Can radiowaves penetrate the ground and find aluminium and be reflected to later be analyzed and tell us what lies under the ground of a planet? Can radiotelescopes analyze these radiowaves to know what lies beyond some centimeters or even meters of planets subsurfaces?
 A: Radar techniques are regularly used to probe solar system objects at the Goldstone and Arecibo sites. See https://en.m.wikipedia.org/wiki/Radar_astronomy
I am no expert, but according to this source http://echo.jpl.nasa.gov/introduction.html radar imaging techniques can be used to construct 3 dimensional geological models of asteroids and to say something about the surface composition and density.
Radar signals will penetrate into the surface and the composition and density will affect the reflected signal. Unfortunately, the bulk structure and roughness of the surface also affects the reflected signal, so disentangling the two is difficult.
This paper by Thompson et al. (2006) discusses how long wavelength ($>30$ cm) radar can probe the "physical structure and dielectric properties of planetary regoliths... up to tens of metres below the surface".
http://onlinelibrary.wiley.com/doi/10.1029/2005JE002566/full
Of course, it helps if you can get closer to the target, since the returned signal scales as the reciprocal of the fourth power of distance. The MARSIS instrument on the Mars Express satellite was specifically designed to use ground penetrating radar to investigate the subsurface composition of Mars. https://en.m.wikipedia.org/wiki/MARSIS
A: They can, as per @Rob Jeffries' very nice answer. Just wanted to add something, about generally the wavelength/frequent dependence of penetration, and how imaging is done. 
See the Wikipedia article on ground penetrating radar (GPR). https://en.m.wikipedia.org/wiki/Ground-penetrating_radar. At 30 cm wavelength for wet clay and dry sand you penetrate about 3 ft and 6 ft repectively. As you wavelength goes higher, i.e., lower frequencies you will penetrate more. At 100 MHz or 3 m wavelength you penetrate about 20 and 60 ft respectively. As you go to higher wavelengths you need bigger antennas, and your resolution decreases. 
The MARSIS radar operates at 1.8 to 5 MHz center freqs, using two 20 m antennas, with 5 watts out of each, and 1 MHz bandwidth. At 2 MHz the wavelength is 150 m, and it penetrates more than the above numbers, typically, and the goal was to go down to the permafrost. With such a high wavelength the structures searched for were at km scale. 
One can get better resolutions and for 2D or 3D object mapping, i.e., imaging, it is done with some version of synthetic aperture radar, in essence a version of tomography. Real apertures can also be use but you need huge antennas. See https://en.m.wikipedia.org/wiki/Radar_imaging. As the source or object moves it measures Doppler change to improve the cross range resolution (yes, looking at an angle), while the range resolution is done with higher bandwidths (of course limited at lower freqs). They do multiple scans so as to tie the pieces together. Ground penetrating radar will do that for the best performance. They will also do that for surface features mapping. Note, the synthetic aperture gets formed by motion, so it is great orbiting not too far in space as the movements can be fairly precise. For terrestrial radar telescopes it uses the motion of the object mapped. For ground based mobile GPR it needs to track its trajectory carefully. 
