Model for estimating barley production based on satellite images
2022
Florin Sala | Mihai Herbei
The present study aimed to analyze and evaluate the barley crops, and also to estimate production by analyzing satellite images. The study was conducted at Didactic and Experimental Resort (DER) of BUASVM Timisoara, Timis County, Romania. The PlanetScope platform was used for the study, with a spatial resolution of 3 m. The satellite images were taken in the PlanetScope Remote Sensing System, at 7 different moments, between 27 March and 27 June 2020. Based on spectral data, MSAVI2 and NDVI indices were calculated. Four plots cultivated with barley were studied (B/A75, B/A80, B/A82 and B/A84). The variation of the MSAVI2 and NDVI indices in relation to the time factor (T, days), over the study interval, was described by the polynomial models of 2nd degree, in statistical accuracy conditions. Based on the regression analysis, the estimation of the production based on the MSAVI2 and NDVI indices was possible under the conditions of R2=0.998, P<0.05 for plot B/A75, R2=0.999, P<0.05 for plot B/A80, R2=0.997, P=0.0577 for plot B/A82 and respectively, R2=0.999, P<0.05 for plot B/A84. From the calculation of the RMSEP index, for the estimated productions, the values were obtained: RMSEP=80.8162 for YB/A75, RMSEP=50.1633 for YB/A80, RMSEP=192.3947 for YB/A82 and respectively RMSEP=112.2899 for YB/A84. The value of RMSEP=50.1633 for YB/A80 confirms that for the plot B/A80 the production estimate was made with the highest precision.
显示更多 [+] 显示较少 [-]