Visualization of Obstacles on Bird's-eye View Using Depth Sensor for Remote Controlled Robot


The bird’s-eye view system is one of the image presentation systems for remote control of robots. However, when there are obstacles that are higher than the floor, there is a problem that the appearances of them are incorrectly shown on the bird’s-eye view image due to the image distortion. That is because the image is generated under the assumption that all objects are present on the floor. In this paper, we propose a method to visualize correctly the areas where obstacles are present on the bird’s-eye view image by using 3D measurement data. Specifically, the information of the robot’s distance from obstacles measured by a depth sensor is superimposed onto the bird’s-eye view image as points with colors obtained by fish-eye cameras. The bird’s-eye view image generated by the proposed method can improve the visibility of the surrounding environment.

Proceedings of the International Workshop on Advanced Image Technology 2017