Object-Based Rate Control for 3D Vision Reconstruction

This paper presents two types of methods for controlling the image acquisition rates of multiple cameras in a 3D vision applications. Both control methods adjust their image acquisition timing based on the object states estimated from the captured images. The first method synchronously controls the image acquisition rates of the cameras based on the current object states in the world/physical coordinate. The second method adjusts the image acquisition rates by the object states in the image coordinate. In this case, each camera independently adjusts its own acquisition rate based on the information obtained from the captured images. The proposed methodologies have been experimentally tested using a two-camera setup within the MATLAB programming environment. Experimental results show that, compared with fixed-rate methods, the proposed approaches achieve satisfactory performance and reduce computational cost and data storage.