Quality Assurance of Hyperspectral Imaging Systems for Plant Phenotyping
Author Information
Author(s): Justus Detring, Abel Barreto, Anne-Katrin Mahlein, Stefan Paulus
Primary Institution: Institute of Sugar Beet Research, Göttingen, Niedersachsen, Germany
Hypothesis
This research proposes an easy to apply quality assurance pipeline for hyperspectral imaging (HSI) systems used for plant phenotyping.
Conclusion
The quality assurance pipeline was successfully applied to evaluate a handheld HSI-system, enabling reliable disease severity estimation of Cercospora leaf spot in sugar beet plants.
Supporting Evidence
- The measurements of the calibration material were highly correlated with those of the non-imaging spectrometer (r>0.99).
- The CNN achieved a precision of over 98% for background and healthy classes.
- The study emphasizes the importance of evenly distributed diffuse illumination for HSI.
- The quality assurance pipeline can be easily applied to any other HSI-system to ensure high quality HSI.
- The CNN pipeline showed a low false positive rate for disease severity results of the non-inoculated variant.
Takeaway
This study created a way to check if hyperspectral imaging systems work well for studying plant diseases, helping farmers know when their plants are sick.
Methodology
The study evaluated the spatial and spectral accuracy of a handheld hyperspectral imaging system using various quality assurance methods, including the sine-wave-based spatial frequency response (s-SFR) and comparisons with a non-imaging spectrometer.
Potential Biases
Potential biases may arise from the reliance on human expert ratings for disease severity assessment.
Limitations
The tested system showed shortcomings in image resolution, sharpness, and illumination.
Participant Demographics
The study involved two sugar beet varieties, with 14 plants split into inoculated and non-inoculated groups.
Statistical Information
P-Value
p<0.05
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website