7results about How to "Accurate depth information" patented technology

The invention relates to the technical field of stereo matching, in particular to a three-eye camera video picture output method and device, and the method comprises the following steps: extracting information from a video picture of a three-eye camera, calculating the Gaussian curvature and average curvature of multiple points in an image, calculating differential geometric characteristics, recording curvature attributes, and generating a curved surface feature set. According to the method, parallax adjustment is carried out on the edge area of the object by applying high-order curvature information, more accurate depth information can be provided at the edge of the object through the edge optimization disparity map, the method is particularly key to reconstruction of a fine structure in a scene, and in the generation process of the global optimization disparity map, through parallax comparison and correction between continuous scene frames, more accurate depth information can be provided at the edge of the object. According to the method, the moving object or scene change can be processed more effectively, the continuity and accuracy of the disparity map are ensured, and the quality of three-dimensional reconstruction and the practicability of applications such as automatic driving and video monitoring are remarkably improved in a dynamic environment.

This invention discloses a method for constructing a realistic depth estimation training dataset based on a 3D mesh and 3D Gaussian splashing, relating to the field of UAV aerial surveying technology. The method involves acquiring aerial image sequences using a visible light camera mounted on a UAV platform, and filtering and preprocessing the image quality; generating a dense point cloud and a 3D mesh; training and generating a 3DGS model based on camera pose and image sequences consistent with the 3D mesh; unifying the 3D mesh, camera pose, and 3DGS model to the same world coordinate system; and using Z-buffer rasterization to output pixel-by-pixel depth and validity masks; generating RGB images of the corresponding viewpoint in the 3DGS renderer; and encapsulating the dataset according to a recommended data organization structure. This invention can accurately generate realistic RGB and geometrically accurate depth, with strict pixel alignment between the RGB images and depth maps, adapting to the needs of real-world application scenarios such as UAV aerial surveying and actual terrain mapping.

The invention discloses a near-infrared two-region multi-channel living body macroscopic stereo imaging system and a near-infrared two-region multi-channel living body macroscopic stereo imaging method. According to the invention, visible light and infrared imaging technologies are integrated, and the coaxial dual-channel lens, the illumination and excitation light source and the stereo display system are adopted, so that high-quality stereo images can be captured by crossing different spectral intervals. According to the system, a coaxial dual-channel lens is used for splitting light, so that synchronous acquisition of near-infrared and visible light signals in the same view field is realized; the image quality is enhanced by using a specific image processing algorithm; integrating image information of different spectrums by using an image fusion technology; a stereo display system is adopted, and a final stereo imaging result is presented through a naked eye 3D display technology. In addition, a binocular depth estimation algorithm is adopted to accurately calculate depth information, and more specific target depth information is further provided to help operation. According to the technical scheme, utilization of macroscopic three-dimensional information of living bodies in multiple fields such as medical diagnosis, biological scientific research and industrial detection can be improved.

The present invention discloses a SAR target positioning and detection method based on radar and optical fusion, which belongs to the field of target recognition technology, and includes the following steps: step 1, setting the center position of the radar and camera field angle on the same vertical line, controlling the radar and the camera to work at the same time, collecting radar data and optical data at the same time, and after the collection is completed, performing signal processing, radar imaging and imaging result enhancement on the collected radar signal to obtain a SAR image; performing data processing on the collected optical signal to obtain an optical image, and extracting optical target depth information at the same time; step 2, combining the data enhanced SAR image and the optical image for inter-frame alignment; step 3, after inter-frame alignment, converting the radar image target and the optical image target to the same spatial coordinate system to complete target positioning alignment; step 4, fusing the optical image target category to the SAR image to complete SAR target detection. The present invention improves the overall detection speed and accuracy.

The invention discloses a copper mine mining quantity estimation method based on binocular image three-dimensional reconstruction. The method specifically comprises the following steps: acquiring binocular images of a camera of a satellite for shooting a target copper mine mining area at two positions; processing the binocular image by using distortion correction and stereo correction to obtain an overlapping region in the binocular image; calculating the depth of each pixel point in the overlapping region in the binocular image, obtaining the depth information of each pixel point, and generating a depth map; identifying the mining area of the binocular image by using a pre-trained yolk v8 model; calculating the mining amount by combining the depth information and the identified mining area; the method has the advantages that the binocular image is used for performing three-dimensional reconstruction on the copper ore area, compared with monocular and multi-view three-dimensional reconstruction, the method is simpler and more efficient, and more accurate depth information can be obtained; the binocular image is used for carrying out three-dimensional reconstruction on the copper ore area, compared with monocular and multi-view three-dimensional reconstruction, the method is simpler and more efficient, and more accurate depth information can be obtained.

The present application discloses a method and related device for target positioning that integrates laser radar and camera, and the method includes: performing rough calibration operations on the selected laser radar and camera to obtain an initial transformation matrix composed of an initial rotation vector and an initial translation vector; performing ICP matching analysis based on the point cloud data between two adjacent frames and the global source point cloud data to obtain the radar rotation vector and radar translation vector with the minimum matching error; triangulating the camera image through the radar rotation vector and the radar translation vector to obtain the camera image depth information; performing relative calculations on the camera image depth information and the preset radar image depth information based on the least squares method to obtain the optimal transformation matrix; using the optimal transformation matrix to perform positioning calculations on the target object to obtain the positioning result. The present application solves the technical problem that the existing technology has limitations to varying degrees, resulting in low accuracy of the positioning result.

The invention provides a non-vision-field imaging method and device based on quasi-Fresnel transformation, and belongs to the technical field of optical non-vision-field imaging. The method comprises the following steps: collecting photon echo data when non-vision-field imaging is carried out on a non-vision-field target as a detection signal; calculating a convolution size parameter of the detection signal based on quasi-Fresnel transform; calculating a two-dimensional complex representation of the detection signal based on the convolution size parameter; calculating a two-dimensional complex representation of the non-view target based on the two-dimensional complex representation of the detection signal; and calculating reflectivity and depth information of the non-vision target based on the two-dimensional complex representation of the non-vision target so as to realize imaging of the non-vision target. According to the method, the two-dimensional features of the surface of the non-vision-field target are fully utilized, the non-vision-field target can be accurately and efficiently reconstructed, and the method has the advantages of being high in calculation speed, small in used memory, low in data dependence and accurate in imaging and positioning.