LDV Capital

View Original

Farm to Fridge: Food and Agriculture are Ripe with Visual Technology Opportunities

LDV Capital invests in people building businesses powered by visual technologies. We thrive on collaborating with deep tech teams leveraging computer vision, machine learning, and artificial intelligence to analyze visual data. We are the only venture capital firm with this thesis. We regularly host Vision events – check out when the next one is scheduled.

©Stijn te Strake/Unsplash

Plants selected by cameras. Cows fed and milked by robots. Proteins acting as sweeteners. Fridges suggesting recipes. Our food system is in for a major makeover over the next five years that will have imaging and computer vision at its core. 

It is a makeover that is far overdue in order for us to feed the 9 billion people expected to inhabit the planet by 2050. Especially while facing climate change, a global health pandemic, and closing borders. It is a makeover that will be enabled by visual technologies to help us produce more food on the same footprint of land, change our diets, and reduce waste.

There are significant business opportunities for startups and technology companies to leverage computer vision and machine learning to modernize the $8 trillion global food and agriculture industry.

Visual technology is any technology that analyzes, organizes, displays, or distributes visual data. Visual data includes everything from photos and videos to satellite, hyperspectral and macromolecular imagery. Typically, visual technologies leverage computer vision, machine learning or artificial intelligence.

Over the next five years and beyond, visual technologies will be critical to feeding the planet and our extensive LDV Capital Insights report elaborates on who will be impacted.

Producing more food without expanding land use requires advanced plant breeding, precision crop agriculture, and improved animal care all powered by visual tech.

Global land use for crop and animal agriculture has grown to over 37% of the total usable landmass. It is one of the major contributors to deforestation and greenhouse gas emissions. In order to feed ourselves without causing further degradation, farmers and agronomists are deploying visual tech solutions to produce more food on the same amount of arable land.

©Meric Tuna/Unsplash

For centuries, farmers have bred plants for traits such as taste, resilience and yield. Developing a higher-yield plant breed currently takes 5-10 years, according to USDA research, even with the adoption of next-generation sequencing (NGS) to select for traits. NGS uses fluorescence imaging to read the genetic makeup of plants and other organisms.

Over the next five years, computer vision and machine learning will help automate phenotyping to significantly aid breeders in correlating genes to physical traits. This will shorten the breeding cycle by two or more years, reduce breeding land use, and have a significant impact on the $42B commercial seed market.

©LDV Capital Insights 2020

Once seeds are planted, it is imperative to monitor fields to manage pests, irrigation and nitrogen needs to improve the yield of planted fields. Sensors, many of them visual, are being deployed to capture and track data for better decision-making and yield estimation. Drone, plane and satellite imagery is analyzed by computer vision and machine learning companies like FluroSat and Descartes Labs, to help make field-level decisions.

Stationary sensors like lux meters, equipment-mounted sensors such as near-infrared sensors, and soil sampling using chemical analysis and/or NGS will also help precisely analyze crop needs.

Collecting and analyzing the data is only as viable as the ability to act on the data to greater control the crops as they grow in the field. Data in combination with equipment-mounted cameras will enable further adoption of variable rate technologies such as fertilizer applicators, precision planters, and automated weeding systems. Together they will ultimately reduce harvest loss and improve yield allowing farmers to produce more food on the same amount of land.


As experts in visual technologies, we at LDV Capital produce an annual proprietary LDV Insights report, where we deep dive into a sector where visual technology will revolutionize businesses. Previously, we covered healthcare, manufacturing, logistics, and more.

We look forward to hearing your insights, learning about your startups, and reading your research papers on how businesses are addressing these challenges and opportunities.


According to Jorge Heraud, CEO of Blue River Technology which sold to John Deere in 2017, “Within agriculture, John Deere's innovation efforts focus on improving the efficiency of tasks performed by the farmer, such as precision herbicide spraying and automated tuning of combine harvester operating settings. Perception will change every product that Deere makes.”

©LDV Capital Insights 2020

Controlled Environment Agriculture (CEA) like greenhouses and vertical farms, most of which leverage LED lighting to grow plants, will also be part of the solution to more food production without the land increase. Lighting tends to be the highest operating and capital cost for most CEA.  Due to lighting advances, lettuce and other leafy greens are starting to be grown at price points competitive to organics at vertical farms, such as Aerofarms and Square Roots. “Lighting recipes” will also help CEA farmers create better-tasting products and a more usable yield.

In animal agriculture, around 6% of a herd dies every year before it can be consumed as food. RGB and thermal cameras will track animals 24/7, evaluating the health of the animals and getting in front of diseases. When coupled with robotics, visual monitoring will enable automated feeding systems, like Cainthus, which will help optimize feed for the animals' health. In dairy, robotic milking systems that use RGB and time-of-flight cameras will enable cows to choose when to milk then correlate feed information to milk quality and production.

In both outdoor and indoor growing as well as animal agriculture, visual technologies will be crucial to extending the yield of crops and animals in order to produce more food on the same amount of land.

New ingredients and food products must be developed with help from microscopy and imaging to help change our diets.

Changing our diets is also part of the solution. There is a lot of hype around plant-based proteins and lab-grown meats as the greatest opportunities for dietary change. However, discovering new proteins and functional nano-scale ingredients will lead to the creation of food that is healthier, better for the environment, better tasting, and more efficient to produce. It includes opportunities much larger, and lower hanging, than just traditional meat alternatives.

