A Portable High-Resolution Snapshot Multispectral Imaging Device Leveraging Spatial and Spectral Features for...

Area of Science:

• Agricultural Science
• Plant Science
• Remote Sensing

Background:

• Spectral imaging is crucial for plant phenotyping, particularly for assessing corn leaf nitrogen status.
• Spatial variations within multispectral images offer stronger signals for nitrogen estimation, but current technologies lack high resolution and ease of use for large leaf segments.

Purpose of the Study:

• To develop a proximal multispectral imaging device capable of capturing high-resolution images of large corn leaf segments.
• To improve the accuracy and efficiency of corn nitrogen estimation by addressing limitations of existing imaging technologies.

Main Methods:

• Developed a novel device using airflow for autonomous leaf positioning and flattening, employing a transmittance imaging regime for uniform lighting.
• Captured high-resolution, high-precision multispectral images of corn leaves within six seconds.
• Conducted a field assay with six nitrogen treatments, collecting 10 images per plot and analyzing spatial-spectral features.

Main Results:

• Analysis using average vegetative index showed limited differentiation between nitrogen treatments.
• Extracting and combining spatial and spectral features significantly improved nitrogen treatment classification accuracy.
• Spatial-spectral analysis methods further enhanced classification accuracy compared to the average index.

Conclusions:

• The developed high-resolution, high-throughput imaging device offers significant advantages for distinguishing nitrogen treatments in corn.
• Combined spatial-spectral analysis of detailed leaf images enables more precise corn nitrogen status classification.
• This technology facilitates improved crop management through accurate phenotyping.