The accuracy really depends on what you are examining. For example, if you are looking at large-scale features, than coarser resolution is probably very accurate, but if you are looking for small scale features than higher resolution is needed because coarser resolution will not be accurate. So your assumption is correct in the sense that the lower the resolution, the less accurate the imagery will be for smaller-scale feature identification.

In terms of achievable spatial resolution with drones, PIX4D has a calculator you can use that can tell you the maximum ground sampling distance (GSD) - spatial resolution - based on your camera dimensions and the altitude of the drone being flown.

https://support.pix4d.com/hc/en-us/articles/202560249#gsc.tab=0

For a more general publication on resolution vs. accuracy you can look at:

Chen, D., Stow, D. A., & Gong, P. (2004). Examining the effect of spatial resolution and texture window size on classification accuracy: an urban environment case. International Journal of Remote Sensing, 25(11), 2177-2192.

Dear Deepak,

Your question is not entirely clear to me, because I don't know what you mean by "spatial accuracy". Is it the accuracy with which you will detect a particular object in the scene, or do you mean the accuracy with which you'll be able to define the geographical location of that object in the scene?

http://www.edc.uri.edu/nrs/classes/NRS409/RS/Lectures/HowRemoteSensonWork.pdf

http://gis.humboldt.edu/club/Images/Documents/lecture7.pdf

Thank you all for your answers and recommendations. I would like to elaborate on my question a bit further.

So, I am using a drone-mounted sensor (spectrometer) which can be simplified as a single pixel camera with an ultra-high spectral resolution. The footprint size of this sensor (or spatial resolution) is about 1.5 by 3.0~m (for my flying configuration). What I am interested in is to know if I can validate the computed geographic location of the footprint (expected uncertainty in computed footprint geographic location is about 15~cm). For this, I need to be able to pick some geographic references using the sensor. However, I think that the sensor spatial resolution (1.5 x 3 m) would have some limitation on the spatial accuracy of the picked geographic reference. I want to be able to define this limitation. An example: Using sensor with a spatial resolution of 1~cm, I would be able to pick the center of a ping pong ball with higher confidence as opposed to using a sensor with a spatial resolution of 5~cm. Hope it makes sense.

Thanks.

Hello Deepak,

I do not know about drone applications, but for GPS a paper of a former co-worker of me might be interesting: https://www.ncbi.nlm.nih.gov/pubmed/22829049 

Good luck!

Gijs