Scaling allows to reduce the camera resolution by an arbitrary non-integer factor during rectification. The camera raw images stay at their original size, but the rectified images, disparity map and point map will be scaled by the specified factor to improve stereo matching runtime. This allows you to choose your own tradeoff between image resolution and performance.



A positive real number between 0.1 and 1.




Setting a new Scaling factor immediately clears and resizes the affected image nodes.


As Scaling only affects the rectified images you might set a new Scaling factor and rerun ComputeDisparityMap without capturing a new image pair! You could therefore use Scaling for fast object detection in low resolution, and then perform measurements in higher resolution by setting Scaling to 1 without the need to capture an additional image pair.

Effects on other parameters

You might want to adjust DepthChangeCost, DepthStepCost and the PostProcessing parameters for optimal 3D data results after changing the scaling factor. The disparity range is automatically adjusted to keep the originally configured measurement volume. You can read the actual disparity range that is used on the scaled images from the corresponding scaled nodes.