Implementation and Verification of a Virtual Testing System Based on ROS and Unity for Computer Vision Algorithms

Abstract

With the development of artificial intelligence technology, Computer vision algorithm is playing an increasingly important role. Computer vision algorithm testing is a vital link to ensure the safe and reliable operation of agents. However, the traditional testing methods based on real scenes provide single samples and are challenging to obtain ground truth, which makes it inefficient to test computer vision algorithms. To solve this problem, the testing method of computer vision algorithms using virtual scenes instead of real scenes has been applied. In this paper, Unity and robot operating system (ROS) are selected to build a virtual testing system named URCV for computer vision algorithms. The feasibility of the system and the influence of virtual scene elements on the testing of the monocular ORB_SLAM2 algorithm are verified, including the rendering path of Unity's RGB camera, texture accuracy, and illumination model.