980results about "Details involving processing steps" patented technology

The invention relates to the technical field of rock mechanics, and discloses a rock three-dimensional reconstruction method and system based on multi-source information fusion, and the method comprises the steps: obtaining and preprocessing data, carrying out the spatial feature learning of a fusion feature vector through a 3D-CNN network, and constructing a three-dimensional voxel model of rock microscopic damage; converting the fused image data into a point cloud model of the underground cavern surrounding rock structure by adopting a three-dimensional reconstruction algorithm based on point cloud, and constructing a digital twin framework of the underground cavern surrounding rock structure based on an implicit surface reconstruction algorithm; feature parameters output by the three-dimensional voxel model and the digital twinning framework are used as input, and the optimal supporting opportunity and supporting parameters are output through an LSTM-CNN fusion model; in the underground engineering construction process, surrounding rock deformation data are collected in real time, and a supporting scheme is adjusted in real time through a depth deterministic strategy gradient algorithm; according to the method, the scientificity and timeliness of support design under complex geological conditions can be improved.

The invention discloses an AR home experience method in a large scene. On the basis of combination of a natural feature identification-based three-dimensional registration method and a binocular tracking positioning and local mapping method, the camera attitude is estimated by using feature points of a real-time scene and corresponding three-dimensional points thereof under the binocular trackingpositioning and local map construction technology. According to the mode, on-site environment features shot in real time are used as recognition tracking objects, a virtual home model can still be normally positioned and tracked under the condition that no identification graph exists, the problems that an existing AR home experience application is small in use range and poor in stability are solved, and therefore the AR home experience of virtual and real fusion can be met in a wider range and more truly.

Systems and methods for four-dimensional analysis of a patient's coronary arteries and myocardial wall. An example method includes accessing angiographic x-ray images of one or more arteries, with the x-ray images depicting the arteries from respective angles, with each x-ray image being associated with a cardiac phase of cardiac phases, and with the cardiac phases being included in a cardiac cycle. A three-dimensional model of the arteries is generated for each cardiac phase based on a subset of the x-ray images which are associated with the cardiac phase. A four-dimensional representation of the arteries is generated throughout the cardiac cycle based on the three-dimensional models for the cardiac phase. Information associated with the four-dimensional presentation is presented.

The invention relates to the technical field of digital surveying and mapping of constructional engineering and construction quality control, and discloses a house surveying and mapping system for constructional engineering, which comprises a data acquisition module for synchronously generating a panoramic image signal and an inertial measurement unit signal; the preprocessing module receives a panoramic image signal and an inertial measurement unit signal, and fuses the panoramic image signal and the inertial measurement unit signal through a visual inertial odometer to generate a pose calibration signal; the fusion modeling module receives the purification point cloud signal and the pose calibration signal, and generates building component parameterized data based on a deep learning component recognition algorithm; the result output module receives the building component parameterized data and generates a construction deviation analysis report based on a preset tolerance threshold value; the intelligent control module generates an acquisition quality feedback signal through a real-time quality evaluation algorithm to the data acquisition module so as to adjust a scanning path, and sends a dynamic sampling rate control signal to the preprocessing module so as to optimize a calculation load. The problems of low house surveying and mapping efficiency, poor modeling precision and weak field real-time feedback are solved.

The invention discloses an intelligent park management system based on cloud computing and Internet of Things, and relates to the technical field of park management systems, and the system comprises a dynamic digital twinborn body module which comprises a heterogeneous data fusion layer which integrates a BIM model, LiDAR point cloud, an unmanned aerial vehicle inspection image and distributed loT sensor data, and eliminating timestamp deviation of multi-source data by adopting a space-time alignment algorithm, and generating a millimeter-level precision three-dimensional park twin model. Through the construction of the high-precision digital twinborn model, the panorama and equipment operation state of the park can be accurately presented in real time, the timestamp deviation and spatial registration error of multi-source data are eliminated, the real-time performance and accuracy of the data are improved, a solid data basis is provided for subsequent intelligent simulation and early warning decision, and the response and monitoring capabilities of the system are improved.

