Recent progress on grapevine water status assessment through remote and proximal sensing: A review

• Published in: Scientia horticulturae 2024-12, Vol.338, p.113658, Article 113658
• Main Authors: Abbatantuono, Francesco, Lopriore, Giuseppe, Tallou, Anas, Brillante, Luca, Ali, Salem Alhajj, Camposeo, Salvatore, Vivaldi, Gaetano Alessandro
• Format: Article
• Language: English
• Subjects: machine learning; Precision viticulture; Satellite; Spectral bands; Thermal indices; Vitis spp L
• ISSN: 0304-4238
• DOI: 10.1016/j.scienta.2024.113658

•The use of satellite platforms is promising due to lower costs and easier image access;.•The best results are obtained using CWSI or single bands and machine learning;.•Red-edge, NIR, Green, SWIR, and red bands are sensitive to water stress;.•Traditional indices (especially NDVI) provide inferior performances;.•Choose a reliable parameter (SWP) and enlarge datasets to obtain robust models. According to modern precision agriculture principles, remote and proximal sensing can be extraordinarily useful tools for sustainable water resource management in viticulture. More than one hundred papers were read and cataloged to outline the most effective methodology (comprised of platforms, cameras, indices, single bands, and statistical methods) for monitoring water status in different wine grape varieties located in different areas. Satellites and airplanes can monitor areas at the regional or larger scale; however, while satellite images can be free, airplane imagery can be more expensive. The use of satellite platforms is particularly promising, especially due to recent technical progress aimed at improving spatial and temporal resolution. In addition, unmanned aerial vehicles (aka drones) equipped with thermal, multispectral, and hyperspectral cameras have provided excellent results. Proximal thermal and spectral cameras (e.g., handheld or installed in tractors) can be an inexpensive alternative but often present similar problems to traditional methods (e.g., time-consuming). The best results were obtained from thermal indices (e.g., Crop Water Stress Index) and the use of machine learning (ML) algorithms on individual bands and indices obtained with hyperspectral or multispectral cameras carried on drone or satellite platforms. [Display omitted]