Accuracy assessment is an integral part of LULC classification studies. Usually studies use some numbers of sampling points or ground truth points or even random points to compare the LULC classification and the real land cover. Is there any study available which has used a buffer area (any area) instead of a point to compare the classified LULC to the actual land cover on the ground. Please share the reference for any such studies, would love to know about it.
Dear Bhattacharjee,
The classics methodologies for RS accuracy assessment are:
CONGALTON, R. G.; GREEN, K. Assessing the accuracy of remotely sensed data: principles and practices. New York: CRC Press, 2009. V.2, p.183.
CONGALTON, R.; GREEN, K. Accuracy assessment and validation of remotely sensed and other spatial information. International Journal of Wildland Fire Scientific, v. 10, p. 321–328, 2001.
You can also do it by spatial agreement (area-by-area) of your classified mapping and the reference data. I suggest you take a look at this article:
OLIVA, P.; SCHROEDER, W. Assessment of VIIRS 375m active fire detection product for direct burned area mapping. Remote Sensing of Environment, v. 160, p. 144–155, 2015.
Regards,
- Badges
- Science topic
- Similar topics
- Geoscience
- Remote Sensing