The invention provides an AR-based bridge crack detection analysis method and system, and relates to the technical field of bridge detection analysis, and the technical scheme is characterized in that the method comprises the steps: monitoring the ambient light parameters of a bridge detection region; adjusting illumination output to stabilize imaging illumination conditions of the crack region; acquiring a crack image, and acquiring polarization information of the crack; stokes parameters are calculated, and a polarization degree image and a polarization angle image are generated; extracting a crack two-dimensional contour based on the polarization degree image and the polarization angle image; fusing the two-dimensional contour of the crack with the depth data to generate a three-dimensional model of the crack; and based on the crack three-dimensional model, crack geometric parameter quantitative analysis and multi-stage evolution state evaluation are carried out. The AR-based bridge crack detection and analysis method and system provided by the invention have the advantages of overcoming the influence of outdoor complex illumination conditions on image acquisition quality, improving the accuracy of crack recognition and three-dimensional modeling, and supporting reliable multi-stage evolution state evaluation.

In various examples, a framework is or provides an end-to-end solution that includes multi-sensor capture, data processing, inferencing, synchronization, alignment, and 3D rendering for multi-modal perception fusion pipelines. A multi-modal perception fusion pipeline may include a mixer, an aligner, an inference environment, and a multi-view renderer. The mixer may merge sensor data from different data sources into a single HashMap frame. The aligner may use calibration data for sensor-to-sensor coordinate transformations. The inference environment may receive multi-modality data and use custom preprocessing and custom postprocessing to generate inference results. The renderer may generate different sensor data renderings. The framework may include an application that uses configuration data to generate or configure a custom multi-modal perception fusion pipeline. The inference environment may access inference models using a uniform inference interface and support remote inference, allowing the pipeline to become an API client of the inference models.

The invention relates to the technical field of ruin three-dimensional construction, in particular to a ruin three-dimensional model construction method fusing SLAM point cloud and image features, which comprises the following steps: acquiring a depth image and RGB data, aligning and eliminating inconsistent data according to a timestamp, calculating a frame rate to node ratio, establishing a binding relationship between an image and a point cloud, and extracting three-dimensional coordinate points. Recognizing and classifying geometric structures, evaluating texture and edge response, adjusting resolution, analyzing static and dynamic structures, eliminating dynamic interference, calculating geometric curvature and fitting error, optimizing three-dimensional model display, and obtaining detail reduction degree scores. According to the method, through accurate synchronization of the point cloud data and the RGB image, strict verification of timestamp matching and data consistency is allowed, data errors and noise are significantly reduced, the accuracy and reliability of the model are ensured, structural elements are identified, dense areas of geometric structures are refined and analyzed, richer data level support is provided, and the method is suitable for large-scale popularization and application. And higher-quality data is provided for archaeological research.

The invention provides a crack detection method and system for a double-block sleeper, and relates to the technical field of rail transit disease detection. The method comprises the following steps: inputting a multi-view image into a pre-trained multi-view fusion model to generate an initial three-dimensional reconstruction model of the double-block sleeper; fusing the ultrasonic detection data, the infrared thermal imaging data and the initial three-dimensional reconstruction model to obtain a target three-dimensional reconstruction model; extracting crack characteristics of different scales from the target three-dimensional reconstruction model, and identifying and marking a suspected crack region based on the extracted crack characteristics of all scales; a deep learning model is adopted to detect real cracks in the suspected crack area; and under the condition that a plurality of real cracks exist, the crack distribution condition is analyzed, the damage degree of the double-block sleeper is evaluated, and a crack detection report of the double-block sleeper is generated according to the damage degree evaluation result. According to the technical scheme provided by the invention, the crack detection efficiency and effectiveness are improved.

The invention relates to the technical field of three-dimensional measurement, and discloses a double-zoom system and a three-dimensional information acquisition method thereof, and the three-dimensional information acquisition method comprises the steps: adjusting a projection unit and a first micromirror array, and enabling a projection end to work in a coaxial light mode; controlling the second micromirror array to zoom by adopting a preset second zoom parameter, and processing the two-dimensional image data of the imaging unit in real time to generate a two-dimensional ultra-depth-of-field image; adjusting the parameters of the second micromirror array until the image acquired by the imaging unit is clear; a preset first zooming parameter is adopted to control zooming of the first micromirror array, and scanning information and image data are obtained in real time and processed to obtain surface height information; generating three-dimensional information containing texture information in combination with the two-dimensional super field depth image and the surface height information; according to the method disclosed by the invention, the problem of high calculation error in the absence of 2D texture information in the traditional 3D imaging technology is solved, and the height information can be accurately acquired even in an area with scarce texture information.

