The CDC recently released a report that goes into detail on a study conducted from April 28 - May 3 to understand how many people in Dekalb and Fulton might actually have had COVID-19. They did this by conducting a special kind of survey to determine the prevalence of the virus -- essentially "how many people have or have had the virus".
The results? They estimated the prevalence of the virus in those two counties is around 2.5%. What this basically means is as much as 2.5% of the county populations have COVID antibodies in their system.
So, knowing this, a bit of simple math:
Fulton's population is 1,050,110 and Dekalb's population is 756,558. With a prevalence of 2.5% this means as many as 18,931 people in Dekalb had the virus and 26,253 people in Fulton had the virus when the study was completed on May 3rd.
If you work backwards from the prevalence and the known cases on May 3rd (2184 cases for Dekalb and 2981 for Fulton) you can deduce that for every reported case, there were approximately 8.6 cases in Dekalb and 8.8 in Fulton. Is the factor still 8.6 and 8.8 respectively? At this point we can only guess but since Georgia hasn't really moved much over the last few months in testing, it's reasonable to assume the factor is roughly the same.
Note: As of this writing on July 21 Dekalb has 9,800 reported cases and Fulton has 13,379 reported cases
A little more detail on how the study was conducted
Data like this does a great job of helping us understand how widespread COVID-19 actually is.