Top 10 Fastest-evolving AI applications in crop farming
As you read this, numerous innovative AI applications for enhancing crop planting, performance, and harvest are being released worldwide. From advanced sowing, pruning, and harvesting robots to AI-powered nanosensors, the future of farming is evolving rapidly.
According to a team of scientists based in the US (for example Dr. Hamrani and Dr. Krishnaswamy Jayachandran at Florida International University), Morocco, Saudi Arabia and Spain, “it’s anticipated that AI will revitalize both existing and new” areas of agriculture, from traditional row cropping to vertical farming. This team just published a review of AI farming applications, noting that investment in this sector is expected to reach $2.6 billion USD by 2025.
At its core, the value of AI is its capacity to analyze large amounts of data very quickly and provide real-time or very fast support for specific decisions. It can also provide a great deal of support for deciding on the best overall longer-term strategic approach to specific problems on farms.
Here are 10 of the most critical and fastest-evolving AI technologies being used in crop farming today, or coming soon.
1. AI seeding and planting
Software programs are increasingly being used in real-time to ensure depth of seed placement, seeding rate and more is optimized at planting, through harnessing data from soil moisture sensors, weather stations, soil test results and more. These programs are rapidly approaching or are currently at AI level.
Hamrani and Jayachandran explain that AI is also being developed to choose at planting, “the crop varieties best suited to the specific conditions of a field, and varietal selection for drought resistance, pest resistance and higher yield.”
2. Irrigation
AI and IoT technology (integrated with pumps and valves, for example) are enabling real-time management of irrigation systems to maximize crop performance while conserving the most amount of water.
In their paper, Hamrani and Jayachandran add that “many technologies have been developed to control the communication between machines and different nodes settled in agricultural farms. These machine-to-machine technologies are efficient for monitoring soil moisture content and temperature at periodic intervals to automate the irrigation with precise requirement.”
3. Soil health monitoring
An AI firm based in California, Saiwa, notes that wireless IoT soil sensor networks throughout a field can “continuously transmit data to a central gateway or cloud platform, where it is processed and analyzed using AI algorithms to detect patterns, identify potential issues and make informed decisions.”
Specifically, and AI system called Convolutional Neural Networks (CNN) are able to classify high-resolution aerial soil images as well as scans from ground robots to determine soil types and flag visual patterns indicative of soil health issues. Saiwa explains that “object detection models identify the coverage of vegetation, crop residue, gully erosion and more from such images. Spectral signature models analyze elemental compound mixtures in a location using sensor time series or spectral camera data.” In addition, ‘class activation map’ approaches on hyperspectral data enable AI to pinpoint important soil bands.
At the same time, ‘recurrent neural networks’ can monitor temporal changes in moisture, salinity and nutrient levels in fields using delayed sequences, reports Saiwa, uncovering critical seasonal effects on soil. “In contrast, graph neural networks analyze the intricate connections and interactions among soil structure, crop biology and added amendments that affect soil quality.”
4. Automated weed, insect and disease management
AI systems are in active development that integrate visual systems that distinguish weed species from crops, and the presence of diseased crop plants or insect-damaged plants, as well as insects themselves. AI discernment using vast databases of reference images is supported by decision trees, random forests and neural networks.
“The AI-based disease detection process involves image collection, image labelling, data splitting and storage, and dividing the data set into training, validation and testing data subsets,” explain Hamrani and Jayachandran in their paper. “The model is trained and validated using the data subsets, as defined, and then the model results are tested against the third data set to provide the decision on whether the disease exists in the crop or not.”
The AI then makes rapid recommendations for actions that could be taken to deal with weeds, insects and disease. AI systems can also be connected to IoT technology in the field that determines whether fungal pathogen spore levels in the air (or number of insects or weed pressures) are approaching set thresholds, and combined with weather data, recommend when and what crop protection products to spray.
5. Autonomous robots, drones and tractors for crop management and harvesting
The day is coming when a farm AI system will be able to monitor the crop for threats and automatically oversee spraying of the correct product by large tank-carrying drone sprayers or another type of robot sprayer.
AI-based robots are already revolutionizing harvesting as well. According to Jayachandran and his colleagues, there is already “a remarkable example” of a sowing, pruning, and harvesting robot designed to work efficiently in dense vegetation. This robot, developed in 2023 by scientists at Sony and several Japanese institutes, “minimizes its impact on the environment and exhibits impressive obstacle avoidance capabilities, reducing operating time by 49 % compared to traditional controllers.” There are also efforts happening to integrate spectrometers on AI-enabled grape harvesters that can estimate properties such as sugar and acid concentrations.
