In the face of climate change and labor shortages, the agriculture sector is turning to technology to ensure stable crop production. A recent study published in *Frontiers in Plant Science* offers a promising solution for indoor farming, particularly in controlled environment agriculture (CEA). The research, led by Jung-Sun Gloria Kim from the Department of Biosystems Engineering at Seoul National University, introduces a phenotyping-based growth stage classification pipeline for basil that could revolutionize precision agriculture.
Traditional growth stage classification relies on time-based criteria, which often fail to capture the physiological nuances of plants and lack reproducibility. Kim’s study addresses these limitations by focusing on morphological traits, specifically the number of leaf pairs emerging from the shoot apex. This trait, the researchers found, can be consistently observed regardless of environmental variations or leaf overlap, making it an ideal indicator for non-destructive, real-time monitoring.
The pipeline employs low-cost fixed cameras to capture top-view images of plants under various artificial lighting conditions. Using YOLO (You Only Look Once), an object detection system, the system automatically detects multiple plants. K-means clustering then aligns the positions of these plants to generate an individual dataset of crop images and leaf pairs. A regression model is subsequently trained to predict leaf pair counts, which are converted into growth stages.
The results are impressive. The YOLO model achieved a high detection accuracy with a mean average precision (mAP) of 0.995 at an intersection over union (IoU) threshold of 0.5. The convolutional neural network regression model demonstrated a mean absolute error (MAE) of 0.13 and an R² of 0.96 for leaf pair prediction. The final growth stage classification accuracy exceeded 98%, maintaining consistent performance in cross-validation.
“This pipeline enables automated and precise growth monitoring in multi-plant environments such as plant factories,” Kim explains. “By relying on low-cost equipment, it provides a technological foundation for precision environmental control, labor reduction, and sustainable smart agriculture.”
The commercial implications of this research are significant. For instance, accurate growth stage classification is crucial for nutrient management, harvest scheduling, and quality improvement. Automating these processes can lead to increased efficiency and reduced labor costs, making indoor farming more viable and scalable.
Moreover, the study’s focus on low-cost equipment and real-time monitoring aligns with the growing trend of smart agriculture. As the sector continues to evolve, such technologies could become integral to precision farming, enabling growers to optimize resources and enhance productivity.
The research also opens up new avenues for future developments. For instance, the pipeline could be adapted for other crops, expanding its applicability across various CEA settings. Furthermore, integrating this technology with other smart agriculture tools could create a comprehensive, automated decision pipeline for growers.
As the agriculture sector grapples with the challenges posed by climate change and labor shortages, innovations like Kim’s pipeline offer a beacon of hope. By harnessing the power of AI and low-cost technology, the sector can strive towards more sustainable and efficient practices, ensuring stable crop production for years to come.