Artificial intelligence for better crops

The technology could change how farmers protect their crops, by detecting plant diseases much earlier. But the challenge is to develop tools that are affordable and effective.

swarms of Locusts destroy crops in east africa, Corn rootworms wreak havoc In the Midwest of the United States. Pests destroy rubber trees in Brazil and Destroy potatoes in south India. Unpredictable and irregular weather patterns that cause them Climate change It will only exacerbate these problems – and scientists say it increases the potential for crop diseases to strike and cause serious damage .

A warm winter can enable the pest to invade new areas. The armyworms that feed on maize and millet and the refractotrophic fruit fly that feed on fruits and vegetables have them Spread to new locations as a result of the high air temperature. Desert locusts that destroy entire crops when they congregate, they are expected to strike new territories while changing their migration routes. It’s a serious problem in a world where he’s valued 700 million to 800 million people faced hunger In the year 2021 and as the world population is increasing.

Plant pathologist Karen Jarrett of the University of Florida, Gainesville, believes that artificial intelligence (AI) could be very valuable in combating these pests. If agriculture is equipped with cost-effective AI tools that can identify crop diseases and pest outbreaks early in their development, farmers and others can spot problems before they take off and cause real damage, she says — a topic she and her colleagues have explored. in 2022 Annual review of plant pathology. This conversation has been edited for length and clarity.

You specialize in the study of plant diseases, so let’s dive into this topic from this angle. How do changes in the environment and climate affect plants and the emergence of plant diseases?

Most pathogens have a range of temperatures that they prefer. From the pathogen’s point of view, some years can be better than others. Sometimes a harsh winter or prolonged drought will kill the pathogen. But it won’t happen in a mild year – so the pathogen will thrive, and there may be more diseases in following seasons.

Consider late blight in potatoes. It is a famous example of a plant disease that had a major impact on European society during the mid-1840s. Late blight was one of the driving factors for the Irish Potato Famine, which led to a massive exodus from Ireland.

First, the pathogen was introduced. Then there were some years when weather conditions strongly favored the pathogen: cold, wet weather. As a result, the pathogen thrived, wreaking havoc on the crop. It is estimated that 1 million people were killed and 1 million people fled the country during that time.

Today, as temperatures become more moderate, such as at higher elevations and toward the Earth’s poles, pathogens that favor temperate conditions can move into new areas and become more destructive.

When new crop diseases emerge, how can anyone be sure they are linked to climate change?

Any pestilential crop is a kind of storm. It’s hard to say whether or not an individual storm was caused by climate change, but you can start to draw conclusions.

One thing plant pathologists talk about all the time is the “disease triangle.” Infection with a disease requires three things: a pathogen capable of infecting, a conducive environment, and a host plant that can become infected. If the environment changes, for example through climate change, so that the favorable weather for pathogens becomes more common, it will make it easier for the pathogen to thrive and attack more plants. People’s decisions about how to manage plant diseases are another dimension. Often, many of these components change at the same time, so it is difficult to quantify the amount of damage caused by the pandemic due to climate change.

Let’s add artificial intelligence to this discussion. How can AI help mitigate pathogen threats to crops?

Artificial intelligence is intelligence produced by a machine, such as a computer system that is equipped with learning algorithms that can continue to improve their ability to make predictions as they get more information. These tools are so advanced that they can process massive amounts of information in a matter of seconds. For crop resilience, AI can help by making better crop-monitoring tools, designing better pesticide delivery or harvesting robots, and better software to help breed traits such as disease resistance and drought tolerance. It has a strong social angle, as it can help farmers and policy makers make smart decisions.

Let’s analyze each of those. How has artificial intelligence been used in surveillance technologies and what technologies are there? Can you explain?

If you think of an epidemic in an area, in the early stages, the disease is only present in a few locations. And after that, it will begin to grow rapidly. There is a possibility for monitoring to use remote sensing techniques such as Drones and satellite imagery that can locate crops on agricultural lands infected with pathogens. AI tools can already use image analysis to detect changes in a file Coloring leaves, flowers or fruitsand even their shapes or sizes.

Identifying diseases and taking action early on can make managing an epidemic easier. In the past, satellite data was very poor: You couldn’t get a resolution high enough to diagnose a problem. But the resolution keeps getting better. As a result, their potential for use in surveillance is increasing.

How exactly does AI use image analysis in these tools?

