Economic studies show labor accounts for approximately 60% of the variable costs in apple production in the U.S. Among the labor costs, fruit harvest, tree pruning, and fruit thinning account for 44%, 25%, and 21 %, respectively. Facing a shortage of farm workers, the apple industry has been seeking aggressively to automate these manual tasks.
Several units of the robotic apple fruit harvesting systems developed by FFRobotics, an Israel-based company, will reportedly be available commercially this year. With this exciting development, many are asking about when robotic apple pruners will be on the market. Following is a brief update on the current status of robotic pruner development, and a genetic solution that could be a viable alternative to robotic pruners in the apple orchard.
To keep apple trees in optimal shape for fruit production with high yield and high quality, horticulturists have developed multiple orchard training systems that are being used in modern orchards, such as tall spindle, V-trellis, super spindle, solaxe, 2-D fruiting wall, and bi-axe systems. In any of the orchard systems, tree pruning is an indispensable component as it helps remove non-productive branches, controls overall canopy shape and size, and increases sunlight interception.
In a tall spindle system, for example, tree pruning occurs in both dormant and growing stages. Dormant pruning is conducted usually by human hands with the focus on removing large branches to achieve branch renewal. However, summer pruning is executed by a tractor-powered hedge trimmer that removes all parts of branches exceeding a predetermined distance from the trunk. In other orchard systems, dormant pruning is performed more or less similarly.
As you see, dormant pruning is essentially a process involving a series of decisions. You must determine which tree and which branch should be pruned, to what extent the tree should be pruned, and at which point of branch the cut should be made. Therefore, dormant pruning requires clear pruning rules, proper training, and skills. Consequently, it is laborious. For these reasons, dormant pruning is the targeted application in robotic pruner development.
Replacing human pruners, robotic pruning systems must have the artificial intelligence and ability to detect branches, locate the cutting sites, and make the cuts quickly. Efforts by Amy Tabb (a USDA-ARS Research Engineer at the USDA Appalachian Fruit Research Station in Kearneysville, WV) to develop a robotic system for apple tree dormant pruning were highlighted in American Fruit Grower’s June 2017 issue. To date, the prototype remains under development.
However, great progress has been made in developing grapevine robotic pruners. For example, Vision Robotics Corporation, San Diego, CA, has been testing its commercial prototype, the Intelligent Autonomous Grapevine Pruner, in vineyard trials for years.
Another grapevine robotic pruner prototype has been developed in New Zealand. In vineyard conditions, it can prune one vine in two minutes, similar to human pruners. This prototype requires reconstruction of 3-D models of the entire vine using cameras and computers, which is a major difference from the Vision Robotics system.
Notably, both grapevine robotic pruner prototypes use an enclosure that shuts out sunlight to create a controlled environment where robotic pruning is accomplished. This is caused mainly by the limitations in the current vision technologies. The enclosure provides a consistent background for cameras and computers to better detect branches.
Engineers attribute their greater success in developing grapevine robotic pruners to grapevines’ relatively simpler canopy architecture. Although the basic requirements for robotic pruners are likely similar, the complex structure of apple trees would slow down and/or obstruct the process of reaching the cutting site and making the cuts.
If apple tree robotic pruners must also use similar enclosures that block sunlight, which will be limited in their dimensions, most of the modern orchard systems mentioned above would not be compatible, except for the 2-D fruiting wall system, which is designed for robotic apple harvesters.
Changing Tree Structure
A genetic solution for reducing pruning labor cost is to grow columnar apples.
Columnar apples originated from ‘Wijcik McIntosh,’ a ‘McIntosh’ sport mutation discovered in the 1960s. Differing from ‘McIntosh,’ ‘Wijcik McIntosh’ grows considerably fewer branches and bears fruit from short spurs on the main trunk or primary branches. These characteristics make columnar apple trees require minimal pruning, reducing pruning cost.
The undesirable characteristics of ‘Wijcik McIntosh’ include low vigor and a strong tendency toward biennial bearing. However, these undesirable traits can be disassociated with the desirable traits in improved columnar apples.
The columnar growth habit is caused by increased expression of the Co gene on chromosome 10. The gene is present not only in ‘Wijcik McIntosh’ and ‘McIntosh,’ but also in other apples of standard growth habit, such as ‘Gala,’ ‘Fuji,’ and ‘Honeycrisp.’ However, its expression is repressed in these standard cultivars.
To have a columnar version of these widely grown varieties, a possible approach is to elevate the Co gene expression levels in the shoots, which seems doable now even without introducing foreign DNA based on the latest advances in plant biotechnology.
The goal of the 2-D fruiting wall system is to have a simple canopy for the existing apple cultivars so that robotic pruning and harvesting are feasible. However, creating and maintaining such 2-D fruiting walls requires intensive investment as well as labor.
Growing new and improved columnar apples or a columnar version of the widely grown apple cultivars accomplishes the same goal of 2-D fruiting walls without extensive wiring and tiring. Given the natural 2-D fruiting walls offered by columnar apple trees, it is time to consider columnar apples as an important alternative for reducing production costs.