Abstract
Autonomous robots that replace humans are currently in high demand in the manufacturing sector worldwide. However, these robots cannot be used in unknown environments because most of them are controlled by sequence controls. To solve this problem, cameras are needed as vision sensors that can recognize the outside environment. It would enable industrial robots to more effectively perform the handling tasks by combining the image recognition technique with the feedback and sequence control techniques. In this search, we aim at improving the camera calibration accuracies in order to accurately calculate the shape and position of an object from the image information obtained by two cameras. Then, we focus on the positions of two cameras. We have proposed the positions of two cameras in the past to improve these calibration accuracies. However, we require wide floor ranges for the two cameras in the past reserch. By reducing the floor area of an industrial robot, it is expected to improve productively and investment costs. Therefore, improvement of the position of cameras of the past study is needed. So, in this reserch, we newly propose the positions of two cameras in a small area by taking into consideration calibration accuracies. First of all, we examine the influences of the positions of two cameras for calibration accuracies. Next, we determine the optimal camera position from a simulation based on the modeling image noise obtained through experimentation. After that, we verify the effectiveness of the positions of two cameras by sorting operation accompanied by identification task and robotic grasping task.