Structural computational biology and Computational Protein Design (CPD) use computer vision to examine changes in protein structure. Some companies, like Amai Proteins, are using CPD to discover new proteins that can be new ingredients, such as a sugar alternative.

Nanoparticle Tracking Analysis is the observation of atoms, molecules, or macromolecules illuminated via laser and viewed under a microscope. This will reveal the opportunities for nanoparticles to do everything from improving food texture and nutrition to acting as an antimicrobial agent in packaging. 

Cellular agriculture is the growing of meat, eggs, and dairy products in a cell culture instead of sourcing from animals. It is a fast-growing space that relies on microscopy for the research and discovery (R&D) phase. 

Products like rennet, a protein that operates as an enzyme which is found in the stomach lining of a calf and critical to cheesemaking, was one of the first cellular ag ingredients to be approved by the FDA in 1990. But thanks to microscopy, both plant-based rennet and lab-engineered rennet have been discovered and are now an ingredient in 80-90% of US cheese because it is purer, more reliable, and cost-effective.

In the lab-grown meat sector of cellular ag, microscopy is being deployed to identify and replace growth factors as they are the most expensive attribute of cell-based meat production. In the future, machine learning will be deployed to assess molecular structures rapidly, further assisting in efforts to make cellular agriculture commercially viable.

Ground meat will remain the focus of the lab-grown meat industry over the next five years because there is still discovery to be done to understand which bioprinting techniques, predominantly laser-assisted vs thermal, will be best to create muscle fibers and fat pockets necessary for creating tissue structure.

Changing our diets, not just by limiting our meat consumption, but also by improving the ingredients and nutritional value of our food will play a major role in efficiently producing food for humanity.

Better food safety and food waste management leveraging light and camera technology will help cut the estimated 30% of our food supply that is lost and wasted.

Moving food from field to table, an incredibly large amount of food is lost. Some are due to food safety, such as contamination, other is due to production errors, logistics, food spoilage, and plate waste. Cutting food waste at all junctures can be aided by visual technologies.

©LDV Capital Insights 2020

Nearly a quarter of food product recalls are due to operational errors such as product contamination, foreign bodies, spoilage, and unauthorized ingredients. Currently, quality assurance (QA) in food processing and packaging leverage 2D x-rays to identify large contaminants like metals but most QA is still done by human visual inspection. Produce packhouses and food processors are rolling out high-tech scanning systems utilizing RGB, near-infrared spectroscopy (NIRS), 3D x-ray, 2D & 3D imaging, and hyperspectral imaging to more accurately detect flaws. Machine vision will enable a more thorough QA process by improving efficiency during sorting and processing in packhouses and at food processors. These developments can lead to less discarded produce at processing facilities.

Pulsed light technology is gaining momentum as a replacement for heat processing in a variety of food processing applications which will help extend the shelf life of food.

Optical biosensors and fluorescence imaging will be deployed to rapidly detect pathogens such as E. coli and Salmonella faster and earlier. This can stifle outbreaks before they begin. Combined with better visualization of supply chain data, product recalls, which typically cost companies more than $30 million per product, can quickly identify and contain the spread. Over the next five years, the cost and time savings from the adoption of these visual technologies will make food safer for human consumption and save food companies millions in recalls and settlements. They will also reduce waste during the production process leading to additional revenue.

The USDA estimates that 31% of food waste occurs at the retail and consumer level in the US totaling more than $162 billion. The adoption of robotics and the Internet of Eyes at grocery stores can aid retailers in bringing products to shelves faster and doing automated promotion of items that are reaching expiration dates. The rise of in-store fulfillment for online grocery and click-and-collect models that will leverage robotics and computer vision will also help.

“To augment our goods-to-person robotics platform we use computer vision and machine learning to allow for robotic arm based pick and pack. This automated "pre-pick" is initially focused on boxed items that currently account for ~20% of grocery sales. By the end of 2020, we expect to be able to leverage robotic arm picking for 40% of items and by 2021 we will likely be able to handle ~60-80% of “pre-pick” items in grocery,” says Steven Hornyak, CCO of Fabric.

Imaging systems with computer vision and machine learning will help track and reduce food waste in commercial kitchens. Trash bins with cameras, like Winnow’s, can reduce overproduction costs by recognizing and quantifying food scraps. Intelligent food imaging systems, like Walmart’s Eden, will help prioritize the flow of perishable goods and ensure they are used before expiration.

©LDV Capital Insights 2020

Smart consumer refrigerators with cameras, computer vision, and machine learning can keep track of content. Already, top smart fridge companies like Samsung and Whirlpool, have refrigerators that analyze buying habits and suggest recipes based on inventory. They can also watch for signs of spoilage, enabling on-time consumption of food and less waste. The global market for smart consumer kitchen appliances is projected to grow at a 22% CAGR through 2025 which represents business opportunities as well as the opportunity to cut consumer food waste.

A 70% increase in demand for an $8 trillion industry is a tremendous growth opportunity that will require computer vision, machine learning, and AI solutions.

From farm to fridge, by 2025 visual technologies will be deployed across the food system to enable it to meet the changing demands of the world. 

Everywhere cameras, visual sensors, computer vision, and light are incorporated represents an opportunity for new businesses to grow and current companies to invest and improve.

Hopefully, the adoption of visual technology innovations across breeding, growing, harvesting, processing, and serving food over the next five years will help make headway towards a better food system for all.

See this gallery in the original post