Abstract: A method includes: obtaining a registration depth map and a target spectral image; obtaining depth information of a target point from the registration depth map, and obtaining a first grayscale value of the target point; calculating 3D coordinate information of the target point according to the depth information of the target point, a first internal parameter of the imaging spectrum device, and first coordinate information of the target point in the target spectral image; determining normal vector information corresponding to the target point according to neighboring points of the target point; inputting the 3D coordinate information and the normal vector information into a full-light model of a diffuse reflecting plate to obtain a second grayscale value; and obtaining reflectance of the diffuse reflecting plate, and calculating reflectance of the target point according to the reflectance of the diffuse reflecting plate, the first grayscale value, and the second grayscale value.

Abstract: A light source spectrum acquisition method includes: acquiring a multispectral image, and determining a multispectral response value of each pixel in the multispectral image; reconstructing an RGB image according to the multispectral image; converting the RGB image into a grayscale image; and determining a target region in the grayscale image, wherein in the target region a grayscale value is less than or equal to a threshold, and calculating a light source spectrum response value according to multispectral response values of pixels of the multispectral image corresponding to the target region.

Abstract: A method includes: obtaining a first image including a foreground, a second image including a background, and soft segmentation of the foreground; and inputting the first image, the second image, and the soft segmentation into an image matting network, and outputting a foreground segmentation of the first image. The image matting network includes at least one stage network that includes a context combining module, a stem block, and a predictor module. The first image, the second image, and the soft segmentation of the foreground are inputted into the context combining module, the context combining module is configured to output a low-order feature and a high-order feature after feature exchange, the stem block is configured to fuse the low-order feature and the high-order feature based on an attention mechanism to obtain a fused feature, and the predictor module is configured to output the foreground segmentation according to the fused feature.

Abstract: A method includes: obtaining and splicing a plurality of frames of point cloud data to obtain a complete point cloud model of an indoor structure; extracting point cloud plane features corresponding to the complete point cloud model, and classifying the point cloud plane features into indoor structure types comprising a floor, a wall, and a ceiling; projecting point cloud data of the wall onto a plane of the floor to generate a planar grid graph; and generating a two-dimensional plane view of the indoor structure based on the planar grid graph.

Abstract: A method for correcting interference fringes includes: obtaining correction parameter sets of different photographing distances; obtaining a to-be-corrected image and calculating an average depth value of the to-be-corrected image; selecting a target correction parameter set corresponding to the average depth value from the correction parameter sets of different photographing distances; and correcting first pixel values of to-be-corrected pixels at the different coordinate positions in the to-be-corrected image according to different target correction parameters corresponding to the different coordinate positions in the target correction parameter set, to obtain a corrected image. Each of the correction parameter sets includes different correction parameters corresponding to different coordinate positions. The average depth value includes an average value of depth values corresponding to a plurality of to-be-corrected pixels in the to-be-corrected image.

Abstract: A method for calibrating an optical axis includes: acquiring a first speckle image and a second speckle image captured respectively before and after a speckle emission device is rotated; extracting at least two first speckle points in the first speckle image and second speckle points in the second speckle image corresponding to the at least two first speckle points, to obtain at least two line segments formed by lines connecting the first speckle points and the second speckle points corresponding to the first speckle points in the same image coordinate system; and calculating a perpendicular bisector of each of the line segments, and determining an optical axis reference point according to an intersection point of the perpendicular bisectors. The method provided in embodiments of this application can reduce the impact of a jig tolerance on the optical axis calibration, and improve the accuracy of the optical axis calibration.

Abstract: A method for correcting interference fringes includes: obtaining interference fringe images with different photographing distances; obtaining a to-be-corrected image including to-be-corrected pixels, and calculating a depth value of each to-be-corrected pixel in the to-be-corrected image; selecting, from the interference fringe images with different photographing distances, an interference fringe image corresponding to the depth value of each to-be-corrected pixel as a target interference fringe image; extracting first pixel values of target coordinate positions in the target interference fringe image; and correcting second pixel values of to-be-corrected pixels according to the first pixel values corresponding to the to-be-corrected pixels to obtain a corrected image.

Abstract: This application discloses a calibration method, a calibration system, and a calibration board. The calibration board includes checkerboard cells arranged on a surface of the calibration board, where at least one dot is arranged in each of the checkerboard cells, the dots of the calibration board include at least a first feature dot and a second feature dot, the first feature dot and the second feature dot are arranged in the checkerboard cells according to a random rule or a specific rule, and a diameter of the first feature dot is greater than a diameter of the second feature dot. The calibration board in this application improves the stability and accuracy of calibration board detection by utilizing characteristics such as a high detection accuracy and a strong anti-blur capability of a dot mark, combining with corner points of checkerboard cells.

Abstract: The specification provides a depth imaging method and device and a computer-readable storage medium. The method includes: controlling an emission module comprising a light emitting device to emit at least two speckle patterns that change temporally to a target object; controlling an acquisition module comprising a light sensor to acquire reflected speckle patterns of the at least two speckle patterns reflected by the target object; and performing spatial-temporal stereo matching by using the reflected speckle patterns and the at least two reference speckle patterns, to calculate offsets of pixel points between speckles of the at least two reference speckle patterns and speckles of the reflected speckle patterns, and calculating depth values of the pixel points according to the offsets.

Abstract: An optical information detection method includes: acquiring a first image captured by a first imaging device, wherein the first image comprises a target full-field projection pattern; extracting features from the first image to obtain first feature information; acquiring second feature information and first graphic information of a reference full-field projection pattern, wherein the first graphic information comprises zero-order information and/or secondary information; calculating a first mapping relationship between the first feature information and the second feature information; mapping the first graphic information to the target full-field projection pattern according to the first mapping relationship, to obtain second graphic information corresponding to the target full-field projection pattern; and calculating target optical information according to the second graphic information.