RAPID-MOLT: A Meso-scale, Open-source, Low-cost Testbed for Robot Assisted Precision Irrigation and Delivery
To study the automation of plant-level precision irrigation, we present a modular, open-source testbed that enables real-time, fine-grained data collection and irrigation actuation. RAPID-MOLT costs USD $610 and is sized to fit a small laboratory floor space of 0.37 m^2. The functionality of the platform is evaluated by measuring the correlation between plant growth (Leaf Area Index) and water stress (Crop Water Stress Index) with irrigation volume. In line with biological studies, the observed plant growth is positively correlated with irrigation volume while water stress is negatively correlated. This testbed is the basis for ongoing work on learning-based irrigation controllers.
The contributions are as follows:
- A hardware system design that enables both plant level sensing with multispectral imaging and also precise irrigation control with Raspberry Pi actuated solenoids.
- A software infrastructure and data processing pipeline to evaluate irrigation controllers, including the automatic extraction of Leaf Area Index and Crop Water Stress Index from RGB and thermal images.
- Experimental data from two experiments with different crops, reflecting plant growth and water stress data in response to different irrigation schedules.
As seen in the flow chart above, each platform unit functions together as follows. Plant and environmental images and data are collected by the sensing unit, and sent to the cloud processing unit for the creation of data-specific irrigation schedules. These schedules are relayed to the irrigation unit for irrigation deployment to the potted plants.
The software is grouped in 1) sensing, 2) data processing, and 3) actuation.
- The sensing folder contains the code for the android app (uploades the thermal and rgb images to the cloud) and the arduino uno.
- The data processing contains two core files:
- downloadImagesFromGoogleDrive.py downloads the images from the cloud to the local computer
- process_images.py extracts the LAI and leaf temperature out of the images
- The actuation folder contains the scripts and files for the raspberry pie to execute the derived irrigation schedule.
The Data folder contains an example of images from one experiment and the files resulting from the image processing.
CAD models for construction can be found on this repository. Code for the software infrastructure and experimental data will be posted soon.