Estimativa de orientação com uma bússola visual baseada em cores

Abstract

Humans and animals make use of many senses to interact with the environment around them. Computers, in turn, interact through input and output devices such as speakers, keyboards, and monitors, and their interactions have many more limitations. But currently, devices like cameras and microphones have been added to computers, increasing their interactivity. With this advancement, there are autonomous robots equipped with sensors, such as sound, touch and vision, the latter through one or more cameras. A robot, to be considered autonomous, must make decisions without human intervention and, in order to have excellence in what it does, must be provided with a good form of guidance. This work presents an alternative visual compass to stipulate the orientation of robots and autonomous vehicles, so that they can get around in a scenario (environment) and do the work in which they were built to perform. The proposed approach works with 360o perspective of environments, in which information is extracted from the color changes that occur. With color change counts, the process is repeated with images received from the robots or vehicles in motion, and the color changes of this image with those stored from the 360o perspective. The result of the color change comparison is used to estimate the degree of similarity between the images and thus to present the angle at which the image of the moving robot is present in the 360o perspective. Robots or autonomous vehicles can use this value presented in their decision making.