Unmanned aerial vehicles (UAV s) are relatively new inventions, designed in an era where the available technologies did not allow them to be properly built and operated. Recently, with the technological progress, UAV s are no longer an idea, becoming a reality. Nowadays, the use of computational techniques that allow such devices to execute tasks autonomously has been an intense research field. Typically, such techniques are based on computational algorithms executed in processing units, embedded or not in a drone. Although, in most applications, the current state of the art is capable of meeting the requirements associated to the computational processing capacity required to execute these algorithms, it is observed that the flight autonomy time still represents a big challenge, due to technological limitations, associated primarily to the motors efficiency, the energy storage capacity in batteries and the thrust-to-weight ratio. In this way, weight reduction techniques of the aircraft that allow an autonomous and reliable flight are indispensable. The IMAV (International Micro Air Vehicles Conference and Competition) is one of the most important international competitions aimed at fostering the development of key technologies to the development of intelligent drones. The requirements of the competition demand the development of autonomous, light-weight and energetically efficients systems. The present work belongs to this scope, aiming to contribute to the development of an autonomous UAV to IMAV2018, that will take place in Melbourne, Australia, during November 2018. In particular, this work aims to contribute to the robust navigation of the system through one of the elements that compose the IMAV2018 s challenge, focusing on the development of image processing algorithms capable to implement a line-follower, besides recognizing and avoiding possible obstacles that may be on the UAV path. The developed computational algorithms are described and evaluated. The results shows that the techniques used in this project has great potential to perform the autonomous navigation of the drone, based on computational vision.