The model for optimal design of robot vision systems based on kinematic error correction

An active vision system is a robot device for controlling the optics and mechanical structure of cameras based on visual information to simplify the processing for computer vision. In this paper, we present a kinematic model for the optimal design of such active vision systems. We first build a generic kinematic model for robot structure error analysis using a Denavit–Hartenberg transformation matrix, differential changes for this transformation matrix and link parameters. We then extend it to analyze an active vision system using algorithms for estimating depth using stereo cameras. This model is generic and is suitable for analysis of any active vision system. Since we can employ it to analyze errors based on variations of link parameters when we use an active vision system to estimating depth, we can combine it with a cost-tolerance model to implement an optimal design for active vision systems. In this way, we can not only save manufacturing cost and implement design for manufacturing (DFM) but reduce or avoid calibration work for an active vision system. Our algorithm also works for a binocular head and on even more complex tasks. Based on our approach, we have created a software tool that functions as a C++ class library. We also demonstrate how to use this software model to analyze a real system TRICLOPS, which is a significant proof of concept.