Remote sensing with Land Cover Classification
LULC Classification and Ground Truthing In this exercise, we are using imagery located at Pascagoula, MS, to digitize the land cover surrounding the area and classify it by land use and description. In order to do this, I decided to use a land cover legend to digitize the different land types using a polygon feature class. The different land types were given different colors to differentiate the symbology of the land cover. I then used 30 sample points called "truthing points" to test the accuracy of my land cover results. I simply dropped 30 points on my image and then exported them to google earth using a kmz tool. I then zoomed in on each sample point and viewed the google Earth Image in 3D to see fi my land cover was classified accurately or not. I then coded my results in my truthing feature class to identify which points were accurate and wh9icvh ones were not. Close Up View Of one Sample Point The final map below shows the sample points in accuracy in comparison t