6. Grain and produce grading and sorting
Several companies have already marketed fruit and vegetable grading and sorting systems that employ visual analysis by AI. These systems are able to assess thousands of high-resolution, multi-channel fruit or vegetable images every second, cross-referencing them with images in extensive databases to make precise grading decisions.
In addition to size and ripeness, AI programs can detect defects such as splits and punctures. Today’s systems are also customizable for grading parameters to cater for seasonal dynamics and market preferences.
Similarly, AI systems can compare grain samples to databases to grade grain based on given quality parameters. At least one grain-grading AI system is already available as a smart-phone app.
8. Drought, flood and water stress prediction
Hamrani and Jayachandran report that AI systems such as CNN, Long-Short Term Memory network, and Wavelet decomposition functions combined with the Adaptive Neuro-Fuzzy Inference System, are being used in flood and drought forecasting.”
With accurate drought prediction, farmers can choose crops and/or varieties that perform best under water stress conditions. They can also employ practices such as controlled drainage to ensure fields receive more moisture, or invest in more irrigation systems.
Data from drone and satellites, combined with meteorological and soil composition data, can also predict water stress, preventing a farmer from applying in-season nitrogen (side-dress) applications that will go to waste. Deep learning techniques combined with image processing can aid in identifying water stress beyond the limitations of traditional image processing. Examples of deep learning models used for this purpose include AlexNet, GoogLeNet and Inception V3.
8. Energy Allocation
AI-based predictive analytics are being developed for ‘smart energy planning’ on farms, promoting the use of renewable energy sources and energy conservation techniques. AI can accurately predict energy demands, detecting potential inefficiencies in energy usage and analyze historical energy use patterns.
AI-powered robots and autonomous vehicles can perform tasks such as planting, harvesting and weeding with greater efficiency than less-automated or non-automated systems, explain Hamrani and Jayachandran, reducing the energy needed for these operations. They add that “AI can also determine the transportation and distribution of agricultural products, reducing fuel consumption and energy costs.”
9. More stable biomass revenue
AI can be used to provide a more stable revenue stream for farmers growing dedicated biomass crops for biofuel, heat and electricity production. The opportunities for ‘new’ oilseed crops in the production of sustainable aviation fuel and other fuels are outlined in a new US Department of Agriculture report released in March 2024. When grown as winter ‘intermediate’ crops, domesticated pennycress, carinata and camelina generate added revenue for farmers, improve soil health (functioning as cover crops) and help mitigate food versus fuel concerns.
However, crop residues also represent a source of revenue for farmers. For example, wheat straw gathered after the grain is harvested can be used to make paper products.
Use of crop AI is being applied to ensure biomass crops and residue are used to the fullest extent, promoting stable, long-term biomass-based energy production and therefore stable, long-term income for farmers. “Practically, agricultural biomass can feed large scale of bioplants (100 MW and more),” state Hamrani and Jayachandran. “Thus, pyrolysis and gasification technologies been widely developed, explaining the reason for huge AI studies coupled with these systems.”
Combined heat and power (CHP) systems can also be fed with crops or crop residues (systems which can also support further food production when they heat greenhouses). AI can be used to maximize CHP efficiencies and to oversee biomass supply versus demand (forecasting required acreage for the coming year, for example, based on electricity needs), therefore maximizing long-term, stable income streams for farmers.
10. AI and Nanotechnology
AI, combined with nanotechnology, is predicted to take farming to an entirely new level. Applications in agriculture include plant sensor development through which the plant can itself sense exterior stressors like humidity levels.
“AI-driven smart delivery systems use nanotechnology to release fertilizers, pesticides and herbicides in a controlled manner, minimizing waste and environmental impact,” Hamrani and Jayachandran explain. “AI integrates with nanosensors to monitor soil health, crop conditions and environmental factors in real-time, enabling precise interventions. AI-powered nanosensors detect nutrient levels, pH, moisture content and the presence of pathogens in the soil and crops, providing accurate data for decision-making. AI optimizes the formulation and application of nanofertilizers and nanopesticides, improving nutrient uptake and pest resistance while reducing environmental impact.
Join 17,000+ subscribers
Subscribe to our newsletter to stay updated about all the need-to-know content in the agricultural sector, two times a week.
