DATALOGIC IP TECH S.R.L.

The disclosure relates generally to systems and methods for quickly and efficiently estimating motion of objects in a three-dimensional environment. The disclosed systems and methods include generating a voxelized map of the three-dimensional environment including voxelized objects of interest. Each voxel in the voxelized object is associated with a signature identifier that may be used to track motion of the voxelized objects at different timepoints. By creating data structures of the three-dimensional environment at different timepoints and comparing changes in the coordinate position of the voxelized objects, motion data of the voxelized objects in the three-dimensional environment may be estimated.

A system and method of detecting an image of a template object in a captured image may include comparing, by a processor, an image model of an imaged template object to multiple locations, rotations, and scales in the captured image. The image model may be defined by multiple model base point sets derived from contours of the imaged template object, where each model base point set inclusive of a plurality of model base points that are positioned at corresponding locations associated with distinctive features of the imaged template object. Each corresponding model base point of the model base point sets may (i) be associated with respective layers and (ii) have an associated gradient vector. A determination may be made as to whether and where the image of the object described by the image model is located in the captured image.

A computer-implemented method of performing a three-dimensional 3D point cloud registration with multiple two-dimensional (2D) images may include estimating a mathematical relationship between 3D roto-translations of dominant planes of objects in a 3D point cloud and bi-dimensional homographies in a 2D image plane, thereby resulting in a 3D point cloud registration using multiple 2D images. A trained classifier may be used to determine correspondence between homography matrices and inferred motion of the dominant plane(s) on a 3D point cloud for paired image frames. A homography matrix between the paired images of the dominant plane(s) on the 2D image plane may be selected based on the correspondence between the inferred motions and measured motion of the dominant plane(s) on the 3D point cloud for the paired image frames. The process may be less computationally intensive than conventional 2D-3D registration approaches.

A computer-implemented method of performing a three-dimensional 3D point cloud registration with multiple two-dimensional (2D) images may include estimating a mathematical relationship between 3D roto-translations of dominant planes of objects in a 3D point cloud and bi-dimensional homographies in a 2D image plane, thereby resulting in a 3D point cloud registration using multiple 2D images. A trained classifier may be used to determine correspondence between homography matrices and inferred motion of the dominant plane(s) on a 3D point cloud for paired image frames. A homography matrix between the paired images of the dominant plane(s) on the 2D image plane may be selected based on the correspondence between the inferred motions and measured motion of the dominant plane(s) on the 3D point cloud for the paired image frames. The process may be less computationally intensive than conventional 2D-3D registration approaches.

A device includes an image sensor and a processor to: receive training images that each include multiple visual features and an ROI; receive indications of locations of ROIs within each training image; perform at least one transform on the multiple visual features and ROI of at least one training image to align the multiple visual features and ROIs among all of the training images to identify a common set of visual features present within all of the training images; derive a converged ROI from at least a portion of the ROI of at least one training image; and generate an anchor model based on the converged ROI and the common set of visual features, wherein the common set of visual features defines the anchor and are each specified relative to the converged ROI, and the anchor model is used to derive a location of a candidate ROI in an image.