A fully automatic real time system for the characterization of automotive headlamps Alessandro Bevilacqua, Alessandro Gherardi and Ludovico Carozza ARCES - University of Bologna Via Toffano 2/2, 41025 Bologna - ITALY {abevilacqua,agherardi,lcarozza}@arces.unibo.it Abstract—In this work we describe a fully automatic real time system, based on a CCD camera, for extracting the geometric pa- rameters of the light proﬁles projected by automotive headlamps. Our system is capable to automatically adjust to a wide range of light sources of different power allowing, through a locally adaptive thresholding algorithm, the automatic detection of the dark-light border of a beam proﬁle as it would be perceived by an experienced human operator. The extracted beam proﬁle is processed in order to compute the position of geometric reference points provided by current regulations (i.e the “elbow” point). In order to validate the measuring system and to achieve accurate ground truth data, the headlamp has been mounted on a special three-axis numerical control unit built on purpose. Experimental results, carried out on a large number of different headlamps, show that our method is able to achieve accurate measurements in compliance with current regulations. Finally, it is worth remarking that our solution is fully automatic and it just requires a simple setup procedure. I. I NTRODUCTION The automated geometric and photometric characterization of light sources in transportation systems is stirring up the industrial research and it is stimulating new proposals in reg- ulation requirements in order to improve safety standards. An important issue is how to perform accurate measurements of the light distribution for these light sources that are outstanding for the road safety. Nowadays, automotive industries and reg- ulation agencies are interested in developing automated tools for the assessment of car headlamps beams properties, in order to prevent drivers in the opposite direction from being dazzled by dipped headlights as well as to check that main headlights direct their beam in the proper direction. Systems currently available on the market require awkward set up procedures and the continuous intervention of human operators. Different techniques and technologies are used. Usually a large number of sensors (photogoniometers, lightmeters, etc.) often arranged in complex and expensive systems are necessary in order to get the accuracy needed [5], [8]. In this paper, we present a fully automatic real time system capable to perform accurate measurements of geometric and photometric regulation parameters for both driving and passing headlamps light beams. The luminous proﬁle is projected onto a panel and acquired and processed by a CCD camera. This research has been partly granted by SIMPESFAIP SPA. This technology is quite cheap and allows the development of algorithms capable to cope with two basic issues. The ﬁrst one is to guarantee the adaptability to a wide range of luminous power, adjusting the radiometric resolution of the acquisition system in order to work with a wide variety of light sources. Headlamps yield a very high dynamic range and CCD based systems need to take into account the correct exposure time in order to achieve a reliable representation of the light distribution so as to perform accurate photometric measurements. The second one is to mimic the response of the human eye even in a highly contrasted and untextured scene such as the one generated by beams projected on a white panel. An automatic segmentation algorithm, based on a locally adaptive thresholding, has been devised, to detect the line corresponding to the light-dark border. The experiments prove that our method works to extract the proﬁle as being perceived by the human visual system with the accuracy required by present regulations of the highway code. In Sect. II, the system is presented. In Sect. III, the al- gorithms devised to perform the characterization of beams proﬁles are described in details. In Sect. IV, the experimen- tal results are presented where measurements achieved after characterizing headlamp beam proﬁles are discussed. Finally, Sect. V draws conclusions. II. THE SYSTEM Pictures of the developed automatic measurement system are shown in Fig. 1. The industrial prototype hosting the automotive headlamps is shown in Fig. 1(a), while Fig. 1(b) shows the Optical Projection System (hereinafter, OPS). It is (a) (b) Fig. 1. Pictures of the system: in (a) the NCU, in (b) the OPS. equipped with a Fresnel lens (whose optical parameters are I2MTC 2009 - International Instrumentation and Measurement Technology Conference Singapore, 5-7 May 2009 978-1-4244-3353-7/09/$25.00 ©2009 IEEE 36