Comparison of vegetation indices acquired from RGB and Multispectral sensors placed on UAV

This manuscript presents a comparison of Normalized Difference Vegetation Index (NDVI) obtained with multispectral cameras versus four indices obtained from RGB sensors for the identification of soil and vegetation in images captured with an unmanned aerial vehicle. This comparison was made using the NDVI as ground truth, obtaining 2 classes of data that would be compared later to the other indexes by counting the pixels corresponding to each class. In the case of the RGB indices, the average was defined as the center of the data and as the cut-off point of both classes. The results of this investigation indicated that it is possible to identify the same spatial patterns using RGB indices, where the TGI index shows the best behavior. However, despite the fact that the pixel count showed similar results, the visual inspection of the results indicated that the RGB indices presented errors when identifying the vegetation, especially in the zone of the row. This indicates that to delimit with precision the areas corresponding to vegetation and soil it is necessary to use more complex clustering techniques.