1 Introduction

Since the years 70, astrometric observations of the Sun are done with the solar astrolabe at Calern station - Côte d'Azur Observatory ([Laclare et al. 1980]; [Chollet 1981]) to estimate its radius and the eventual variations of this parameter. The quality of the visual observations which give, since 1978, the best evaluation of the solar radius, urged us to increase the quality of the instrument and to begin studies in order to automate the observations. This is partially done now: the modern astrolabes use zerodur prisms and the observations are made using CCD video camera ([Sinceac 1998]; [Sinceac et al. 1998a]).

These observations have a great importance in many scientific domains. Firstly, the solar radius is an important parameter for the physical study of this star and its variations, if they are real, are not only important for the solar physics but also for celestial mechanics and the planets motion in relativistic gravitation theory.

In other way, if the quality of these observations can be extended to other objects as stars, planets and small planets, a connection between the existing spatial reference frames could be possible. Few experiments now in progress shows that these objectives are attainable ([Kurzynska 1997]; [Popescu 1998]). A first test done by the authors, with intensified video CCD camera, shows that objects of magnitude as high as 9 or 10 may be observable now.

But, it is not sufficient to have the best instrument, if the subsequent analysis of the observations is not carefully done. The aim of this work is to remove disruptive effects on the observations, which can be resumed as:

The apparent or observed solar radius is always smaller than the true one.

It is well known that a center to limb effect exists on solar images, which shows a decreasing of the solar light intensity ([Allen 1973]; [Pecker & Schatzman 1959]). In the evaluation of the observed radius, this effect is amplified and modified by atmospheric turbulence and transmission.

The true radius needs to be defined. This necessary definition will be derived from one of the adopted parameters of the solar images model. Indeed, the amount of data obtained by CCD observations allows us to construct a model which depends only on 5 parameters. More, we will see that the adopted model gives a proof of the reality of the disruptive effects mentioned above. Taking account of the different effects acting to construct the final image on the CCD, we will define the solar radius as:

the semi-diameter of a large circular source of light emitting with constant intensity,

that is without center to limb effect. We precise here that the observed diameter is always a vertical one and that the observations are made at constant zenith distances.

In order to be clear, a description of observations methodology should be presented briefly (for details, see [Sinceac 1998]; [Sinceac et al. 1998a]). This description will help to understand the way followed to solve the problem.