OpenCV
OpenCV is currently one of the most practical computer vision foundations inside the xtan ecosystem. xtan uses OpenCV for real image handling, stereo-related processing, calibration-oriented workflows, and early geometry-aware perception steps that need a stable and proven library base. At the same time, OpenCV is not the final destination. The longer-term direction points toward a more specialized xtan stack with stronger control over geometric logic, sensor-near processing, and RAW-oriented image concepts that go beyond standard library defaults.
OpenCV as the current working base
OpenCV fits xtan well in the current phase because it provides a broad toolkit for image processing, camera workflows, feature analysis, and stereo-related experimentation. That makes it useful for getting perception pipelines running in a practical way while xtan continues to define its own deeper architectural direction.
Why OpenCV matters right now
xtan depends on real cameras, geometry-aware interpretation, and reliable low-level perception steps. OpenCV supports that work by offering a stable bridge between image acquisition and higher-level spatial logic. It is especially helpful when experiments need to move quickly from raw sensor input toward calibration, tracking, and structured scene understanding.
Why xtan moves beyond OpenCV over time
The future xtan path goes further than a general-purpose computer vision library. A dedicated stack allows tighter control over geometric reasoning, domain-specific stereo logic, and improved RAW handling closer to the sensor data itself. For xtan, OpenCV remains highly useful as a present-day implementation layer, while the longer-term goal is a more specialized perception foundation shaped around xtan's own geometry and imaging requirements.