Winetech Technical Yearbook 2022

BACKGROUND Water is a limited natural resource and a major environmental constraint for vineyard crop production. The unpredictability of rainfall patterns, comb i ned wi t h t he po t ent i a l l y catastrophic effects of climate change, could present dire future scenarios of water shortages for grape farming. Furthermore, major water shortages would negatively affect vegetative growth and grape quality, potentially leading to devastating losses in grape production. 1,2 Losses in crop production would have disastrous effects on job security and national income. It is therefore important to develop management schemes and farming practices that can improve water usage. Developing methods capable of detecting water-stressed vines can help prevent these devastating crop losses and improve B

the management of water resources. Hyperspectral remote sensing (also known as imaging spectroscopy) could provide a potential solution for the monitoring of vineyard water status. Hyperspectral data collection captures a wide range of electromagnetic radiation (typically 350 - 2 500 nm) at narrow spectral intervals (typically 10 nm). The more detailed spectral information can better detect the biochemical and physiological differences of vegetation and therefore lead to more accurate detection of water-stressed vines. 3,4 Hyperspectral data, analysed using machine learning algorithms, may prove to be the most accurate approach for the non-destructive detection of vineyard water stress. MATERIALS AND METHODS ● Ground-based hyperspectral images were captured for a Shiraz vineyard on the Welgevallen experimental farm in Stellenbosch. The SIMERA HX MkII hyperspectral camera (SIMERA Technology Group, South Africa) was used to capture the hyperspectral images. Images comprised 176 spectral wavebands ( p ).

Accuracy vs Dimensionality vs Time

RF Default

RF Op � mised

XGB Default

XGB Op � mised

RF

XGB

0,9

6000 5500 5000 4500 4000 3500 3000 2500 2000 1500 1000

0,8

0,7

0,6

0,5

0,4

Time(s)

Accuracy

0,3

0,2

0,1

500

0

0

All bands

Filter

Filter - Wrapper

Wrapper

Feature Selec � on

FIGURE 1. RF and XGBoost classification results and respective processing times.

TABLE 1. RF and XGBoost important wavebands as determined by the KW, FW and SFFS feature selection approaches. Common wavebands are highlighted in bold. Processing time for model classification is also provided.

KW

FW

SFFS

473.92, 474.74 , 475.58 , 476.41 , 477.25, 478.09 , 478.94 , 479.78 , 480.63, 481.48, 482.34 , 483.20, 484.06 , 484.92, 485.79 , 486.66, 487.53 , 488.41 , 489.29 , 490.17 , 491.06 , 491.95 , 492.84, 493.74, 494.64, 495.54, 497.36, 504.76, 577.17, 578.48 , 579.79, 581.11 , 582.43 , 583.77, 585.12

475.58, 488.41, 578.48, 644.22

RF (nm)

473.92, 474.74, 475.58, 476.41, 478.09, 478.94, 479.78, 480.63, 483.20, 484.06, 484.92, 485.79, 487.53, 488.41, 489.29, 490.17, 491.06, 491.95

Classifier

474.74, 475.58, 476.41, 478.09, 478.94, 479.78, 482.34, 484.06, 485.79, 487.53, 488.41, 489.29, 490.17, 491.06, 491.95, 578.48, 581.11, 582.43

496.45, 521.32, 585.12

XGBoost (nm)

25

WINETECH TECHNICAL YEARBOOK 2022