The invention discloses a data processing method and system for a dental three-dimensional model, and relates to the technical field of dental three-dimensional modeling, and the method comprises the steps: obtaining initial data of oral cavity scanning, recognizing a tooth type in the initial data, and determining a corresponding tooth surface reflection parameter according to the tooth type; generating scanning parameter combinations for different tooth positions according to the tooth surface reflection parameters and the tooth anatomical structure; tooth point cloud data are collected, and a tooth enamel reflective area and a shadow shielding area are identified; performing light reflection removal processing on the enamel light reflection area, performing data completion on the shadow shielding area, and generating corrected tooth point cloud data; and performing registration fusion on the corrected tooth point cloud data to generate a dental three-dimensional model containing a complete dentition. According to the method, the problem of data missing caused by enamel light reflection and adjacent tooth shielding is solved through a differential scanning strategy and regional data processing aiming at specific optical characteristics of teeth, and automatic construction of a high-quality dental three-dimensional model is realized.

The invention provides a NeRF sparse view reconstruction method based on adaptive multi-modal feature fusion, and relates to the technical field of three-dimensional scene reconstruction. The method comprises the following steps: extracting a depth feature and a semantic feature from a sparse input view, and generating a first multi-modal feature through a feature enhancement network; a mixed attention mechanism is combined with channel and space attention to perform dual selective enhancement on the features to obtain second multi-modal features; the weight distribution network based on confidence guidance dynamically fuses the multi-modal features, harmonizes conflicts among the modals through a consistency loss function, and generates final fusion features; and inputting the fusion features into an enhanced neural radiation field rendering module, decomposing a scene into multi-scale representation by using layered feature coding, and optimizing geometry and appearance details in stages by combining a progressive training strategy to realize high-quality three-dimensional reconstruction. According to the method, the problems of geometric distortion, fuzzy texture and inconsistent modes caused by sparse views in a traditional method are solved.

The invention relates to the technical field of equipment management and control, in particular to an intelligent management and control method and system for a smart park, and proposes the following scheme: dividing a region into regular grids based on spatial data and equipment data, and generating an inspection path in combination with dynamic weights. And the inspection equipment collects equipment operation data and environmental parameters according to the path, and transmits the inspection data to the processing equipment in real time. And the processing equipment accurately identifies equipment operation abnormity, environment parameter abnormity or security risks through time sequence analysis and an abnormity detection model, and links the management and control equipment to process an abnormal area. The method improves the park inspection efficiency and the exception handling capability, and is suitable for industrial park scenes.

The invention relates to the technical field of three-dimensional reconstruction, and discloses a monocular depth guided object-level NeRF reconstruction method, which comprises the following steps: firstly, through monocular video sequence input, generating a frame-by-frame initial depth map by using a depth estimation module, and estimating a relative camera attitude between adjacent frames through a relative attitude estimation module; calculating the absolute attitude of the camera in combination with the initial depth map and the relative camera attitude; then constructing a geometrically enhanced NeRF model, and optimizing scene representation through multi-resolution hash position coding and spherical harmonic direction coding; introducing photometric loss, depth contrast loss and density loss to jointly optimize parameters of the NeRF model, and constraining geometric reconstruction of the object by using depth information; and finally, through a four-stage iterative training strategy, alternately optimizing depth estimation, a camera attitude and a NeRF model, and generating an object-level controllable three-dimensional model. According to the method, depth information is fully utilized to optimize the NeRF training process, so that end-to-end reconstruction from a monocular video to an object-level controllable model is completed.

