How computer vision is changing agriculture in 2023 | Hacker News Reader

How computer vision is changing agriculture in 2023

voxel51.com

32 pointssickeythecat6 comments

So who's actually hiring for agtech computer vision in 2023? Would love to get into the space in theory and i have a deep relevant background, just wondering if the industry is getting close to generating some real impact in agriculture.

I may be completely wrong, but look for companies at such and such university research park. I think that progress is mostly going to come from spin outs at universities all over the place, especially the ones that have strong engineering and agriculture components. These are the places where research has always been done and have always been partnering with commercial entities for R&D. The Bay Area VC approach is not likely to capture this market.

You know, UC Davis, Texas A&M, Michigan State, Purdue, Iowa, Illinois, etc. Last year I was in Minneapolis for unrelated reasons. I like visiting museums and things when I have free time in random cities. I went to the state of Minnesota’s natural history museum (in a really nice brand-new building), which is on the edge of the UMN Twin Cities campus surrounded by agriculture test fields. I sat there and watched what appeared to be custom made drones flying around swooping low going from one spot to the next without anyone obviously in control.

There's quite a few startups in this space that use computer vision to some degree, I'm aware of Bilberry which uses their own cameras to get on-field images, segment into invasive/non-invasive and spot-sprays from there.

https://bilberry.io/careers/

There's also FarmWise, which is a bit bigger than BilBerry, they have a mechanical weeder (I guess the AI is similar!)

Funny enough they both sit in Paris, not in the US, which is why you may not have heard about that - the startup hype press seems to be very insular/US-focused.

I feel like all these tech pushes in Agriculture are doomed for non-technical reasons:

* Farming is heavily protected and subsidised. So there is no reason to take risks by trying anything new.

* Farms already overproduce. There is no actual market for more food.

* Farmers have a very conservative, anti-change, culture.

* What market forces are applied are all about getting things as standard as possible. Supermarkets want 1m identical apples that are all the right colour. Your apples tasting better is worthless. And if the change to make them taste 100x better makes them 1 shade to light or dark, they're unsellable.

* Most farmers have zero interest in emissions, sustainability, soil quality etc. They don't think about anything past the next harvest.

* Because of the above factors, there is little profit in farming. That means there is little cash for investments in change and anyone entrepreneurial or profit motivated left the industry long ago.

I am not saying farmers are bad people. I am just saying that trying to bring innovative tech to people who do not want it is going to be an uphill battle and tech people always seem surprised they're not welcomed with open arms...

>* Most farmers have zero interest in emissions, sustainability, soil quality etc. They don't think about anything past the next harvest.

I know quite a few farmers, and I also own a farm. We do have quite an interest in soil quality and sustainability. We understand you cannot tax the land over and over again and expect to have a good harvest 10 years from now. Crop rotation, soil testing, regenerative agriculture, and composting are big themes amongst the [industrial scale] farmers I know. I am a small fry, and those issues are even more important for me.

Big corporations that operate farms are another story however.

I feel like one of the coolest parts about AgTech, including computer vision, is how it is helping align economic incentives and climate/sustainability. It is making a lot of practices that are better for the environment the obvious choice!

Reply on news.ycombinator.com