| dc.description.abstract | The study and use of robotics has been a subject in the academic community since before the 1970s. The use of robotics for there solution of everyday problems has become common in both the academic and industrial worlds, and it is possible to carry out numerous activities with greater efficiency and precision, working indiferente áreas of research, such as medicine, security, construction, among others. Considering this, it is used a mobile robot, created with a Raspberry Pi for processing and controlling the robot. Encoders and webcams sensors are used for monitoring the engines and analyzing the environment, respectively. The aimis to develop a stereo vision module to perform scene depth calculations, allowing the detection of possible obstacles and, thus, facilitating the system’s decision making to avoid these obstacles. In order to make this possible, a methodology was defined to allow the completion of the specific objectives, namely: construction of a support for a stereo pair, realization of the geometric calibration of the stereo pair, definition of the standard test scenario, implementation of na algorithm for depth calculation and, finally, carrying out experiments in the standard test environment, in orderto compare the measurements found by the robot, generating graphs and tables for analysis of the results. Finally, tests were carried out to validate the functioning of the implemented system. In addition to the tests with the robot, a stereo system validation test was carried out with a group of images from the University of Middlebury’s stereo dataset. Withthat, it was possible to create graphic sand images, showing the functioning of the robot in relation to each defined test. After analyzing there sults obtained, it was concluded that the developed robot allows movement in na environment with obstacles and also has na acceptable accuracy forth emeasurement of objects positioned up to 50 cm away from the robot. | |
