of the comparison point cloud data relative to the reference point cloud data is determined, and thereby a dangerous rock body landslide motion trend is determined. The method is advantaged in that a rapid comparison algorithm based on point cloud is proposed, environment restrictions can be prevented, three-dimensional deformation can be acquired, an algorithm structure is simple and convenient, and data processing efficiency is high.

The invention relates to a cloud point filtering method and device, computer equipment and a storage medium. The method comprises that cloud point data is obtained; the cloud point data includes a setof data points obtained from the road scanning environment; a neighborhood of a present data point in the point cloud data is determined; when the counted first number of data points in the neighborhood of the present data point is lower than a preset threshold, the present data point is determined to be a candidate noise point; the candidate noise point as a seed point carries out regional growth to obtain a target growth region; when the counted second number of data points in the target growth region is lower than a preset threshold, the candidate noise point as a noise point is filtered from the point cloud data. The noise point is located preliminarily, and then determined by regional growth and filtered, data points formed by objects in the far place and sparse data points in the point cloud data are protected, and the accuracy of point cloud filtering is improved.

The invention relates to the technical field of detection and provides a ground detection method, a related device and a computer readable storage medium. According to the invention, the method comprises the steps of acquiring a depth map and an attitude angle of a camera; building three-dimensional point cloud in the world coordinate system according to the depth map and the attitude angle of thecamera; acquiring an initial ground area according to the three-dimensional point cloud in the world coordinate system, calculating the dip angle of the initial ground area and determining a ground detection result according to the dip angle. Three-dimensional point cloud in the world coordinate system is built according to the acquired depth map and the attitude angle of the camera and the ground detection is performed according to the three-dimensional point cloud in the world coordinate system, so that there is no need to limit the position and attitude of sensors and the method has universality.

A method for extracting parameters of a cutting tool is provided. The method comprises positioning the cutting tool on a moveable stage, performing one or more rotary scans of a first section of the cutting tool to generate a scanning point cloud, indexing a plurality of points of the scanning point cloud, detecting one or more feature points based on the indexed scanning point cloud, and extracting one or more parameters based on the detected feature points. A system for extracting the parameters is also presented.

The invention relates to a transformer substation intelligent safety monitoring system fusing multi-source safety state heterogeneous data and a safety management and control method. The transformer substation intelligent safety monitoring system comprises a three-dimensional laser point cloud model module, a work ticket management module, an intelligent access control management module, an electronic fence module, an intelligent violation identification module and an early warning management module. The invention discloses a field operation safety management and control method based on a transformer substation intelligent safety monitoring system. The method comprises the following steps: firstly, acquiring heterogeneous data such as personnel position data, field operation video monitoring, two-ticket text information, three-dimensional point cloud and the like by using a Beidou positioning technology, a video monitoring system, laser scanning and the like; then extracting text information of the two tickets by using a natural language processing algorithm based on RPNN, obtaining position information of an operation specified area, and realizing generation of a safety area based on the text information of the two tickets; combining the obtained three-dimensional point cloud data of the production site, and realizing analysis and alarm of an operation safety area through a Beidou positioning technology and a vector criterion; and finally, through a face recognition algorithm based on MTCNN, a target detection algorithm based on RefineNet and a semantic segmentation algorithm based on MaskRCNN, realizing risk recognition of typical violation behaviors of personnel. The system greatly reduces the safety risk of operators, saves human resources, reduces the operation and maintenance cost, and is of great significance for improving the intelligent management level of the transformer substation.

The invention discloses a method for obtaining a tunnel excavation face surrounding rock point cloud model. The method comprises the following steps: (1) selecting camera parameters according to the tunnel environment; (2) selecting shooting time according to a tunnel construction procedure; (3) selecting a shooting mode, a shooting position and a mark point position according to the excavation length; (4) selecting shooting parameters according to the environment of the tunnel excavation section; and (5) acquiring a surrounding rock image set after tunnel excavation, and processing the imageset to obtain a three-dimensional point cloud model of the surrounding rock after tunnel excavation. Compared with the prior art, the method has the advantages of being rich in collected data size, low in cost, rapid to operate, high in flexibility and the like, a geological engineer can directly shoot and record when observing surrounding rock through geological sketch after excavation, the siteconstruction progress is not affected, shooting personnel can flexibly change the shooting angle and position, and the problem that rock outcrop is mutually shielded is solved as much as possible.