Well, it’s a lot of work in the beginning. First, people have to collect and curate thousands of photos of healthy and diseased plants in a range of conditions. So collecting and organizing these images takes time and investment. Algorithms are then developed to learn from these images of healthy and diseased plants, to identify signatures of disease.

Many diseases have distinct symptoms that can be detected with the naked eye. So if you have a drone, for example, that can go and take pictures in large fields, those pictures can be compared and analyzed using AI to efficiently diagnose visible crop disease.

For example, our co-author Michael Selfage In Colombia, this technology has been worked on to identify diseases in bananas. In Florida, some farmers have invested in surveillance drones. Currently, some farmers are scanning images from drones themselves, to get a quick view of their orchards. This will likely be gradually replaced by automated image analysis of orchard videos as image analysis develops further and can find diseased plants efficiently.

But there are also issues with safety regulations, because unplanned use of drones can lead to safety risks for the public as well. It is still a young industry. But because there are so many advantages, I think they will only expand, as policies strike a balance between protecting the public and providing benefits in agriculture.

And how can AI be used with automated tools to help crop resilience?

Agricultural robots are a growing field at the moment. An interesting example of artificial intelligence that already exists is the separation of healthy fruits from those infected with pathogens or damaged.

Fruit can often be distinguished as diseased or not, based on color and shape. These AI tools can process those images faster and more consistently so that discolored, lower-quality fruits — often infected with pathogens — are separated automatically.

Also, there is the idea of ​​using drones that can collect images, analyze them and then take immediate action based on the analytics – for example, to make a decision to spray an insecticide. I think these tools will probably be ready for more widespread use in the near future, and you will again need good policies.

Tell me more about how AI tools can help plant breeding and make strains more resilient.

You can think of plant breeding in part as a numbers game, because you have to breed plants and process a lot of individual offspring when you breed to get a trait. Crop breeders are looking among these offspring to find good traits for further development.

Plant breeders can use AI tools to predict which plants will grow rapidly in a given climate, which genes will help them thrive there, and which genes crossed between plant parents will most likely result in better traits. Traits can relate to speed of growth, cooking characteristics, yield, and resistance to pathogens. Crop breeders inoculate the offspring with pathogens and see which ones are resistant, and what genes are associated with resistance.

Artificial intelligence can speed up the analysis of a large number of genetic sequences related to these characteristics and find the correct combination of DNA sequences that you need to obtain a desirable trait. Image analysis is increasingly being used to characterize offspring in breeding programs for major economic crops such as wheat, maize and soybeans.

How did farmers start integrating AI tools around the world?

People have been working on disease image analysis tools so farmers can take a picture of their plant and then get an assessment using the phone. For example, PlantVillage Nuru It is a mobile application that uses image analysis to diagnose potential diseases in crops. It uses machine learning and thousands of images of crop diseases collected by experts from around the world. The images are analyzed by AI and support farmers in making informed decisions about crop management.

In general, image analysis for disease diagnosis is not 100 percent accurate, but it can provide a level of confidence to help farmers diagnose their crop diseases and understand the uncertainty.

What are some of the challenges involved in developing these types of AI tools?

For one thing, you need a lot of data for the AI ​​system to learn from. To create an image analysis tool for diagnostics, you need to include a representative group of crop varieties, which can have a wide variety of shapes and colors. One of the big challenges is getting enough of these images labeled correctly for use in the Image Analysis for Learning tool.

Another big issue is the cost. There can be a lot of tools out there that do what you want them to do, but is the benefit they bring great enough that it’s worth the cost investment? I think there are a lot of AI tools that are already useful, but they may not be profitable for farmers yet. Many current applications exist in situations where high-value materials are being processed, such as in post-harvest fruit handling and crop breeding.

Another type of challenge is training and capacity building so that the use of such tools is not dependent on a single expert but is used on a larger scale. The challenge for AI and new technologies in general is to ensure that the costs and benefits are distributed fairly across society.

What is your ideal vision for securing a climate-resilient food security system in the future?

To be resilient in the face of climate change, our food systems must be designed to respond quickly to new challenges. We can predict some future challenges, but some changes are likely to come as a surprise. Education and capacity building are key to resilience, along with effective collaboration locally and globally. international proposal for A global surveillance system for plant diseases It is an inspiring vision.

For food security in general, we need to support science education and capacity building, to make the best use of our existing technologies and support the development of better ones. We need to work towards food systems that minimize the negative impacts of farming on wild lands and maximize the benefits to human health.

This article originally appeared in Well-known magazine, an independent journalistic endeavor from Annual Reviews. Sign up the news.

Leave a Comment