Systems and techniques for inspecting vehicles. Some embodiments provide a system for vehicle inspection, including a support member having a first portion and a second portion angled relative to the first, and a plurality of sensor arrays coupled thereto. The plurality of sensor arrays may include respective sets of cameras oriented in multiple directions. The plurality of sensor arrays may be coupled to the support member at different positions. The plurality of sensor arrays may include: a first sensor array having one or more wheels of a vehicle in its field of view (FOV) when imaging the vehicle, a second sensor array having vehicle side in its FOV when imaging the vehicle, and a third sensor having a vehicle roof in its FOV when imaging the vehicle. The system may also include a processor configured to control the plurality of sensor arrays to capture images of the vehicle.

The invention relates to an occupancy prediction method, system and device for automatic driving and a medium. The method comprises the following steps: acquiring a multi-view image sequence, extracting multi-view image features through a two-dimensional convolutional network, converting the multi-view image features into three-dimensional voxel features, and projecting to generate bird's-eye view features at the current moment; inputting the aerial view features at the current moment into a scene memory gate module for dynamic fusion with a historical feature map, and generating an updated historical feature map; converting the updated historical feature map into a three-dimensional semantic feature, and fusing the three-dimensional semantic feature with a three-dimensional voxel feature to generate an enhanced three-dimensional feature; performing multi-scale geometric coding processing on the enhanced three-dimensional features to generate three-dimensional geometric features; performing multi-level semantic coding processing on the updated historical feature map to generate two-dimensional semantic features; and fusing the three-dimensional geometric features and the two-dimensional semantic features, and outputting a three-dimensional occupation prediction result. The prediction method can consider high precision, low parameter quantity and real-time reasoning under limited computing resources.

Methods, apparatuses and systems for extrinsic parameter optimization used in image-based pose estimation of a boom are disclosed. The method includes generating a 3D model of the boom, generating 3D keypoints using the generated 3D model of the object (206) and generating predicted 2D keypoints from the generated 3D keypoints and at least one initial extrinsic parameter. The value of the initial extrinsic parameter is optimized using an iterative process such that the variation between the predicted 2D keypoints and location of the boom is minimized.

The invention relates to the technical field of land layout, and discloses a BIM platform-based land reclamation project automatic rapid layout method and system, and the method comprises the steps: obtaining the site topographic data of an excavation region, obtaining an original landform BIM model of a field based on the data, carrying out the screening, enabling the screened original landform BIM model of the field not to comprise a super-slope region, and carrying out the screening of the original landform BIM model of the field; arranging a terrace according to the contour line to obtain a field piece line, performing grouping based on the contour line interval where the height of the point cloud data is located, calculating a field surface design elevation, and establishing a preliminary terrace model on the basis of the original landform BIM model of the field piece based on the field piece line, the gross width of the terrace and the field surface design elevation; and establishing a ridge and ridge model on the basis of the preliminary terrace model according to the field surface clear width, the ridge width and the ridge height. According to the invention, a convenient, rapid and intelligent field distribution scheme is provided for a land leveling construction project in the hilly area, so that fine improvement construction of the terrace type terraced field in the hilly area is realized.

The invention belongs to the field of CT (Computed Tomography) tomography reconstruction technology and artificial intelligence, and discloses a finite angle CT reconstruction method based on combination of a three-dimensional conditional diffusion model and synchronous iteration. The CT is an imaging technology which utilizes X-rays to irradiate a target from different angles and acquire projection, and obtains an internal three-dimensional structure through reconstruction. Different from traditional CT depending on nearly full-angle scanning, the method only collects limited-angle projection, and achieves fault reconstruction under the limited-angle condition through a three-dimensional condition diffusion model of space domain-frequency domain two-way decoding and by means of structural generality priori of a workpiece. Meanwhile, projection and fault data consistency correction is carried out in combination with a synchronous iteration reconstruction technology, the advantages of an iteration method in the aspect of physical mechanism characterization are exerted, interpretable physical constraints are provided for a deep learning network, and therefore the reliability and precision of a reconstruction result are improved.

