??AI for Controlled Environment Agriculture: Monitoring Bok Choy Growth ??

Background

The rapid adoption of Controlled Environment Agriculture (CEA) and Soilless Growing Systems (SGS) presents new opportunities to enhance the year-round production of high-quality specialty crops through precision agriculture solutions.

CEA-SGS allows for optimized environmental conditions, minimizing pest and pathogen pressures while reducing the reliance on agrochemicals.

These systems also improve the efficiency of water and fertilizer use, leading to higher crop yields and quality. However, effective crop monitoring remains a key challenge, as traditional methods require specialized personnel and do not provide frequent data collection to capture the dynamics of plant growth.

Emerging smart technologies, including the Internet of Things (IoT), artificial intelligence (AI), and computer vision, offer solutions for automated and continuous crop monitoring. IoT enables seamless data exchange across interconnected devices, facilitating real-time plant growth tracking.

Computer vision techniques provide non-destructive, efficient, and accurate assessments of plant growth parameters, offering an alternative to traditional manual methods. Despite these advances, segmentation in computer vision still presents challenges, particularly as plants grow and overlap within confined spaces, making it difficult to distinguish individual plants.        

Methodology

This study aimed to develop an IoT-enabled computer vision system for continuous plant growth monitoring in CEA-SGS. The research introduced a recursive image segmentation model designed to track the growth of baby bok choy plants cultivated in a Nutrient Film Technique (NFT) system. The specific objectives included:

1. Designing and implementing an IoT-based system to capture top-view images of plants at predetermined intervals and synchronize the data in real-time.
2. Developing a Recursive Segmentation Model (RSM) that integrates segmentation results from previous images to enhance tracking accuracy over time.

An experimental setup was established at the Penn State Greenhouse Facility, where baby bok choy plants were grown under controlled conditions.

The IoT system captured images at six-hour intervals over 27 days, accumulating 108 pairs of colour images and point clouds.

The recursive segmentation approach leveraged Meta AI’s Segment Anything Model (SAM), using segmentation masks from earlier images as prompts for subsequent frames to ensure accurate plant tracking throughout the growth cycle.

The segmentation performance was evaluated using Intersection over Union (IoU) scores, comparing the RSM to other segmentation models.

Findings

The results demonstrated the effectiveness of the RSM in monitoring plant growth with high accuracy.

• The model started with an IoU score of 0.99 in the early growth stages and maintained a robust performance with an IoU score of 0.90 in later stages.
• Comparisons with conventional segmentation methods indicated that the recursive approach significantly improved segmentation accuracy, particularly when plants overlapped as they matured.
• The study confirmed that leveraging temporal continuity in segmentation improves precision and reduces the need for manual annotation.

Practical Value

The findings of this study highlight the practical applications of integrating IoT and computer vision in CEA-SGS. The automated system developed provides a reliable and efficient alternative to traditional crop monitoring methods, allowing for:

1. Enhanced Decision Support – The continuous and accurate tracking of plant growth can inform better decision-making in crop management, enabling timely interventions and resource optimization.
2. Improved Resource Efficiency – By reducing manual labour and increasing monitoring frequency, growers can enhance productivity while minimizing water and nutrient waste.
3. Scalability and Adaptability – The recursive segmentation approach can be applied to a variety of CEA systems and crop types, offering a scalable solution for precision agriculture.

In conclusion, this research underscores the potential of IoT and AI-driven solutions to revolutionize agricultural monitoring. By leveraging automated and recursive segmentation, the system developed in this study contributes to the advancement of smart farming technologies, ultimately improving the sustainability and efficiency of controlled environment agriculture.

Reference

??Case Study: How AI can practically help with assessment of biochar applications on plants?

Based on collaboration between Petiole Pro and Earth Biochar, we would like to briefly introduce the comparison of Cucumber Leaf Analysis. It was done based on a video records from cucumber trials with control plants and plants treated with different amount of biochar products.

Personal thank you Nadav Ziv, as an expert in biochar applications, for collaboration and soon we will publish this case study in more details.

?? What's On?

If you'd like to receive the regular 'AI in Agriculture' newsletter in your inbox, simply add your email to my mailing list.

Join over 10,730 readers who enjoy weekly updates on AI advancements in agriculture!

Petiole Pro - Free Web-Tools for Plant Phenotyping

Check our own discoveries at Petiole Pro.

You can find this poster in better resolution at our ResearchGate profile.

To access Petiole Pro leaf area web tool, go https://leaf-area.petiole.pro/

Petiole Pro: Leaf Greenness Measurement (DGCI) is available at https://leaf-dgci.petiole.pro/