In-Orchard Sizing of Mango Fruit: 2. Forward Estimation of Size at Harvest

Abstract

Forecast of tree fruit yield requires prediction of harvest time fruit size as well as fruit number. Mango (Mangifera indica L.) fruit mass can be estimated from correlation to measurements of fruit length (L), width (W) and thickness (T). On-tree measurements of individually tagged fruit were undertaken using callipers at weekly intervals until the fruit were past commercial maturity, as judged using growing degree days (GDD), for mango cultivars ‘Honey Gold’, ‘Calypso’ and ‘Keitt’ at four locations in Australia and Brazil during the 2020/21 and 21/22 production seasons. Across all cultivars, the linear correlation of fruit mass to LWT was characterized by a R2 of 0.99, RMSE of 29.9 g and slope of 0.5472 g/cm3, while the linear correlation of fruit mass to )2, mimicking what can be measured by machine vision of fruit on tree, was characterized by a R2 of 0.97, RMSE of 25.0 g and slope of 0.5439 g/cm3. A procedure was established for the prediction of fruit size at harvest based on measurements made five and four or four and three weeks prior to harvest (approx. 514 and 422 GDD, before harvest, respectively). Linear regression models on weekly increase in fruit mass estimated from lineal measurements were characterized by an R2 > 0.88 for all populations, with an average slope (rate of increase) of 19.6 ± 7.1 g/week, depending on cultivar, season and site. The mean absolute percentage error for predicted mass compared to harvested fruit weight for estimates based on measurements of the earlier and later intervals was 16.3 ± 1.3% and 4.5 ± 2.4%, respectively. Measurement at the later interval allowed better accuracy on prediction of fruit tray size distribution. A recommendation was made for forecast of fruit mass at harvest based on in-field measurements at approximately 400 to 450 GDD units before harvest GDD and one week later.

Management Information Systems for Tree Fruit–2: Design of a Mango Harvest Forecast Engine

“…Mango fruit mass increases linearly in the month leading up to harvest, except if growth is disturbed via a change in tree physiology, e.g., in water stress or sourcesink balance [35,36]. For nondisturbed conditions, size measurements (S) taken on two occasions (t 1 and t 2 ) can therefore be used to estimate rate of growth, m (e.g., in g/day), as follows:…”

“…Hectre [34] provides an example of commercial software for monitoring fruit size distribution of fruit in field harvest bins using a mobile device and machine vision. Mango fruit mass increases linearly in the month leading up to harvest, except if growth is disturbed via a major change in tree physiology, e.g., in water stress or sourcesink balance [35,36]. For nondisturbed conditions, size measurements (S) taken on two occasions (t1 and t2) can therefore be used to estimate rate of growth, m (e.g., in g/day), as follows:…”

Fruit Sizing in Orchard: A Review from Caliper to Machine Vision with Deep Learning

“…The need for such fine-tuning depends on the level of prediction precision and accuracy required by the user. For example, Amaral and Walsh [14] noted an increase in error in using a single model across the mango cultivars 'Honey Gold', 'Calypso', and 'Keitt', as opposed to the use of cultivar-specific models, with RMSE increased from 20.4 to 35.4 g on fruit with an average mass of approximately 550 g. However, the lower RMSE was deemed adequate for the harvest management task.…”

“…However, imaging of fruit hanging on tree from one camera perspective can be used to measure fruit length (L), and either width (Wi) or thickness (T) if fruit orientation is controlled, or a value between fruit width and thickness if orientation is not controlled. Estimation of fruit weight (W) in these applications requires the use of a simplified relationship, e.g., W = kLWi 2 for fruit in a controlled orientation, as used in a mobile phone based sizing application [23], with some increase in prediction error, e.g., an increase in RMSE from 23.9 to 25.0 g for mango fruit with average mass of 482 g [14].…”

“…As the caliper method is labor-intensive, alternative methods are sought [14]; however the method is commonly used to verify the truth of alternate procedures, i.e., as a validation device [14,23,[25][26][27][28].…”