The invention discloses a road scene space rapid investigation system combining RTK and image reconstruction, relates to the technical field of road scene space investigation, and is used for solving the problem of poor road mapping splicing effect. According to the method, an RTK track in road inspection is collected and synchronized with an image frame, after abnormal points are removed, continuous control points are extracted to construct a road topology skeleton, a space is divided into a plurality of mapping units based on the control points and direction vectors, redundant overlapping areas are set, and a mapping task package with structural continuity is generated; a road structure area in an image is extracted through semantic segmentation and edge detection, a structure mask is constructed and projected to a three-dimensional space, and point cloud density enhancement and reconstruction optimization are guided, so that block boundary structure anchor points are further identified, local rigid body registration is completed, and a global error minimization model is constructed. The posture consistency and the space continuous splicing of all the image blocks are realized, and the exploration efficiency of the road scene space is improved.

The invention provides a point cloud downsampling method applied to unmanned aerial vehicle obstacle avoidance, and aims to solve the problems of low mapping efficiency and geometric detail loss caused by large point cloud data volume and non-uniform density in environmental perception. The method comprises the following steps: preprocessing an original point cloud through a regular grid to balance density, and extracting an effective flight area of the unmanned aerial vehicle by adopting bidirectional cloth simulation filtering and combining with a flight height; multi-level geometric features including elevation gradient, curvature and roughness are calculated, and point cloud priority scores are fused and generated based on principal component analysis; voxel grids are divided according to the size of the unmanned aerial vehicle, the sampling rate of each voxel is adaptively set in combination with the priority, farthest point sampling is executed in the feature space to select representative points, and finally a simplified point cloud model is output. The method can be integrated to a mapping module of a multi-aircraft cooperative obstacle avoidance and planning system, is suitable for unmanned aerial vehicle path planning tasks in a complex three-dimensional environment, and has the characteristics of light weight, accuracy, high efficiency and the like.

Systems and techniques for inspecting vehicles, such as cars or trucks. Some embodiments provide a method for inspecting a vehicle using a vehicle inspection system. The vehicle inspection system may include a vehicle exterior inspection system having a plurality of exterior sensor arrays, the plurality of exterior sensor arrays including respective sets of sensors oriented in different directions. The vehicle inspection system may include a vehicle undercarriage inspection system having an undercarriage sensor array including cameras configured to capture images at different angles, and a sensor configured to output vehicle position / motion data. In some embodiments, the method includes: moving the vehicle and / or vehicle inspection system; triggering capture, by the exterior sensor arrays and undercarriage sensor array, of sensor data about the vehicle; capturing the sensor data; stopping capture of the sensor data; stopping movement of the vehicle and / or vehicle inspection system; and generating a vehicle condition report.

The invention relates to the technical field of three-dimensional point cloud scanning, and discloses a three-dimensional laser scanner and a real-time point cloud scanning method and system, and the method comprises the following steps: setting a dot spacing, setting the rotating speed of a drive motor, and setting the transmitting frequency of a laser transmitting and receiving module; controlling the motor to rotate to drive the laser to emit laser; when laser starts to be emitted, the reading module collects angle parameters of the coded disc in real time and measures a laser flight time value through the laser flight time measuring module; combining the angle parameter, the laser flight time value and the laser reflection intensity value read in the last time of laser emission into a frame of data and storing the frame of data in an original data buffer area; reading frame data of the original data buffer area in real time, and calculating to obtain a three-dimensional coordinate; and taking the three-dimensional coordinates and the laser reflection intensity value as a frame of data and storing the frame of data in a result buffer area. The problem that the real-time application requirement is difficult to meet in the prior art is solved, and the method has the advantage of being capable of reducing the operand.

A three-dimensional (3D) model generation method includes obtaining, based on a missing region of an initial 3D model of a target object, a camera pose sequence for performing photographing on the target object, where the initial 3D model is obtained by performing 3D modeling based on a first image sequence, and the first image sequence is obtained by performing photographing around the target object; obtaining a second image sequence, where the second image sequence is an image sequence obtained by performing photographing on the target object based on the camera pose sequence; and filling the missing region of the initial 3D model based on the second image sequence to generate a target 3D model.

