depthimage

Depthimage, or depth image, is a two-dimensional image in which each pixel encodes the distance from the imaging sensor to the scene along the corresponding camera ray. Depthimages are produced by depth-sensing cameras such as time-of-flight devices, structured-light systems, or stereo camera rigs, and can also be computed from stereo disparity. They provide direct three-dimensional information without requiring explicit 3D reconstruction.

Data representation: depth values are usually stored as a 2D array with units that are meters or millimeters. Common formats include 16-bit unsigned integers (often representing millimeters) or 32-bit floating-point values (meters). Missing measurements are indicated by zeros, a maximum value, or NaN. Depth images are frequently aligned with color or infrared images, requiring calibration (intrinsic and extrinsic) to map to real-world coordinates.

Processing and challenges: depth images can be noisy and contain holes caused by reflective surfaces, occlusions, or low texture. Processing steps include noise filtering, edge-preserving smoothing, hole filling or depth completion, and alignment with 3D representations such as point clouds or meshes.

Applications: depthimages are used in robotics for navigation and obstacle avoidance, in 3D scanning and reconstruction, in augmented reality and mixed reality, and in gesture recognition. They are commonly converted to point clouds or meshes and registered with color data for richer scene understanding.