Abstract

DIAZ SALAZAR, Aldo A  and  KURKA, Paulo R. G.. Computer vision methods for automotive applications. Tecnia [online]. 2020, vol.30, n.2, pp.74-81. ISSN 0375-7765.  http://dx.doi.org/10.21754/tecnia.v30i2.801.

Recent advances in computer vision are leveraging many technological developments in modern industry and automation. In this tutorial, it is presented a review of computer vision methods and applications relevant to the use of cameras as measurement devices in the automotive industry and robotics. The methods include algorithms for edge and ellipse detection, camera calibration, 3-D reconstruction and stereo vision. The applications are elaborated through simulations of three key problems: Detection of rims in automotive wheels; estimation of the calibration angles of vehicles and; trajectory reconstruction using stereo vision. These applications allow to demonstrate the potential of vision-based technologies in solving complex engineering problems in an automated fashion using cameras as sensors. As a result, three general purpose methodologies are proposed for solving problems of industrial need that would serve as guidelines for further developments in current and other related areas.

Keywords : Computer vision; wheel rim detection; vehicle calibration angles; stereo odometry.

        · abstract in Spanish     · text in English     · English ( pdf )