Methods, systems, and non-transitory computer readable storage media are disclosed for generating three-dimensional meshes representing two-dimensional images for editing the two-dimensional images. The disclosed system utilizes a first neural network to determine density values of pixels of a two-dimensional image based on estimated disparity. The disclosed system samples points in the two-dimensional image according to the density values and generates a tessellation based on the sampled points. The disclosed system utilizes a second neural network to estimate camera parameters and modify the three-dimensional mesh based on the estimated camera parameters of the pixels of the two-dimensional image. In one or more additional embodiments, the disclosed system generates a three-dimensional mesh to modify a two-dimensional image according to a displacement input. Specifically, the disclosed system maps the three-dimensional mesh to the two-dimensional image, modifies the three-dimensional mesh in response to a displacement input, and updates the two-dimensional image.

The present disclosure relates to a method, a device, and a computer program product for image processing. The method includes acquiring a plurality of images and a first video, wherein the plurality of images indicate visual information of a source object from a plurality of perspectives, and the first video indicates animation of a target object. The method further includes generating a three-dimensional model for the source object based on the plurality of images, and generating a plurality of animation models for the target object based on the first video. The method further includes fusing the three-dimensional model for the source object and the plurality of animation models for the target object to generate a second video for the source object, wherein in the second video, the target object in the first video is replaced with the source object.

The invention provides a building structure intelligent design method and system based on 3D modeling, electronic equipment and a storage medium, and relates to the field of building intelligent design, and the method comprises the following steps: obtaining three-dimensional data through an unmanned aerial vehicle, oblique photography and laser scanning, and carrying out data cleaning, conversion and screening; performing feature extraction and semantic segmentation on the point cloud data by using a deep learning model; performing intelligent registration and fusion on the two types of point clouds by adopting a deep learning model; a three-dimensional model is generated based on the registration data, a parameter form is generated in combination with the building planning data, and a component model is constructed; collision detection, smooth verification and mechanical simulation are carried out on the whole model, parameter adjustment and optimization are carried out according to verification results, through automatic data collection, multi-source point cloud fusion, intelligent parameter modeling and multi-dimensional verification and optimization, the problems that a traditional method is low in efficiency and insufficient in precision are solved, and the design efficiency and precision are remarkably improved.

The invention relates to a myocardial resection area determination method and device, computer equipment and a storage medium. The method comprises the following steps: acquiring a heart scanning image, and determining a heart three-dimensional model according to the heart scanning image; determining a heart short-axis section and a heart long-axis section according to the heart three-dimensional model; the clock direction angle of the scalpel is determined on the short axis section of the heart; based on the clock direction angle, the excision depth of the scalpel is determined on the long-axis section of the heart; and performing myocardial resection on the heart three-dimensional model according to the clock direction angle and the resection depth through a scalpel model, determining an updated three-dimensional model after myocardial resection, and marking a resection area on the updated three-dimensional model. According to the scheme, accurate preoperative planning guidance and comprehensive heart structure information presentation are provided for the myocardial rotary cutting operation, and therefore the accuracy and safety of the myocardial rotary cutting operation are improved.

The embodiment of the invention provides a three-dimensional reconstruction method and device, equipment and a storage medium. A mark point can be arranged on the surface of a target object, and structured light is projected to the surface of the target object in at least partial time period of scanning the target object so as to acquire an image frame set of the target object. And reconstructing based on each image frame containing the mark point pattern to obtain a three-dimensional point cloud of each frame of mark point, and reconstructing based on each image frame containing the structured light pattern to obtain a three-dimensional point cloud of each frame of target object. Since the coordinate transformation relationship between the three-dimensional point clouds of each frame of mark point can be accurately determined, and the coordinate transformation relationship between the three-dimensional point clouds of the single frame of mark point and the three-dimensional point clouds of the single frame of target object can also be determined, the splicing of the three-dimensional point clouds of the multiple frames of target objects can be assisted by means of the mark points; and thus, a more accurate splicing result can be obtained.

An embodiment of this application discloses a three-dimensional reconstruction method. The method in this embodiment of this application includes: obtaining an image of a first object and a camera pose of the image; determining a first normalized object location field NOLF image of the first object in the image by using a first deep learning network, where the first NOLF image indicates a normalized three-dimensional point cloud of the first object at a photographing angle of view of the image; determining, from a plurality of three-dimensional models in a model database based on the first NOLF image, a first model corresponding to the first object; determining a pose of the first object based on the first model and the camera pose of the image; and performing three-dimensional reconstruction on the first object in the image based on the first model and the pose of the first object.