The invention provides a face symmetry plane obtaining method based on iterative mirroring and registration, and belongs to the field of computer vision. The face symmetry plane obtaining method comprises the following steps that (1) face data are acquired through 3D scanning, and a point cloud is smoothly scanned by a Gaussian filter; (2) the longitudinal axis is roughly estimated through PCA, the face is axially segmented to acquire slice contour, and the most symmetrical points are determined and the symmetry plane is solved; and (3) a new point cloud is generated by the symmetry plane point cloud, the point cloud and mirroring data are registered through ICP and registration iteration is performed and the symmetry plane is solved. The novel face symmetry plane obtaining method is put forward on the basis of the iterative mirroring and registration method so that the influence of the local asymmetry of the symmetrical object and the asymmetry caused in the scanning process can be effectively avoided and the obtained symmetry plane is enabled to be more accurate.

The invention discloses a point cloud acquisition method, device and system and a storage medium. A six-axis robot carries a single line laser three-dimensional camera to move back and forth on the sliding chute, different rows of the whole scanning area are scanned, large-view-field panoramic scanning is realized; structure is very simple, correspondingly, installation and control are very simple, simplification of the structure and control is achieved, and an overlapped area exists between every two adjacent scans, so that after rough splicing is conducted, the rotation matrix and the translation matrix can be calculated only based on the point cloud of the overlapped area, the number of the point clouds is greatly reduced, the fine splicing time is shortened, and the splicing precisionis improved.

The invention discloses a virtual assembly method and device, equipment and a medium. The method comprises the following steps: determining cloud data of a to-be-registered and a reference point according to collected original point cloud data of a part to be assembled; determining a registration weight between the to-be-registered point and an associated reference point according to the allowableerror of the to-be-assembled part; and according to the registration weight, performing point cloud registration on the to-be-registered point cloud data and the reference point cloud data to realizevirtual assembly of the to-be-assembled part. According to the technical scheme of the embodiment, the assembly error requirements of different assembly parts in the assembly process are met, and theprecision of virtual assembly is improved.

The invention relates to a gear failure part feature reconstruction method based on RE and belongs to the technical field of reverse engineering and the field of gear remanufacturing. The method comprises the following steps of 1, carrying out linear cutting treatment on a failure gear to remove a residual stress layer of the failure gear, and coloring the failure gear; 2, scanning surface characteristics of the failure gear to obtain point cloud data; 3, abnormal point removal, noise reduction, packaging, simplification and hole filling are carried out on the point cloud data on the surface of the failure gear; 4, reconstructing the failure gear based on a reverse engineering principle; and 5, analyzing the precision of the reconstruction model according to the adjacent teeth of the defect part. According to the method, a precision problem of failure part model reconstruction in gear remanufacturing at the present stage is solved, an effective method is provided for feature reconstruction of the gear failure part, a problem that the precision of the gear failure part model reconstruction is low is solved, and a foundation is laid for improving the gear remanufacturing and repairing quality.

The invention discloses a laser radar calibration method and apparatus, an electronic device and a storage medium. The method comprises the steps of obtaining point cloud data of a reference laser radar and a to-be-calibrated laser radar for the same scene; matching the point cloud data acquired by the reference laser radar and the to-be-calibrated laser radar to obtain coordinates of the reference laser radar in a to-be-calibrated laser radar coordinate system; according to the coordinate of the reference laser radar in the coordinate system of the to-be-calibrated laser radar, the absolute coordinate of the reference laser radar in the world coordinate system, the initial installation parameter of the to-be-calibrated laser radar relative to the inertial measurement unit, and the pose of the inertial measurement unit in the world coordinate system when the to-be-calibrated laser radar obtains scene point cloud data; solving optimized installation parameters of the to-be-calibrated laser radar relative to the inertial measurement unit; and calibrating the to-be-calibrated laser radar. According to the method, the rapid calibration of the dynamic laser radar platform is realized under the condition of not depending on manual preset environment feature marks.

