32results about How to "Improve 3D reconstruction accuracy" patented technology

The invention discloses a pyramid Transform-based point cloud reconstruction method, device and equipment, and a medium, and belongs to the field of computer vision. The method comprises the steps that: after point feature data corresponding to each view are obtained, the point feature data are input into a two-channel pyramid network, and point features input in each layer pass through three modules of sampling, a graph neural network and Transformer; when the point feature data passes through the first channel, the updated output of each layer of attention mechanism is input into the next layer, and when the point feature data passes through the second channel, the updated output of each layer of attention mechanism and the updated output of the attention mechanism of the corresponding layer of the first channel located in the same layer as the next layer are spliced and then input into the next layer; and the updated output of the attention mechanism of the bottommost layer of the second channel is converted into a point cloud coordinate to obtain a point cloud result of each view, and an intersection of the point cloud results of each view is taken to represent a target object after point cloud reconstruction. Therefore, the three-dimensional reconstruction precision of the object is higher.

The invention provides a three-dimensional imaging method and device based on unmanned aerial vehicles and a computer readable storage medium, wherein the number of the unmanned aerial vehicles is not less than two. The method comprises the steps: controlling one of the two unmanned aerial vehicles to be located at a first position, and controlling the other unmanned aerial vehicle to be located at a second position; acquiring target images of target objects shot by the two unmanned aerial vehicles at the respective positions, wherein the target objects in the target images of the target objects shot by the two unmanned aerial vehicles at the respective positions at least partially coincide; and generating a three-dimensional image of the target object according to the target images of the target object shot by the two unmanned aerial vehicles at the respective positions. According to the invention, two unmanned aerial vehicles are used for replacing two cameras in a binocular camera, and two target images of the target object are obtained from different positions, thereby obtaining a three-dimensional image of the target object, increasing the distance between phase collection points in a low-cost manner, reducing the demands for imaging resolution, and improving the three-dimensional reconstruction precision.

The present invention provides a UAV-based stereoscopic imaging method and device, the method comprising: acquiring at least two target images of a target object from different positions by a UAV, wherein at least two of the target images are The target objects are at least partially overlapped; and a stereoscopic image of the target object is generated according to at least two of the target images. The two cameras in the binocular camera are replaced by a drone, and at least two target images of the target object are acquired from different positions, thereby obtaining a stereoscopic image of the target object, and the distance between the phase acquisition points is increased in a low-cost manner. The distance reduces the demand for imaging resolution and improves the accuracy of 3D reconstruction.

The invention discloses a binocular visual sense reconstruction method considering three-dimensional distortion, belonging to the computer visual sense measurement technology field. The binocular visual sense reconstruction method considering the three-dimensional distortion introduces a three-dimensional distortion correction function to perform compensation optimization on a visual sense reconstruction result and realizes three-dimensional high accuracy remodeling. The binocular visual sense reconstruction method considering the three-dimensional distortion adopts a traditional binocular visual sense reconstruction method to obtain a three-dimensional reconstruction initial values of measurement points, solves a three-dimensional distortion compensation value on the basis of a relative relation between a standard object and a binocular camera and the reconstruction initial values, thus performs distortion compensation on coordinates of the camera of a tested point which are respectively extracted from a left image and a right image, combines with binocular visual sense theory to reconstruct, and realizes the binocular three-dimensional high accuracy reconstruction. The binocular visual sense reconstruction method considering three-dimensional distortion is high in reconstruction accuracy, can compensate the three-dimensional distortion of the binocular visual sense, realizes form and surface reconstruction of a part having a big curvature and a big size and improves three-dimensional reconstruction accuracy of the binocular visual sense.

A system and method for three-dimensional reconstruction of a microsurgical field, which collects the pattern information of the measured scene through the visible light viewpoint acquisition unit; collects the infrared speckle pattern of the measured scene through the infrared light viewpoint acquisition unit; uses the three-dimensional reconstruction calculation control unit to control the visible light The viewpoint collection unit and the infrared viewpoint collection unit take photos, and fuse the patterns obtained by the visible light viewpoint collection unit with the patterns obtained by the infrared viewpoint collection unit to obtain a three-dimensional reconstruction result. This technical solution introduces multi-viewpoint joint optimization and infrared speckle-based object surface texture enhancement mechanism into high-precision 3D reconstruction. By designing the structure of infrared photosensitive element and speckle projector, the shape structure of surgical field can be accurately obtained. The shape structure is used as the surgical field prior to optimize the three-dimensional reconstruction model under visible light, thereby improving the three-dimensional reconstruction accuracy under the microscope without affecting the main optical path of the microscope.

The invention discloses a method for correcting kinematic errors in a microscopic vision system, which is applied to the microscopic vision system. The method comprises the following steps of: calibrating a camera through a calibration board, establishing a camera model, tracking angular points of the calibration board, generating a tracking trajectory and an ideal revolving trajectory of the angular points through a least square method, then performing error evaluation, establishing an error correcting model, and correcting the microscopic image by virtue of the error correct model. The method provided by the invention can be used for establishing the error correct model for the microscopic vision system only through one-time error evaluation; and the model can be used for any microscopic movement video collected by the system, reducing the movement errors of the system and improving the three-dimensional reconstruction precision, thus being high in practicability.

The photometric stereo 3D reconstruction method based on high-frequency area enhancement based on deep learning, including using photometric stereo system to take several images of the object to be reconstructed, and using deep learning algorithm to output accurate 3D reconstruction of surface normal, including the surface normal generation network is It is designed to generate the surface normal of the object to be reconstructed from the image and lighting; the attention weight generation network generates the attention weight map of the object to be reconstructed from the image; the attention weight loss function is processed pixel by pixel; then the trained network is used Surface normal reconstruction from photometric stereo images. The present invention learns surface normal and high-frequency information respectively through the proposed surface normal generation network and attention weight generation network, and uses the proposed attention weight loss for training, which can improve the reconstruction accuracy of the surface in high-frequency areas such as fold edges . Compared with the previous traditional photometric stereo method, the accuracy of 3D reconstruction is improved, especially the details of the surface of the object to be reconstructed.