ZED Tools 5.5

What’s New

ZED Diagnostic can now optimize a custom object detection ONNX model ahead of time, so an application using it starts immediately instead of optimizing the model on first use. ZED Depth Viewer exposes the new INT8 precision for NEURAL depth, and the ZED Tools support the new ZED X One Core camera. ZED Studio and ZED360 receive stability fixes.

5.5.0

Released: Sep 17, 2026

General

• Added support for the ZED X One Core camera (sl::MODEL::ZED_XONE_CORE), the ZED X One GS variant connected over MIPI through the ZED Link MIPI capture card.

ZED Studio

• Fixed cameras and streams being left open when the app crashed or hung on exit.
• Fixed the video display eventually stopping after repeatedly changing capture format, frame rate or depth mode on the same input.

ZED Depth Viewer

• Added support for the new NEURAL depth precision setting (sl::InitParameters::depth_precision). INT8 trades a small amount of accuracy for a faster depth runtime and a lower memory footprint; FP16 remains the default.

ZED360

• Fixed a camera failing to start when it opened with a warning, which was treated as an open failure and left the camera out of the fusion. Warning codes are now accepted at open and during capture, and the camera runs normally.
• Fixed a crash when subscribing to body-tracking senders over the network, typically hit when Jetson senders publish to a Windows receiver and additional cameras are subscribed. A malformed body or object payload is now dropped with a warning instead of terminating the process.

ZED Diagnostic

• Added the -onnx <path> [WxH] command-line option to optimize a custom object-detection ONNX model ahead of time, so enableObjectDetection() with OBJECT_DETECTION_MODEL::CUSTOM_YOLOLIKE_BOX_OBJECTS, CUSTOM_RFDETRLIKE_BOX_OBJECTS or CUSTOM_BOX_OBJECTS_AUTODETECT starts instantly instead of optimizing the model on first use. The optional input size (default 512x512) must match the ObjectDetectionParameters::custom_onnx_dynamic_input_shape used by the application.