The embodiment of the invention discloses an obstacle judgment method, device and equipment and a storage medium, belongs to the technical field of vehicle driving, and can be applied to closed parks such as ports and logistics, or urban traffic or high-speed scenes and the like. The method comprises the steps that detection results of a laser radar and other target sensors on a driving space are fused in a deep learning mode so as to confirm an obstacle, in the implementation process, for laser point cloud obtained by the laser radar, whether the laser point cloud is a noisy point or not is judged from the point cloud level, and the detection accuracy of the laser point cloud is improved. The morphological characteristics of the driving space are determined from the global level, whether the point locations conform to the time sequence characteristics of the obstacles or not is judged, whether the sensing characteristics of the obstacles exist in the driving space or not is determined through a target sensor, and the driving space is determined by judging the dynamic characteristics of the driving space and at least one sensing characteristic. Whether there is an obstacle in the driving space is determined, and the accuracy of determining whether there is an obstacle in the driving space is improved.

The invention provides a workpiece allowance rapid solving method based on three-dimensional measurement point cloud data, and the method comprises the following steps: scanning a workpiece through ascanner, and obtaining the point cloud data; filtering and denoising the point cloud data, and solving a point cloud normal by adopting a kd-tree method; importing the three-dimensional point cloud data into a theoretical model of the workpiece, and completing registration of the filtered and denoised three-dimensional point cloud data and the theoretical model by adopting a method of combining NDT and ICP; carrying out space division on the point cloud data by using an octree data structure, constructing a hash table, and carrying out triangular patch coding on a divided space block to finishpoint cloud data preprocessing; performing dimensionality reduction on the preprocessed point cloud data, and screening out and removing triangular patches irrelevant to margins of points to be selected; applying the vector product and the cross product to the remaining triangular patches to quickly calculate the allowance and judge the effectiveness. The method has the beneficial effects that large-scale point clouds are encoded through space division of the octree and the hash table, so that the search efficiency is effectively improved, and the allowance of the workpiece is quickly solved.

The embodiment of the invention discloses an object recognition method and system, electronic equipment and a readable storage medium. The object recognition method comprises the steps that a plurality of depth camera devices acquire point cloud frames at different time points, and the point cloud frames comprise at least one object; screening out a target object from at least one object in the obtained point cloud frame of the specific time point through a specific camera device in the plurality of depth camera devices, and reading features of the target object; and based on the features of the target object identified by the specific camera device in the obtained point cloud frames at different time points, determining a depth camera device which identifies the target object again after the target object disappears from the point cloud frames obtained by the specific camera device, the features of the target object can be recognized under the condition of guaranteeing user privacy under the condition of pedestrian recognition.

The invention discloses a subway nut looseness detection method based on view angle conversion. The method comprises the following steps: 1) mapping an acquired nut depth map into a point cloud; 2) coloring the point cloud of the nut; 3) performing visual angle conversion according to the prior information of the nut and the camera position; 4) extracting a marking line in the nut image after view angle conversion; and 5) performing looseness detection according to the form of the marking line. The invention aims to solve the problem of detection caused by angle limitation when a patrol robot below a subway chassis shoots a nut image. Specifically, aiming at the particularity that the working environment of the metro chassis inspection robot is narrow and the expansion and contraction of a mechanical arm with a camera are limited, the problem of low recognition rate caused by a poor angle of a shot nut is solved. By adopting the method disclosed by the invention, the proportion of manual detection is reduced, the automation level is improved, and a safer, more reliable and more intelligent subway operation environment is constructed.

The invention provides a point cloud data processing method and apparatus, an electronic device and a storage medium. The method comprises the steps of S1, obtaining point cloud data of n objects of a target type; s2, for each point cloud data, inputting the point cloud data into an encoder for encoding to obtain a first implicit code vector corresponding to the point cloud data; s3, for each point cloud data, inputting the point cloud data and the first implicit code vector corresponding to the point cloud data into a reversible decoder, and performing a forward process of a target flow to obtain a target matrix; s4, calculating a loss value based on the Gaussian distribution matrix and the target matrix; and S5, if the loss value is greater than a preset loss value, optimizing the encoder and the reversible decoder, and repeating the steps S1 to S4 until the loss value is less than the preset loss value, so as to take the current reversible decoder as the trained reversible decoder. According to the invention, the point cloud reconstructed by the reversible point cloud decoder after training has richer details.