Upp wants to add more broccoli to vegetable protein mix through large automation

What is automation for? Harvest more broccoli than human workers can, according to an AgTech startup in Shropshire, UK, that uses computer vision AI and proprietary farm machinery to boost crop yields.

His argument is not only that his AI-powered specialized harvester makes harvesting a family crop more efficient, but also that the process reduces waste by extracting more nutritious protein from a field of broccoli without the need for a man-made army. . employees.

According to Upp, the smart machines he is developing will allow broccoli growers to harvest more from the vine than would be possible with human field workers, because the combination of AI and tractor tools can do it all: detect, cut, lift and fully automated. transport. , with a speed of up to 3 km/h.

This AI-based approach allows farmers to “recycle” 80% of the broccoli plant (i.e., extra stems and leaves) that usually remains in the field as waste, and sell it as a side product, according to Upp. shape, suggesting it is similar to pea protein.

The startup’s conceptual system, which CEO and co-founder David Whitewood developed for TechCrunch with the help of Lincoln University technologists, includes a tractor equipped with a 3D camera and an on-board computer with computer vision AI. Model trained to recognize when broccoli heads are the right size to harvest (with superhuman accuracy), along with a patented harvesting and cutting tool (patent pending) pulled by a tractor.

“The main task of broccoli harvesting is recognizing which heads are ready to be harvested. So we teamed up with the Agricultural Products team at Lincoln University, which developed machine learning and AI,” he explains. “We tested many cameras with it and tackled the difficult problem of occlusion. [where leaves may partly obscure the camera’s view of the broccoli head].

“They used a depth sensor camera with the 3D part to scale this head. Because we don’t cut all the heads, we just cut them to the right size, as the supermarkets ask… Then it says “cut” and that sends a signal to our on-board computer and then we activate our patented mechanism that grabs the plant, which would be the same like a human grabbing the stem of the plant and then a very sharp blade flies in and cuts in a fraction of a second. And then the plant is dug up.”

The extra edible plant parts harvested in this way are not destined for supermarket shelves, where the strict cosmetic standards that grocers often impose on their suppliers greatly contribute to food waste by refusing to stock fruits and vegetables. The idea is that it is converted into an ingredient rich in protein and nutrients and sold to the food industry.

Upp envisions dried broccoli protein being used in a variety of products, from sports protein drinks to convenience foods and baked goods.

The pieces of broccoli he wants to recycle contain 30% protein by dry weight, according to the startup’s website, and are also packed with nutrients (vitamins A, B, C, E, K, calcium, iron, potassium, phosphorus, zinc). as well as rich in dietary fiber.

Upp seems to have had no trouble stimulating the food industry’s initial interest in recycled edible plant protein; Whitewood noted that he’s already signed a trio of industry partnerships (he can’t name names yet, but says one is a global “functional drinks” giant; another is a major British food brand; and the third is a bakery that specializes in confectionery).

“They are very interested in the healthy aspects of broccoli,” he continues. “They’re interested in it being clean and sustainable… So they’re excited, let’s say. I don’t think we’ll have a problem with a market for it if we take it off the field.”

During processing, Upp works with experts at the James Hutton Institute in Dundee to determine how “to best recover the fractions of this plant that make it particularly suitable for the food industry,” says Whitewood.

With the growing demand for alternative plant proteins, Upp is developing what it calls a “circular plant protein” as the food industry looks for ways to reduce its reliance on animal protein as much as possible while reducing its environmental footprint. pressure on farmers and food companies to meet climate targets.

As such, the startup predicts that its AI-harvested broccoli protein could become a multi-billion dollar market in the coming years.

On the marketing side, he claims an additional environmental benefit: he suggests that broccoli protein is cleaner than pea protein (it’s four times less carbon intensive to produce), while also arguing that the problem is avoided. . So the argument is that this is an even greener plant-based protein.

A possible PR problem is that automating the broccoli harvest will inevitably lead to some displacement of (human) workers.

Whitewood says the system will replace about seven field workers, but notes that “warm bodies” are still needed at the packhouse to pack broccoli products for retail. “Seven hard-to-get people,” he adds, describing the grueling work farmers often do and arguing that these aren’t the kind of jobs anyone will miss. “No one wants to do this job. Even China and India are struggling to get people to do this… This is the 21st century and we still expect people to do this. It’s crazy.”

Founded in 2022, while the startup’s technology has reached concept stage, it is gearing up for the next phase to refine a robust technology that can be commercially deployed, with several protein field pilots planned for the year in the UK, Spain and California.

Expect to start commercial production (and generate revenue) by the end of 2024; Expected sales in the three pilot stores will exceed £50 million by 2027.

The company was founded last year as an offshoot of another British AgTech company called Earth Rover, where Whitewood was CEO before joining Upp as a co-founder when they decided to split into two separate companies.

Today, the startup is announcing an initial investment of £500,000 from decarbonisation, sustainability and social impact investor Codo Beach Capital to fund field trials ahead of its planned commercial launch later next year.

Whitewood says the first commercial use of the technology will likely be in Spain or the UK due to seasonality, before Upp continues to roll out automated crop optimization for broccoli growers in California.

Why has no one ever thought of getting more out of broccoli plants? Whitewood says people have been thinking about the possibilities of doing this for over a decade, but he suggests that given the selective harvesting required, as well as the need to separate the harvested crop, with one part (the broccoli crown) simply “very difficult”. goes to supermarkets (to be sold fresh) and the rest has to be processed further.

“It sounds simple: many people have tried and many have failed,” he suggests. “Only when you have a specialized combine harvester that can handle all the volume can you suddenly take care of the rest. You need automation, and you need major automation. Small robots cannot handle crops of this scale and volume… They need machines the size of a farm.”