Development of a Stereovision-Based Tracking System by Applying a Hybrid Autofocus Technology

Abstract

This article presents an intelligent surveillance system based on stereovision for dynamic tracking of moving objects with autofocus and lossless zoom capabilities. The core technology of the system combines a modified stereovision algorithm with an autofocus search method. First, in order to reduce the effect of changing the lens focal length, the modified stereovision algorithm provides an obvious advantage in accuracy of depth estimation. Then the estimated depth not only can be used for adjusting the appropriate zoom level of cameras but also for shortening the search range of focal points. Moreover, the autofocus search method can simplify the image sharpness curve to a single peak. Therefore, by using spline interpolation, the focus position can be efficiently and accurately estimated. Finally, experimental results show that the proposed scheme has the ability to improve the accuracy of depth estimation and provide more clear images of moving objects.