Abstract
The management conditions of the sugarcane activity result in an increase in the risk of fire. The environmental and social impacts led to the implementation of legal instruments aimed at gradually eliminating the burning of sugarcane straw. The incidence of fire in sugarcane plantations may give rise to criminal liability for the typical adequacy of the conduct of third parties to the crime of fire provided for in art. 250 of the Penal Code or, even, for ideological falsehood (art. 299 of the Penal Code) in the event that the producer registers a criminal occurrence as a way of exempting himself from administrative sanctions. Given the relevance in the context of criminal expert activity, it is necessary to improve methodologies able to provide subsidies to evidentiary instruction. In this work, Principal Component Analysis was used to evaluate multispectral data from optical orbital sensors, as well as those obtained from aerial photogrammetric surveys carried out by remotely piloted aircraft. The spectral indices of vegetation (NDVI), moisture (NDII), senescence (PSRI) and slope were evaluated in 7 (seven) plots (A-F) of sugarcane culture. A matrix with 884 rows (sample points corresponding to image pixels) and 4 columns (response variables) was created, with the data being autoscaled. The PCA results revealed that two principal components represent 97% of the original data variance. Despite the lack of significant divergences in the values of NDVI, NDII and PSRI between the plots, there was a clear difference in the terrain slope variable when comparing plot A in relation to the other plots, which weakens the hypothesis of accidental fire incidence.