Flexibility in parameters is a high note for photogrammetry. Literally using any camera in the world, you can achieve a 3D mesh result with texture. But sometimes it's not good enough to just achieve a result - you want the best result. 

Technology is always changing, especially in regards to computing power, which does directly affect photogrammetry

results. Likewise camera technology changes in regard to megapixel resolution, which nets us more points in the photogrammetry point clouds. At a basic level, photogrammetry algorithms analyze every pixel of every photo taken, and counts the occurrence of references to the neighbor pixels' absolute differences. This is part of how the X, Y, Z translation position of any given vertex in the point cloud is determined.

Photogrammetry scan with color texture on.

Photogrammetry scan with color texture off.

So although megapixel is a formidable factor in the end resulting mesh, I wouldn't go there immediately. There are some basic fundamentals that have to be taken into account prior to just parking a truckload of megapixels into your project. For example, the lighting in which you use on your project is highly consequential in the quality and consistency of your mesh and texture. If there's not enough light, the image becomes degraded with noise, and your resulting 3D mesh will similarly have that noise baked in. Likewise, an overexposed image with have highlights that occlude the actual data you were trying to photograph, and it results in noise as well in the 3D mesh.

A camera's settings determine how it interacts with light, and so the study of light is very important, and to observe what's happening with the light when you setup a scenario. It's often a desire to have your 3D mesh with a flat lit texture, and that can be challenging to setup. The lighting has to be soft and spread out.

Photogrammetry scan of a strawberry.

Each picture can be used a few different ways in photogrammetry. It could be alignment, detail, or coverage a name a few. A wider framed shot with a good depth of field would be helpful for alignment whereas a tighter framed shot with a narrow depth of field would establish quality.