AFM microscopy shows the particle/grain distribution at the surface or close to surface. It also depends on the type and materials of your sample. FeSEM or TEM are another microscopes that can show the particle distributions.

Thanks Hajalilou, that is very interesting. How do these techniques quantify the heterogeneity? Do they use any heterogeneity indices?

Its quantity is obtained from XRD pattern by comparing the peak's intensity (your sample) with the reference one (ICDD)

Method for surface heterogeneity measurement of topography using ArcMap software: Create or obtain a digital elevation model and examine the raster stats, which lists the standard deviation of all elevation points in the DEM. SD of elevation values is a commonly-accepted indicator of surface roughness (hetereogeneity) in hydrology studies.

You could also use photogrammetry or laser scanning (depending on the size of your object and the resolution required) to build a 3D model and then measure surface area and many other related metrics. 

Many industries have their own protocols to quantify the degree of dispersion of fillers in polymers.. but there are some ASTM standards too ex: ASTM-D2663-14 for dispersion of carbon blacks in rubbers, ASTM-D5596 in CB in polyolefins etc. 

The problem is on what is you want to measure (quantify) and at what scale. Many "heterogenous" surfaces seem to display self-similarity properties and can modelled (i.e quantified) using fractal or multi fractal processes. This for instance has been applied to the profile of coral communities. The chain-link comparing the length of a chains placed on the bottom with that of a straight line between two points provide two of the points of the "fractal description" of the bottom. Using a chain with different sizes of links provides finer and finer description of the heterogeneity....

At two D, the altitudes of peaks and valleys seems often to follow a multi fractal pattern. This can also be measured but required then 2D or at least several 1D profiles...

I like the simple way of AGRRA methodology. From  10 m long transects  (each meter marked) the maximum Reef Relief is obtained within a 1m radius of the 1m-, 3m-, 5m-, 7m-, and 9- m marks along the transect line. For doing that, at each one of the 5 points measure the height of the tallest coral or reef rock above the lowest point in the underlying substratum. In other words, five positions are selected and within 1 m diameter of these positions, a measure si made from the deepest part of the relief and highest one using a pvc tube marked in cm. After a number of 10 m transects a number of 50 measures have been taken, which are averaged and  Standard deviation is calculaded. With these measures, rugosity can be compared among sites. Excuse me for my imperfect Englsh.

For a very rough scale...on a stretch of rocky shore for example, ecologists have used a technique called "tape and chain".  One lays a chain of fixed length along the stretch of interest following up and down each contour until you run out of chain.  A tape is then pulled taut between the endpoints of the chain.  The ratio of chain length to tape length is used as a crude measure of surface rugosity.  Luckhurst & Luckhurst 1978 is a citation....but there's not much more to the technique than I just described. 

I had heard of folks using a polymer layer on a rock surface that was peeled off, sectioned, and scanned to then use Image J to calculate a similar metric of micro scale heterogeneity.