ZED cameras are passive depth sensors based on stereo vision: depth is computed from the two synchronized RGB images, exactly like human binocular vision. Unlike active sensors, they do not emit any laser, IR pattern, or Time-of-Flight signal. This design brings several concrete advantages.
No interference, indoors and outdoors
Since ZED cameras do not project any light, multiple cameras can operate in the same area at the same time without any cross-interference; this is a key requirement for multi-camera and multi-robot setups. For the same reason, they are not affected by sunlight and work reliably outdoors, where IR-based active sensors typically fail or see their range drastically reduced.
Long depth range
Active sensors are limited to short ranges for physical reasons (emitter power, eye safety, IR reflectivity). Stereo vision scales with resolution and baseline instead, so ZED cameras can measure depth at much longer distances:
- ZED 2i: depth range up to 20 m, with detection capabilities at even longer distances;
- ZED X: from 0.3 m to 20 m with the 2.2 mm lens, and from 1 m to 35 m with the 4 mm lens;
- ZED X Mini: from 0.1 m to 8 m, ideal for short-range manipulation and precision tasks;
- ZED X One: monocular GMSL2 modules that can be paired in a custom virtual stereo rig, with a user-defined baseline to extend the depth range even further.
AI-powered depth: the Neural Depth Engine
The ZED SDK includes the Neural Depth Engine, an AI-based stereo matching technology that produces dense, accurate, and temporally stable depth maps, even on low-texture surfaces (white walls, floors, uniform objects) that are historically challenging for stereo vision. Depth quality keeps improving with each SDK release through simple software updates, with no hardware change required.
Perfect image and depth alignment
Because depth is computed directly from the images, the left image and the depth map are natively and perfectly aligned: the same pixel coordinates address the same physical point in both. There is no need for extrinsic registration between an RGB sensor and a separate depth sensor, and no risk of misalignment over time.
High resolution and frame rate
ZED cameras provide high-resolution depth compared to active sensors: up to 2K on USB 3.0 models, and 1920x1200 at 60 fps with global shutter sensors on the ZED X family, for sharp, distortion-free images even on fast-moving platforms.
A complete software ecosystem
The ZED SDK provides much more than depth, all natively aligned to the same camera reference frame:
- Positional Tracking for visual-inertial odometry and SLAM;
- Spatial Mapping for real-time 3D mesh and point cloud reconstruction;
- Object Detection and Body Tracking with 3D localization;
- Fusion API to merge data from multiple cameras, and Global Localization to fuse visual-inertial tracking with GNSS data;
- Official integrations for ROS 2, NVIDIA Isaac ROS and Isaac Sim, Unity, Unreal Engine, and more; see the full list in the documentation.
Industrial-grade hardware options
Beyond the USB 3.0 lineup, the ZED X family offers a GMSL2 interface for low-latency, reliable transmission over long cables (up to 15 m), hardware synchronization of multiple cameras, IP67-rated enclosures (ZED X and ZED X Mini), and built-in vibration-resistant IMUs, making them suitable for 24/7 deployment on industrial robots and vehicles in combination with NVIDIA Jetson platforms and the ZED Box ecosystem.