1 of 9

LED Model for Optical Sensor Detection

Adé Kolade (Mechanical Engineering)

Dr. Cherie Kagan, Electrical and Systems Engineering

1

This material is based upon work supported by the IoT4Ag Engineering Research Center funded by the National Science Foundation (NSF) under NSF Cooperative Agreement Number EEC-1941529. Any opinions, findings and conclusions, or recommendations expressed in this material are those of the author(s), and do not necessarily reflect those of the NSF.

IoT4Ag

JULY 28, 2021

REU PRESENTATION

2 of 9

Introduction and Problem Statement

2

Colorimetric optical sensors can

Directly measure plant stress indicators (e.g. plant moisture, temperature)
Be read by existing drone imaging systems

IoT4Ag

JULY 28, 2021

REU PRESENTATION

Current industrial ag practices typically employ crop imaging but lack direct information on plant stress indicators. That is, they can see from the photos when a group of plants is dead, but not when they are starting to exhibit warning signs. The purpose of this project is to codesign optical sensors and the methods for optical detection and data collection.

The picture in the middle is of our gold nanopillar resonator array (SEM images shows the circular gold nanopillars). You can see the diffraction of light in that middle image – colors of blue, green, red come from light diffracting due to the periodic structure. This sample was made with nanoimprint lithography, there are some imperfections as you can see the holes in the sample, which we are aiming to address. These sensors can be made to be 1.5”x1.5”, but the samples we make with electron beam lithography are much smaller (1mm x 1mm).

Precision agriculture (PA) is an approach to farm management that uses information technology ( IT) to ensure that crops and soil receive exactly what they need for optimum health and productivity. The goal of PA is to ensure profitability, sustainability and protection of the environment.

Once upon a time, there were several industrial farms that the world relied on to produce a great amount of our food supply.

One day, scientists proved that among other issues with industrial farming methods, 10-40% of all crops were lost to pests and pathogens.

At first, they established aerial and satellite crop imaging as a common practice to monitor crops. After a while, however, they realized simple imaging does not provide direct information about the plants’ health.

Then, it was hypothesized that colorimetric optical sensors, that is, physical sensors that change color, could enhance the existing optical crop monitoring with direct and specific information about crop health.

Finally, one morning, scientists produced a proof of concept for a metasurface hydrogel

3 of 9

Overview of the Technical Approach

3

A proxy sensor test bed has been used to assess conditions under which optical signals from the sensor can be detected. Initially, LEDs have been used as a proxy for the metasurface sensors:

IoT4Ag

JULY 28, 2021

REU PRESENTATION

4 of 9

Overview of the Technical Approach

4

Took images using a tripod to vary height, and a robot arm to vary angle:

IoT4Ag

JULY 28, 2021

REU PRESENTATION

5 of 9

Results: Image Processing

5

Height: 57.5 in

Lights On

Blink period: 100ms

Height: 27.5 in

Lights On

Blink period: 100ms

Image processed using Matlab to find RGB values

IoT4Ag

JULY 28, 2021

REU PRESENTATION

6 of 9

Summary

Positioning of LEDs on the plant and coverage by leaves affected the images. This is important as it helps determine the mode of detection - type of camera (RGB vs IR) and type of robot (ground vs aerial)

Image Processing:

Images that were processed were cropped to minimize noise, which may not be done in practical use.
Green LED appeared ‘more green’ when camera was closer to the plant.
‘Noise’ from background leaves did not seem to change significantly as a function of camera height.
This shows that camera/detector height is critical to accurate data acquisition and processing.

6

IoT4Ag

JULY 28, 2021

REU PRESENTATION

7 of 9

Future Work

Fixed LED position

Plant and LEDs moved over time, creating slight inconsistencies

Greater variance of camera position

Would like to vary over 20-30 feet

Spectrometer

Direct spectroscopic information (as opposed to RGB images)

Image processing

Use more advanced processing techniques for analysis

Use passive optical metasurfaces

7

IoT4Ag

JULY 28, 2021

REU PRESENTATION

8 of 9

Thank you!

Questions?

8

This material is based upon work supported by the IoT4Ag Engineering Research Center funded by the National Science Foundation (NSF) under NSF Cooperative Agreement Number EEC-1941529. Any opinions, findings and conclusions, or recommendations expressed in this material are those of the author(s), and do not necessarily reflect those of the NSF.

IoT4Ag

JULY 28, 2021

REU PRESENTATION

9 of 9

Backup Slides: Matlab Code

9

IoT4Ag

JULY 28, 2021

REU PRESENTATION