Researchers developed a smart computer system that can tell farmers whether their potato plants have the right amount of potassium—a nutrient plants need to grow big and strong. Using special cameras that see colors humans can’t see, combined with artificial intelligence, scientists created a tool that checks potato nutrition without damaging the plants. In a two-year study, they found that potatoes grew best with a specific amount of potassium fertilizer, and their new technology could predict this need with about 82% accuracy. This invention could help farmers use fertilizer more efficiently and grow healthier potatoes while protecting the environment.

The Quick Take

  • What they studied: Can scientists use special cameras and computer learning to figure out if potato plants have enough potassium nutrition without harming the plants?
  • Who participated: Potato field experiments conducted over two years with 15 different treatment combinations (3 water levels × 5 potassium fertilizer amounts)
  • Key finding: A computer model using special camera images predicted potassium nutrition status with 82% accuracy, and potatoes grew best with 300 kg of potassium per hectare under full watering conditions
  • What it means for you: Farmers may soon be able to use cameras and computers to check if their potatoes need more potassium, helping them save money on fertilizer and grow better crops. However, this technology is still being developed and isn’t widely available yet.

The Research Details

Scientists conducted a two-year experiment in potato fields where they tested different amounts of water and potassium fertilizer. They created 15 different treatment combinations by mixing three water levels (full water, 80% water, and 60% water) with five different potassium amounts (none, 100, 200, 300, or 400 kg per hectare). Throughout the growing season, they took special photographs of the plants using cameras that can see colors invisible to human eyes. They also measured the actual potassium levels in the plants and compared these measurements to what their computer model predicted. The computer model used a machine learning technique called Random Forest, which is like teaching a computer to recognize patterns by showing it many examples.

This research approach is important because potassium is expensive and using too much or too little wastes money and harms the environment. Current methods to check potassium levels require damaging plants or waiting for lab results. This new technology could let farmers check their plants instantly without harm, making farming more efficient and sustainable.

The study was conducted over two years, which strengthens the findings by showing results were consistent across different seasons. The researchers used multiple measurement methods and compared predicted values to actual plant measurements, finding strong agreement (correlation above 0.87). The computer model’s predictions were accurate within about 5%, which is good for practical farming use. However, the study was conducted in a specific region with specific potato varieties, so results might differ in other locations or with different plants.

What the Results Show

Potatoes produced the highest yield (about 59,500 kg per hectare) when farmers applied 300 kg of potassium per hectare with full watering. Interestingly, applying more potassium than this amount actually reduced yields, showing that more fertilizer isn’t always better. The researchers created a mathematical formula to calculate the ideal potassium level for plants at different growth stages. When they tested their computer prediction model, it correctly predicted potassium nutrition status about 82% of the time using the measured method and 80% of the time using the theoretical method. The computer model worked best when it combined information from special cameras (which see plant colors) with texture patterns from plant images, using a machine learning technique called Random Forest.

The study found that most plant color measurements (vegetation indices) were significantly related to potassium status, but texture patterns from images were even more strongly related. Different combinations of camera measurements and texture features worked well for prediction, giving farmers flexibility in what equipment they might use. The researchers created visual maps showing potassium nutrition across entire fields, which could help farmers see which areas need more attention.

This research builds on previous work showing that special cameras can help monitor plant nutrition. However, this study is novel because it combines both color information and texture patterns from images, which improved prediction accuracy compared to using either method alone. The use of machine learning to predict potassium status is relatively new in potato farming and represents an advance over traditional soil testing methods.

The study was conducted in one specific region with particular soil and climate conditions, so results might differ elsewhere. The researchers tested only certain potato varieties, so the technology may need adjustment for other types. The study doesn’t explain how the technology would work in real farming conditions with varying weather or different field sizes. The computer model requires special cameras and software that most farmers don’t currently have, which could limit practical use. Additionally, the study doesn’t compare costs of this new technology versus traditional fertilizer management methods.

The Bottom Line

Based on this research, farmers should consider applying about 300 kg of potassium per hectare under normal watering conditions for best potato yields. As this technology becomes available, farmers may want to explore using camera-based monitoring systems to check potassium status during the growing season. However, farmers should wait for this technology to be tested in their specific region before making major changes to their fertilizer practices. Confidence in these findings is moderate to high for the specific conditions tested, but lower for different regions or potato varieties.

Large-scale potato farmers who want to optimize fertilizer use and reduce costs should pay attention to this research. Environmental managers interested in reducing fertilizer runoff would benefit from this technology. Potato researchers and agricultural technology companies should consider developing practical versions of this system. Small-scale farmers might find this technology too expensive currently. Home gardeners growing potatoes don’t need this level of precision.

If farmers adopt this technology, they could see benefits within a single growing season by adjusting fertilizer amounts based on real-time monitoring. However, the technology itself is still in the research phase and may take 2-5 years to become commercially available and affordable for most farmers.

Want to Apply This Research?

  • Users could track their potassium fertilizer applications (amount in kg per hectare) and corresponding potato yields (kg per hectare) to see if they’re in the optimal range of 250-350 kg per hectare. They could also record irrigation levels to understand how water and potassium interact.
  • Farmers using a nutrition tracking app could set a target of 300 kg potassium per hectare and receive reminders to monitor their plants during the growing season. They could photograph their potato plants weekly and compare visual changes to the app’s guidance on what healthy potassium levels look like.
  • Over an entire growing season, farmers should track potassium application dates, amounts, and plant appearance at regular intervals (every 2-3 weeks). They could record yield results at harvest and compare them to previous years to see if adjusted potassium amounts improved outcomes. This creates a personal database showing what works best for their specific fields.

This research describes an experimental technology for monitoring potato potassium nutrition and is not yet widely available for commercial farming. Farmers should continue following their local agricultural extension recommendations for fertilizer use. Before making significant changes to fertilizer practices based on this research, consult with local agricultural experts familiar with your specific soil, climate, and potato varieties. The optimal potassium amount may vary significantly based on local conditions, soil type, water availability, and potato variety. This study was conducted in specific conditions and results may not apply to all farming situations. Always conduct small-scale tests before implementing new fertilizer strategies across entire fields.