I think it should be around 5% of the total area

I think this video will help you. https://www.youtube.com/watch?v=enrWhgIOqWQ

You might find this helpful. http://www.fao.org/3/X6831E/X6831E12.htm#TopOfPage

There are several techniques, based on objectives, stand diversity in species, age, height, density, etc. Sometimes topography or soils are a helpful determinant in separating forest types, such as bottomland or riparian forests as compared to upland forest types. Advice from a forest statistician can be helpful in designing a sampling plan for research level work. Unless the forest type is pretty uniform, stratifying the area into at least the primary and sometimes secondary forest types, and then random sampling within each forest type is a fairly common approach foresters use. Sometimes foresters divide the forest into forest stands or compartments which are fairly uniform in forest type, age, etc.

Foresters have various tools that help evaluate and measure forest stand attributes. For example, a prism is often used to help estimate stand basal area, and the prism uses a variable plot size to sample small trees close to plot center, and capture all the larger trees at distances that vary with tree diameter. Normally foresters and soil scientists use aerial photos to help divide or stratify area into similar units, and then collect a number of samples from each unit needed based on the area of that stand or unit type.

There are books and guides written on the subject, portions often taught in forestry schools. If conducting research, you pronbably want to be consistent with a standard method, to save alot of explaining.