Hacker News Reader

Comment by mrfusion | Hacker News Reader

It seems like picking a lot of fruits are easily in the realm of what computer vision is capable of these days.

What would be the first steps to doing a start up in this space? Anyone interested?

I think the first step would be to work in the area for a while and learn about the real issues. A lot of industries look really easy when you take a quick look from the outside but reality is a little more complex.

there are startups. The issue is rarely technological capability as much as it is adoption, reliability, and durability. Not impossible just hard. There is a lot of work in the space being done by John Deere and others...a LOT of farm analytics, but most of their parts are still cast iron for a reason.

A lot of the issues currently are related to the mechanical handling of the fruit. It's very important not to bruise or damage the product. Buyers generally have strict rules on exactly how the fruit is picked, e.g., how much stem is still attached to the fruit, for both weight and quality reasons.

Other issues arise if the orchard is on uneven ground, or their are irrigation lines in the way.

Certainly not insurmountable, but not easy. Then there's the distinct problem of adoption.

potentially robotics could increase your yields by making uneven plots of land more viable for planting.

Find the easiest adaptation for the baseline of picking the fruit. E.g. a robot that picks fruit really well.

I’m definitely interested in the farm/cv realm. It takes funding though.

Reply on news.ycombinator.com