Artificial Intelligence Models Help Farmers Plan Harvest Timing
AI models and camera systems are enabling farmers in Washington to predict ideal harvest dates and manage weather risks such as extreme heat.
Joel Carter from Okanagan Specialty Fruits in Washington State noted that extreme heat forced them to halt apple harvesting early in the morning last year, drawing attention to the importance of artificial intelligence in this process.
The Impact of Extreme Heat on Harvest
On the first day of the apple harvest in Washington State, even though the fruit was ripe and pickers were ready, temperatures reaching 38 degrees Celsius eliminated safe working conditions. Joel Carter of Okanagan Specialty Fruits stated that they stopped work at 10:00 AM during such extreme heatwaves to ensure worker safety.
The Role of Artificial Intelligence in Agriculture
Joel Carter emphasized that artificial intelligence models factoring in weather conditions are extremely useful in predicting ideal harvest dates. Stating that knowing when the fruit will ripen is not enough, and that one must also know how long it can be harvested, Carter noted that these models help at this stage.
Technology Investments and Productivity
The company, which produces sliced apples in Washington for sale to institutions such as hotels and schools, owns over 1,250 acres of apple orchards. While investing in technology to maximize efficiency, the firm also cultivates genetically modified apples that do not brown after being cut.
Vivid Machines Cameras and Artificial Intelligence
Okanagan Specialty Fruits is testing cameras from the Canadian firm Vivid Machines by mounting them onto tractors. Collecting images of apple orchards while the tractor moves, the artificial intelligence detects buds, flowers, or fruits in these images.
Accuracy of Predictions and Historical Data
Joel Carter stated that the Vivid system currently provides them with crop forecasts and harvest dates, noting that the system can pick out even very small flower buds that are difficult to spot with the naked eye. However, he added that the prediction accuracy of the system depends directly on the quality of the historical data fed into it.