Fully Autonomous Imaging Pipeline Developed to Assess Leaf Health
Outcome/Accomplishment
Researchers at the U.S. National Science Foundation (NSF)-funded Internet of Things for Precision Agriculture (NSF IoT4Ag) Engineering Research Center (ERC) have successfully evolved a fully autonomous imaging pipeline for better assessing leaf health. The team—based at the University of Pennsylvania (Penn) and the University of Florida—successfully demonstrated the feasibility of a low-cost, lightweight, compact robot that has been equipped with an RGB camera and sensor localization algorithm, as well as a hyperspectral imaging system that efficiently images the localized sensor, acquiring sensor resonance information. Sensor resonance was then measured using a SpectralZoom camera.
Impact/Benefits
NSF IoT4Ag is working to enable Internet of Things (IoT) technologies for precision agriculture that will address the challenges of ensuring food, energy, and water security for decades to come. The team’s LEAFSENSE technology combines artificial intelligence and smart sensors to enable precision agriculture, optimizing productivity and reducing costs while minimizing the use of energy and water resources and the impact of agricultural practices on the environment. LEAFSENSE leverages colorimetric sensors to autonomously map spatial and temporal characteristics of local leaf health levels at low cost and without a power source. While these passive, metasurface-based colorimetric sensors are cost effective and efficiently capable of directly measuring crop health, they necessitate a customized imaging system for agricultural environments.
Explanation/Background
The team included Drs. Sanjeev Koppal (Florida), Vijay Kumar (Penn) and Cherie Kagan (Penn). Experiments were conducted both indoors and outdoors involving autonomous navigation, RGB sensor localization, hyperspectral image capture of leaf sensors, and leaf sensor resonance acquisition from hyperspectral images. RGB sensors are optoelectronic devices that detect visible light and measure the intensity of red, green, and blue wavelengths. Hyperspectral images are rich digital pictures where each pixel contains a continuous spectrum of light across hundreds of narrow, adjacent wavelengths to record a unique chemical fingerprint for every single point in the scene. The outdoor experiments measured sensor resonance using a SpectralZoom camera that can optically magnify distant or microscopic targets while simultaneously analyzing chemical and material properties. The work was accepted into the International Conference on Robotics and Automation (ICRA) Workshop in 2025 and is currently under review for IEEE Robotics and Automation–Letters.
Location
Philadelphia, Pennsylvaniawebsite
Start Year
Microelectronics and IT
Quantum, Microelectronics, Sensing, and IT
Lead Institution
Core Partners
Fact Sheet
Outcome/Accomplishment
Researchers at the U.S. National Science Foundation (NSF)-funded Internet of Things for Precision Agriculture (NSF IoT4Ag) Engineering Research Center (ERC) have successfully evolved a fully autonomous imaging pipeline for better assessing leaf health. The team—based at the University of Pennsylvania (Penn) and the University of Florida—successfully demonstrated the feasibility of a low-cost, lightweight, compact robot that has been equipped with an RGB camera and sensor localization algorithm, as well as a hyperspectral imaging system that efficiently images the localized sensor, acquiring sensor resonance information. Sensor resonance was then measured using a SpectralZoom camera.
Location
Philadelphia, Pennsylvaniawebsite
Start Year
Microelectronics and IT
Quantum, Microelectronics, Sensing, and IT
Lead Institution
Core Partners
Fact Sheet
Impact/benefits
NSF IoT4Ag is working to enable Internet of Things (IoT) technologies for precision agriculture that will address the challenges of ensuring food, energy, and water security for decades to come. The team’s LEAFSENSE technology combines artificial intelligence and smart sensors to enable precision agriculture, optimizing productivity and reducing costs while minimizing the use of energy and water resources and the impact of agricultural practices on the environment. LEAFSENSE leverages colorimetric sensors to autonomously map spatial and temporal characteristics of local leaf health levels at low cost and without a power source. While these passive, metasurface-based colorimetric sensors are cost effective and efficiently capable of directly measuring crop health, they necessitate a customized imaging system for agricultural environments.
Explanation/Background
The team included Drs. Sanjeev Koppal (Florida), Vijay Kumar (Penn) and Cherie Kagan (Penn). Experiments were conducted both indoors and outdoors involving autonomous navigation, RGB sensor localization, hyperspectral image capture of leaf sensors, and leaf sensor resonance acquisition from hyperspectral images. RGB sensors are optoelectronic devices that detect visible light and measure the intensity of red, green, and blue wavelengths. Hyperspectral images are rich digital pictures where each pixel contains a continuous spectrum of light across hundreds of narrow, adjacent wavelengths to record a unique chemical fingerprint for every single point in the scene. The outdoor experiments measured sensor resonance using a SpectralZoom camera that can optically magnify distant or microscopic targets while simultaneously analyzing chemical and material properties. The work was accepted into the International Conference on Robotics and Automation (ICRA) Workshop in 2025 and is currently under review for IEEE Robotics and Automation–Letters.