Comparison of Vegetation Indices of Crops from Satellite Images PerúSAT-1 and UAV

Vegetation indices mainly identify plant health conditions, such as bioindicators monitoring crops using satellite images. The most used indices in the agriculture sector are Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and Soil Adjusted Vegetation Index (SAVI). These indexes can be obtained from the visible and infrared bands from sensors aboard platforms such as satellites or Unmanned Aerial Vehicle (UAV). The present research was carried out on an area of horticultural crops in the Carabayllo, Lima (Peru) district. Images from the Naomi sensor of the PeruSAT-1 satellite and the MicaSense Altum multispectral sensor of UAV are used, where PeruSat-1 is an Earth Observation Satellite that belongs to the Peruvian government, and it is operated by Comisión Nacional de Investigación y Desarrollo Aeroespacial (CONIDA). These images from the study area were obtained at different wavelength amplitudes for each spectral band. The best environmental conditions were considered, such as the low cloudiness and proximity of the capture dates of both sensors, to reduce the spectral variations of the crops and their phenological state. However, it is important to convert digital levels to surface reflectance levels for the same radiometric scale for spectral response analysis with both sensors. The images underwent radiometric and geometric correction, then the generation of vegetation indices to obtain segments with similar spectral characteristics for each class, from which the values were extracted using random points (50 points per class). The results are analyzed using a Pearson sample linear correlation, where it is observed that the vegetation indices obtained from PeruSAT-1 and UAV images show a positive correlation around r > 0.8 for the variables of NDVI, SAVI, and EVI.