Abstract: Described are systems, methods, and apparatus for generating motion extracted images having a high dynamic range (“HDR”) based on image data obtained from one or more image sensors at different times. The implementations described herein may be used with a single image sensor or camera that obtains images at different exposures sequentially in time. The images may be processed to detect an object moving within the field of view and pixel information corresponding to that moving object extracted. The non-extracted image data may then be combined to produce a motion extracted HDR image that is substantially devoid of the moving object.

Abstract: Described are systems and methods for generating high dynamic range (“HDR”) images based on image data obtained from different image sensors for use in detecting events and monitoring inventory within a materials handling facility. The different image sensors may be aligned and calibrated and the image data from the sensors may be generated at approximately the same time but at different exposures. The image data may then be preprocessed, matched, aligned, and blended to produce an HDR image that does not include overexposed regions or underexposed regions.

Abstract: A time-of-flight camera may be used to generate depth images of a scene, where the scene includes retroreflective materials. The time-of-flight camera may capture first sensor data by illuminating the scene at a first energy level, and by exposing a sensor for a first duration. If portions of the sensor are saturated by light reflected from retroreflective materials, corrupted pixels are detected within the first sensor data. A second set of sensor data may be captured by illuminating the scene at a second energy level, or by exposing the sensor for a second duration. The corrupted pixels may be identified and masked from the first sensor data, which may be blended with corresponding portions of the second sensor data to generate a depth image.

Abstract: Described are systems, methods, and apparatus for generating motion extracted images having a high dynamic range (“HDR”) based on image data obtained from one or more image sensors at different times. The implementations described herein may be used with a single image sensor or camera that obtains images at different exposures sequentially in time. The images may be processed to detect an object moving within the field of view and pixel information corresponding to that moving object extracted. The non-extracted image data may then be combined to produce a motion extracted HDR image that is substantially devoid of the moving object.

Abstract: Described are systems and methods for generating high dynamic range (“HDR”) images based on image data obtained from different image sensors for use in detecting events and monitoring inventory within a materials handling facility. The different image sensors may be aligned and calibrated and the image data from the sensors may be generated at approximately the same time but at different exposures. The image data may then be preprocessed, matched, aligned, and blended to produce an HDR image that does not include overexposed regions or underexposed regions.

Abstract: A shifted-lens imaging device mounted in association with a storage unit captures images of interactions by personnel with items on the storage unit. The images are processed to identify the items and/or the personnel, and to store an association of the personnel with the items in one or more data stores. The shifted-lens imaging device is mounted to the storage unit directly, such as by one or more extensions, and aligned to include a planar area of interest associated with the storage unit, such as a frontal area of the storage unit, within a field of view. A lens of the shifted-lens imaging device is shifted in parallel with the imaging sensor by an offset or distance that may be selected based on a focal length of the imaging device and a distance between the imaging device and dimensions of the planar area of interest.

Abstract: A plurality of cameras obtain images in an environment. Calibration data about the translation and rotation of the cameras is used to support functionality such as determining a location of an object that appears in those images. To determine the calibration data, images are processed to determine features produced by points in the environment. A cluster is designated that includes the images of the same point as viewed from at least some of the cameras. A triangulation matrix and back-projection error are calculated for each cluster. These triangulation matrices and back-projection errors are then perturbed to find the translation and rotation that minimize eigenvalues and eigenvectors representative of the feature. The eigenvalues and eigenvectors are then used to determine the rotation and translation of the respective cameras with respect to an origin.

Abstract: Described are systems, methods, and apparatus for generating motion extracted images having a high dynamic range (“HDR”) based on image data obtained from one or more image sensors at different times. The implementations described herein may be used with a single image sensor or camera that obtains images at different exposures sequentially in time. The images may be processed to detect an object moving within the field of view and pixel information corresponding to that moving object extracted. The non-extracted image data may then be combined to produce a motion extracted HDR image that is substantially devoid of the moving object.

Abstract: A method for controlling movement of at least one movable object, a computer readable storage medium and an apparatus configured to control movement of at least one movable object. Position information of the at least one movable object is retrieved. The retrieved position information is compared with a viewing zone of the camera to determine if the movable object is about to interfere with the viewing zone. An alert is generated for the at least one movable object in case the movable object is about to interfere with the viewing zone (4a, 4b) and movement of the at least one movable object is controlled responsive to the alert.

Abstract: A method and an apparatus for extracting feature correspondences from images are described. An image dataset, feature points of the images and preliminary correspondences of the feature points are acquired (10) as input data. At least one cluster of the feature points is generated. In a same cluster, each feature point is coupled to at least one other feature point as preliminary feature correspondences. For each cluster, primary feature correspondences of the feature points are determined by determining consistency measures between every two feature points in the cluster. The cluster is then segmented by maximizing an average of the consistency measures of the cluster.

Abstract: A method for extracting keyframes from a sequence of frames for a computer vision application using structure from motion, the keyframes being a subset of representative frames from the complete sequence of frames, and an apparatus configured to perform the method are described. A subset selector selects a subset of keyframes that closely match a current camera position from already available keyframes. A determination unit then determines whether a current frame should be included a bundle adjustment keyframe set and/or a triangulation keyframe set.

Abstract: A method and an apparatus for camera calibration are described. The method and the apparatus use an image dataset in which a calibration object is captured by a camera. 2D and 3D correspondences are acquired from the image dataset, as well as reprojection errors of the 2D and 3D correspondences. A reliability map of a retinal plane of the camera is generated using the acquired reprojection errors, which indicates a reliability measure of the geometrical information carried by each pixel of the retinal plane of the calibrated camera.

Abstract: An apparatus and a method of determining a sequence of transitions for a varying state of a system, wherein the system is described by a finite number n of states, and wherein a transition from a current state to a next state causes a cost in dependence of a distance that is dependent on a previous state, the current state, and the next state.

Abstract: A method and an apparatus for camera calibration using a color-coded structure are described. The camera calibration method comprises: receiving from a camera an image, in which a color-coded structure is captured; determining a plurality of coordinate points in the image; and performing camera calibration using the coordinate points. The color-coded structure includes at least two sets of patterns in different color channels.

Abstract: A method and an apparatus for bilayer image segmentation are described. A set of segmentation seeds for the image is generated by analyzing a depth histogram of the image. Then a segmentation map is generated by minimizing an objective function, which models a directed flow from the foreground segmentation seeds towards the background segmentation seeds.

Abstract: A method and an apparatus for bi-layer segmentation of an image or a sequence of images are described. A classifier is derived based on depth data of the image and another classifier is derived based on color data of the image. The image is then segmented by maximizing a weighted sum of matching scores derived from the classifiers based on depth data and color data of the image. The classifier based on color data of the image is derived using color sampling subsequent to generating an initial segmentation of the image.

Abstract: A method and an apparatus for bilayer image segmentation are described. A set of segmentation seeds for the image is generated by analyzing a depth histogram of the image. Then a segmentation map is generated by minimizing an objective function, which models a directed flow from the foreground segmentation seeds towards the background segmentation seeds.

Abstract: A method and an apparatus for bi-layer segmentation of an image or a sequence of images are described. A classifier is derived based on depth data of the image and another classifier is derived based on color data of the image. The image is then segmented by maximizing a weighted sum of matching scores derived from the classifiers based on depth data and color data of the image. The classifier based on color data of the image is derived using color sampling subsequent to generating an initial segmentation of the image.