rangePx

RangePX is a software framework and data format used in range imaging and depth sensing to represent per-pixel distance information. It refers to a collection of tools, libraries, and conventions for capturing, processing, and exchanging depth data derived from stereo cameras, structured light, time-of-flight sensors, or fused multi-sensor systems.

At its core, a range image stores a distance value for each pixel, typically in meters, with special values indicating missing or invalid data. RangePX specifies standard encodings for these values, along with accompanying metadata such as camera intrinsics, sensor extrinsics, timestamps, and scene confidence. It supports multiple data channels, including range, intensity, and confidence, to aid processing.

The typical workflow involves data acquisition from one or more range sensors, pre-processing to handle noise and alignment, range computation or extraction, optional sensor fusion to improve accuracy, and post-processing such as filtering, downsampling, and visualization or export to common formats like PNG, PLY, or CSV.

RangePX is used in robotics and automation for navigation and obstacle avoidance, in autonomous vehicles for scene understanding, and in 3D mapping and augmented/virtual reality pipelines where depth information enhances realism and interaction.

The project originated in academic and industrial settings and has evolved as an open ecosystem with software implementations in languages such as C++ and Python. It emphasizes interoperability, modularity, and clear documentation to facilitate integration with existing perception stacks.

Related topics include depth sensing, depth maps, point clouds, LiDAR, stereo vision, and 3D reconstruction.