Grapevine Yield and Leaf Area Estimation Using Supervised Classification Methodology on RGB Images Taken under Field Conditions

Abstract

The aim of this research was to implement a methodology through the generation of a supervised classifier based on the Mahalanobis distance to characterize the grapevine canopy and assess leaf area and yield using RGB images. The method automatically processes sets of images, and calculates the areas (number of pixels) corresponding to seven different classes (Grapes, Wood, Background, and four classes of Leaf, of increasing leaf age). Each one is initialized by the user, who selects a set of representative pixels for every class in order to induce the clustering around them. The proposed methodology was evaluated with 70 grapevine (V. vinifera L. cv. Tempranillo) images, acquired in a commercial vineyard located in La Rioja (Spain), after several defoliation and de-fruiting events on 10 vines, with a conventional RGB camera and no artificial illumination. The segmentation results showed a performance of 92% for leaves and 98% for clusters, and allowed to assess the grapevine’s leaf area and yield with R2 values of 0.81 (p < 0.001) and 0.73 (p = 0.002), respectively. This methodology, which operates with a simple image acquisition setup and guarantees the right number and kind of pixel classes, has shown to be suitable and robust enough to provide valuable information for vineyard management.

Use of multi-spectral airborne imagery to improve yield sampling in viticulture

“…Other studies propose the use of alternative sources of information to facilitate or automate bunch weight estimates. This is the case of image analysis that has been proposed to detect, count and estimate the weight of clusters (Diago et al 2012;Reis et al 2012;Nuske et al 2011;Serrano et al 2005;Dunn and Martin 2004) or to estimate the number and the volume of berries (Grocholsky et al 2011;Rabatel and Guizard 2007). Other authors have proposed continuous weighing devices positioned on the trellising system (Blom and Tarara 2009) at a specific location in the field.…”

Modern viticulture in southern Europe: Vulnerabilities and strategies for adaptation to water scarcity

“…Manual ground-based and aerial manned and unmanned remote sensing measurements are being progressively implemented in modern viticulture not only in research but also in commercial vineyards to monitor plant stress and or to assess canopy and/or berry traits Grant et al, 2007;Grant, 2012;Fuentes et al, 2014;Fernández et al, 2013;Jones and Grant, 2015). These new approaches combine the use of different types of detectors and spectral wavelengths ranging from visible (red, green, blue) (RGB) and infrared thermal imaging to multispectral and tomography measurements (Leionen et al, 2006;Diago et al, 2012;Fuentes et al, 2012;Costa et al, 2013;Jones and Grant, 2015;Rustioni et al, 2014). Robots and unmanned Aerial Vehicles (UAVs) have been recently applied in precision viticulture (Baluja et al, 2012; ZarcoTejada et al, 2009 ZarcoTejada et al, , 2012Gago et al, 2015).…”

Vineyard yield estimation by automatic 3D bunch modelling in field conditions

“…Furthermore, comparing the coefficients obtained with other yield estimation studies, this method can be cataloged as a great accurate one. Related to this assertion, Diago et al (2012) achieved a R 2 of 0.73 between the observed and predicted yield values. Dunn and Martin (2004) succeeded a R 2 of 0.85 between grape weight and the ratio of grape pixels to total image pixel.…”

“…Moreover, 2D computer vision techniques have also been applied to individual strains for the identification of plant elements (Herrero Langreo et al, 2010) or counting individual berries using a flatbed scanner (Battany, 2008). A 2D grapevine yield and leaf area estimation was done by Diago et al (2012), who used a visible light camera to capture images in-field using a white screen behind the canopy. Their approach involves the computation of Mahalanobis colour distance for a supervised classification application.…”