Alternative approach for automatic evaluation of traffic sign visibility

To read the full-text of this research, you can request a copy directly from the authors.


In the present paper we propose an automated low cost method for evaluating the visibility of traffic signs. Since the traffic signs may present difficulties to be seen at night and whose main propose is to advice the motorists, we decided to do the measurement at night, within a car and using the headlamps as illumination sources. In this way we evaluate the surface reflectivity of the sign and the influence of bad orientation or positioning of the sign relative to the vehicle headlamps or road. The visibility parameter of a traffic sign is then calculated from a sequence of images of the sign captured with a colour digital camera at different distances sign-vehicle. The key point for an automated process is to detect and recognise automatically the sign from the captured images. This is done by a software which uses Support Vector Machines (SVM) as a novel classification technique. As example, this technique has been applied successfully over three different signs with three different degrees of surface deterioration. The influence of relative orientation vehicle-sign is also analyzed by a sign placed on a curved road.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

ResearchGate has not been able to resolve any citations for this publication.
Full-text available
Recent availability of video-cameras with CCD-type sensors (charge coupled device) has proved to be particularly stimulating for all those applications requiring photometric measurements, above all for the measurement of luminance values related to the physical and technical qualities of a built environment. This method allows the instantaneous capture of an image, thus enabling collection of luminance values relating to the points of measurement; this in turn leads to the evaluation of luminance distribution and lighting levels of the surfaces that make up the environment.Setting up this system requires the following basic configuration: a photopic filter V(λ), an optic interface, a computer equipped with an appropriate card for the capture and digitalisation of the acquired image (the “frame grabber”) and, finally, suitable software for the processing of collected data.In this article a detailed description of this acquisition system is reported, and subsequently a report on the procedure adopted for its calibration so as to enable the capture of relevant photometric values. Final analysis and validation of results are carried out by means of field test. A case study of CCD photometer application has been then performed using a basic software tool autonomously developed to evaluate indoor lighting level; the luminance map of a diffuse light source has been used as “input” data for the developed software, and the “output” data, i.e. illumination levels, have been then compared with measured values.
Full-text available
This paper describes the methodology developed for characterizing a commercial charge-coupled device (CCD) camera as a luminance meter for analyzing lighting systems and especially for measurements in road light plants. Today, several luminance meters based on commercial CCD cameras are on the market. They are very attractive for the lighting engineer: The availability of a high number of closely spaced small detectors (pixels) on a single chip permits analyses almost impossible with a traditional luminance meter. These commercial-industrial CCD cameras are sold at prices lower than scientific grade ones. They are factory equipped with a dedicated filter to reach the correct photopic sensitivity V(λ), and they are factory calibrated in luminance SI units. The main counterparts in using these cameras are in the difficulties to define the measurement accuracy and the influence of the environment luminance on the measured values of the framed scene, in the low resolution of their A/D converter (usually 8 or 12 bit), and the higher noise level (usually the CCD chip is not cooled). To reach the measurement accuracy required by lighting norms, it is necessary to characterize metrologically a camera and quantify all the possible external influences which could degrade its performances, in real measurement situations, and which could affect the measurement results. A carefully controlled measurement set up and operating procedure could limit the causes of errors and improve the accuracy of measurements obtained in operating conditions. In this way, the measurement uncertainties might be evaluated completely, and considerations on the results could suggest particular operating practices to limit the causes of error due to measurement setup and environmental conditions.
Conference Paper
The characterization of the photometric properties of a road surface is of prime importance in the design of lighting plant and when the real vision condition should be determined by computer simulation. The measurement could be done in laboratory but the in situ measurement are very interested because it permit to test several zone on the road and there is no mechanical starch on the surface of the sample. This work describes an innovative portable system based on a CCD luminance meter able to obtain uncertainty comparable in traditional laboratory systems.
CCD matrix detector for photometry
  • G Rossi
  • P Iacomussi
  • M Sarotto
  • P Soardo
G. Rossi, P. Iacomussi, M. Sarotto, and P. Soardo, "CCD matrix detector for photometry," in Proc. 8th Int. Congr. Metrology, Besançon, France, Oct. 1997, pp. 326-330.
  • J Casas
J. Casas, Optica, 7th ed. Zaragoza, Spain: Librería PONS, 1994, ch. 20.