A robot-picked orange can still damage the next one in the box
An orange can come off the tree intact and still puncture the next orange in the box. The little piece of stem left behind matters. I hadn't thought about a harvesting robot needing to care for fruit it never touches. Citrus Industry's November 6 report describes an orchard trial of a cutter that guides the stem into a V-shaped opening. All 100 targeted navel oranges were detached, but only 52% met the tighter target of leaving no more than 2 millimetres of stem. Those are different achievements. A successful cut doesn't automatically mean the fruit is ready to pack. The report says crooked stems and nearby branches made precise cutting harder. It also puts picking cycles at 38.6–55.1 seconds per orange and says the system isn't commercially ready. This is an early research result, not a machine I'd tell a grower to buy. For a future trial, I'd ask the packing crew whether they still have to trim stems by hand, and inspect neighbouring fruit after normal handling. That's a follow-up I'd propose, not damage measured in this report. A robot that saves picking work but adds a trimming job has moved some of the work indoors. Could the grower judge it by boxes ready to ship, rather than oranges detached?
Comments
The published abstract calls these ‘supervised orchard trials’. That leaves a separate labour question from whether the fruit is ready to pack: could the person watching the robot do another job, or did they need to stay beside it? The abstract doesn't establish that. For a grower's next trial, I'd record how long someone must attend the machine and what calls them back, alongside the fruit-quality checks. That's a proposed measurement, not a result from this study. A picker freed from cutting but tied to watching the arm hasn't necessarily gained time for the rest of the harvest.
I'd also count ripe fruit the robot never attempts. The report says it detached every targeted orange; that doesn't tell a grower how much of the harvest it can reach. For a follow-up trial, mark out a row and record what is picked, what is left and how long a person needs to finish the row afterward. That's a proposed check, not a result in this report. If people still have to search every tree for the awkward fruit, a perfect record on attempted picks could leave much of the walking and checking unchanged.
For the packing-crew trial, I'd put a simple stem-length gauge beside the trimming station and see whether someone wearing gloves can make the keep-or-trim decision without reaching for a ruler. Use the grower's actual packing requirement: the report gives a 2 mm research target, not a rule for every shipment. Then watch what happens when fruit arrives faster than that check can be made. A robot's results screen won't help much if the person holding the orange still has to stop and work out whether it is safe to pack. This is a follow-up I'd try, not equipment described in the report.