Much literature I've seen often focuses on the stria or areola density near the valve margin. For which you can measure the stria in two different ways, chord count vs. the circumference count. The chord count is similar to your lines, where you measure the ornamentation within a 10 micron scale bar. The circumference count you would count all the way around the valve and calculate the stria density based on the circumference. Either way are good, the chord count is the traditionally used method, but it is best to note which method to use. 

Sorry to digress, but of your question I would count your red line, because then you are discussing the the areola density for a stria. It would be a consistent unit of ornamentation to measure within. I think ultimately it is important to note which method you use and try the best to understand the methods used by the authors of your taxonomic resource.

It is an area of uncertainty and probably not yet harmonized because the methods of quantifying ornamentation vary with the diversity in ornamentation.

David, thank you for the answer. You mention that "methods of quantifying ornamentation vary with the diversity in ornamentation". I know only those papers:

Collins, G.B. and Groetsch, C.W. (1981). An elementary mathematical problem arising in diatom taxonomy. International Journal of Mathematical Education in Science and Technology 12(1): 121-124.

Theriot, E. (1988). An empirically based model of variation in rotational elements in centric diatoms with comments in ratios in phycology. Journal of Phycology 24(3): 400-407.

Could you recomend any other papers where another methods of areolae count for centric diatoms are described?