The physical mechanism that governs the incipience and development of leading edge cavitation on hydrofoils is still not fully understood. It involves several parameters such as the pressure level, the nuclei content, the surface roughness and the boundary layer with strong interaction between each other. In the present study we have carried out an experimental investigation to analyze the role of these parameters on the vaporization process. The case study is the water flow over a 2D NACA0009 hydrofoil placed in the test section of the EPFL high speed cavitation tunnel. Wall pressure measurements as well as high speed visualization are carried out for a large variation of hydrodynamic conditions and surface roughness. First the analysis of nucleation effect led us to distinguish between homogeneous nucleation, characterized by free micro bubbles within the liquid stream, and the surface nucleation, which consist of micro sized volume of gas trapped between the liquid and solid surface. The later, which results from a lack of wettability, depends on the surface roughness as well as the physical and chemical properties of the solid (hydrophobic or hydrophilic surfaces).Experimental evidence of the existence of such nucleation is clearly made in both still and flowing liquids. Our experiments have revealed that surface nucleation plays a major role in the cavitation onset and development over lifting hydrofoils. A novel mathematical model is proposed to predict the growth of a single nucleus from a surface roughness element with respect to surrounding pressure and surface tension in still liquid. For moderate Reynolds numbers and low incidence angles, we have demonstrated that the liquid may withstand negative pressure, with absolute values as -1 bar, without vaporization. We have also investigated a particular development of periodic bubble cavitation. We have demonstrated that such bubbles originate from surface nucleation and described the physical process of growth and advection of single surface nucleus. The generation frequency of such bubbles was found to vary substantially from one nucleation site to another. Indeed, besides the pressure level, the size of non wetted volume in the hydrofoil surface is a dominant parameter in the vaporization rate. As the traveling bubbles evolve on the hydrofoil surface, they interact strongly with the boundary layer and outer flow. According to pressure signals, we have shown that while the bubbles grow, they remain slightly above the hydrofoil surface and a moving 3D boundary separation is evidenced in their back. We have also shown that as the pressure level is reduced below a threshold value, which is different from one nucleation site to another, the periodic bubble cavitation turns into attached spot cavitation in a continuous way. The role of the boundary layer state in the cavitation onset and development has been also investigated. We believe that for smooth surfaces (i.e. low Reynolds number based on roughness height), the onset of attached cavitation requires a laminar separation of the boundary layer as already stated by several authors. Nevertheless, this condition is no more required when surface nucleation occurs. In this case, the cavity is continuously fed with vapor generated at its detachment. Finally, the cavitation occurrence on a single indentation on the hydrofoil surface has been investigated. We have demonstrated the significant role of sheer stress on the vaporization process and questioned the pressure based criterion for cavitation incipience. We have shown that for flowing liquids, the criterion based on the maximum tensile stress is more appropriate.