Prediction of the Yield Per Hectare of the Crop of Chili Pepper (Capsicumchinense), by Means of a Simulation Model with GIS. A Case Study in Santo Domingo - San Jacinto Del Bua

Abstract

The present investigation refers to the application of a model of prediction of the yield per hectare of the pepper crop (Capsicumchinense), with the help of GIS tools. Results are delivered in maps which are easy to interpret. The main independent variables will be number, weight and fruit diameters. In addition, the Normalized Difference Vegetation Index (NDVI) will be included as a variable. This index will be extracted from an aerial image captured by a drone with multispectral camera. The aim is to observe the statistical correlation that exists between the Performance and the NDVI. Also a validation of the model is made comparing the estimated performance vs. the real performance. In case there is a deficit in production, corrective measures will be recommended. The methodology begins with the processing of the NDVI index. It will be treated and reclassified. Later regression will be applied in the ArcGis software. The results show that the prediction achieved a visual similarity which is very close to reality, offering ranges of (4125 to 580 kg/Ha) Estimated yield (3594 to 694 kg/Ha) Real yield and a difference of (1039 to −353 Kh/Ha); the R2 in general gave as a result 0.9 in the multivariate regression; the exponential regression between the NDVI and the real yield gave R2 of 0.068, so there is a linear dispersion of data. In general the prediction was very close to reality, and it was very similar in visual terms between the resulting images.