In one embodiment, a method includes accessing first image data generated by a first image sensor having a first filter array that has a first filter pattern. The first filter pattern includes a first filter type corresponding to a spectrum of interest and a second filter type. The method also includes accessing second image data generated by a second image sensor having a second filter array that has a second filter pattern different from the first filter pattern. The second filter pattern includes a number of second filter types, the number of second filter types and the number of first filter types have at least one filter type in common. The method also includes determining a correspondence between one or more first pixels of the first image data and one or more second pixels of the second image data.

Panoramic videos are generated from multiple video feeds in real time received from multiple video cameras, such as an array of video cameras having overlapping fields of view. Texture mapping techniques are employed to correct lens distortion or other defects or deficiencies in the video frames in each of the video feeds caused by optical properties of the corresponding video camera. Video frames of related video feeds, such as the video feeds of cameras in an array of cameras having adjacent and overlapping fields of view, are seamlessly stitched automatically based on an initial manual configuration, again employing texture mapping techniques. A colour profile of the different video frames is normalized to provide a uniform and seamless colour profile of the video panoramic view.

A method for generating a three-dimensional model of an object, by a scanning system including a client-side device including: an acquisition system configured to capture images; and an interaction system including a display device and a network interface includes: capturing a plurality of images of the object by the acquisition system, the images being captured from a plurality of different poses of the acquisition system; computing depth maps from the images of the objects, each of the depth maps corresponding to one of the poses of the acquisition system; combining the depth maps to generate a combined point cloud; and displaying, on the display device, the combined point cloud or a 3D mesh model generated from the combined point cloud.

A 3D display panel assembly includes a display panel and an adjusting panel. The display panel includes a plurality of subpixels arranged in an array, any two adjacent lines along a first direction being respectively a first subpixel line and a second subpixel line, the first subpixel line including first subpixels configured to emit primary light, and the second subpixel line including second subpixels configured to display black when the first subpixel line emits the primary light. The adjusting panel includes a plurality of adjusting units arranged in an array, each line along a second direction including a plurality of continuously arranged first adjusting unit groups. A 3D display device and a method for driving the same are also provided.

A camera module, for a head-mounted display (HMD), includes a first optical module for tracking a hand motion of a user; a second optical module for reconstructing a hand gesture or a step of the user and a space; a third optical module for establishing a three-dimensional (3D) virtual object; and a control unit for integrating with the first optical module, the second optical module and the third optical module to virtualize a body behavior of the user; wherein the camera module is rotatable to maximize a tracking range.

A camera, including: two imaging systems each comprising a different optical path corresponding to a different viewing angle of an object; one or more illumination sources; a mask disposed with multiple pairs of apertures, wherein each aperture of each aperture pair corresponds to a different one of the imaging systems; at least one detector configured to acquire multiple image pairs of the object from the two imaging systems via the multiple pairs of apertures; and a processor configured to produce from the multiple acquired image pairs a multi-layer three dimensional reconstruction of the object.

In one embodiment, methods, media, and systems process and display light field images using a view function that is based on pixel locations in the image and on the viewer's distance (observer's Z position) from the display. The view function can be an angular view function that specifies different angular views for different pixels in the light field image based on the inputs that can include: the x or y pixel location in the image, the viewer's distance from the display, and the viewer's angle relative to the display. In one embodiment, light field metadata, such as angular range metadata and/or angular offset metadata can be used to process and display the image. In one embodiment, color volume mapping metadata can be used to adjust color volume mapping based on the determined angular views; and the color volume mapping metadata can also be adjusted based on angular offset metadata.

This disclosure relates to techniques for generating robust depth estimations for captured images using semantic segmentation. Semantic segmentation may be defined as a process of creating a mask over an image, wherein pixels are segmented into a predefined set of semantic classes. Such segmentations may be binary (e.g., a person pixel or a non-person pixel) or multi-class (e.g., a pixel may be labelled as: person, dog, cat, etc.). As semantic segmentation techniques grow in accuracy and adoption, it is becoming increasingly important to develop methods of utilizing such segmentations and developing flexible techniques for integrating segmentation information into existing computer vision applications, such as depth and/or disparity estimation, to yield improved results in a wide range of image capture scenarios. In some embodiments, an optimization framework may be employed to optimize a camera device's initial scene depth/disparity estimates that employs both semantic segmentation and color regularization in a robust fashion.

An imaging system includes a light source that, in operation, emits an emitted light containing a near-infrared light in a first wavelength region, an image sensor, and a double-band pass filter that transmits a visible light in at least a part of a wavelength region out of a visible region and the near-infrared light in the first wavelength region. The image sensor includes light detection cells, a first filter that selectively transmits the near-infrared light in the first wavelength region, second to fourth filters that selectively transmit lights in second to fourth wavelength regions, respectively, which are contained in the visible light, and an infrared absorption filter. The infrared absorption filter faces the second to fourth filters and absorbs the near-infrared light in the first wavelength region.

An agricultural working machine, in particular a tractor, has at least one agricultural working unit for working a crop field including a multitude of useful plants, and a camera system which includes a 3D camera. The camera system is configured for generating with the 3D camera 3D information regarding the crop field by recording stereoscopic image pairs along two different viewing axes. The viewing axes proceed from optical centers of the 3D camera, which are connected to each other by a baseline that is inclined relative to the horizontal.