Two approaches are suggested for recording the continuum corona in the range < 1.4 Rsun. They are different from the classical coronagraphic ones. Current state in the thin film technology allows discussing a new generation coronagraph with a variable transmission of an entrance aperture. The estimated coronagraphic factor is 2 orders of magnitude higher compared to a Lyot-type coronagraph. Another approach is based on the use of total solar eclipses at near-Mercury orbits. The instrumental background is decreased at least 3 orders of magnitude. That allows using a more simplified optical sketch.

"Weak" magnetic-field diagnostics in faint objects near the bright solar disk are discussed in terms of the level of non-object signatures, in particular, of the stray light in telescopes. Calculated dependencies of the stray light caused by diffraction at the 0.5-, 1.6-, and 4-meter entrance aperture are presented. The requirements for micro-roughness of refractive and reflective primary optics are compared. Several methods for reducing the stray light (the Lyot coronagraphic technique, multiple stages of apodizing in the focal and exit pupil planes, apodizing in the entrance aperture plane with a special mask), and reducing the random and systematic errors are noted. An acceptable level of stray light in telescopes is estimated for the V-profile recording with a signal-to-noise ratio greater than three. Prospects for the limb chromosphere magnetic measurements are indicated.