Estimation of Chlorophyll, Nitrogen, and Magnesium in Green Leaf by a Computer Vision Based in-Situ Biosensing Device: Validation and Interim Analysis

Leafstalk is a compact (327 gm of weight), battery-operated in situ biosensing device to estimate chlorophyll, nitrogen, and magnesium concentrations in green leaves. It has the following components—an LED, an optical sensor, a controlled holder maintaining a fixed distance from the LED and the sensor to place the blade of the leaf, the computing electronics, and a display. It utilizes computer vision-based techniques and machine learning-based models to analyze the light reflected from the leaves. The result of the analysis in terms of indices is shown on the device display. The device is connected to the cloud so that the estimated values along with plant/leaf identifier can be sent to a remote server through the internet. The device provides the calibration features to deliver the best indices for specific types of plants. The performance of Leafstalk has been compared with the Minolta SPAD meter on 1000 data points from 40 leaf samples. The findings revealed a precision with the coefficient of variation (within  ± 3%) and a strong positive correlation coefficient of 0.81 with p-value  <  0.00001. The results underscore Leafstalk’s potential as a valuable tool for researchers, agronomists, and farmers, enhancing their ability to remotely monitor and optimize plant growth, stress responses, and crop yields.