Abstract

Smartphones are an as yet untapped resource available to agriculture. They are ubiquitous across the globe yet have not previously been tested as a resource available to farmers. Imaging methods such as unmanned aerial vehicles (UAV) and satellite imaging have been well-explored and employed in various aspects of agriculture; however, such methods can be cost-prohibitive and at the mercy of another company or agency. If smartphones could be shown to capture color in such a way that relates in a quantifiable way to data measured by laboratory-grade equipment they could prove to be extremely valuable to farmers. Cutting out expensive and specialized technology for a device already sitting in people’s pockets would benefit farmers around the world. Given this idea, three experiments were designed to assess the color capabilities of smartphone cameras in relation to agricultural applications. The first experiment assessed the capability of smartphone cameras to identify the presence of cyanobacteria in a given water sample based on measurements of color and transmission spectra. These data were then related to color captured by four smartphones. Additionally, the measurements were used to create a preliminary customized Color Checker(TM)-inspired chart for use in identification of cyanobacteria. Current techniques employed by the state of New York for identifying cyanobacteria in water are cumbersome, involving week-long testing in government labs. This project is an attempt to simplify the process by using image capture with smartphones. The second assessment was similar to the first, with tomatoes in place of cyanobacteria. Five smartphone devices were used to image tomatoes at different stages of ripeness. A relationship was found to exist between the hue angles taken from the smartphone images and as measured by a spectroradiometer. A tomato Color Checker(TM) was created using the spectroradiometer measurements. The chart is intended for use in camera calibration for future imaging of tomatoes. The final assessment was an online experiment, wherein participants were asked to choose a color from an array generated from images of tomatoes that best represent the color of the tomato. This was a first step toward understanding which characteristics people use to categorize a crop as ripe and how those characteristics are rendered by smartphone imaging.

Publication Date

12-2021

Document Type

Dissertation

Student Type

Graduate

Degree Name

Color Science (Ph.D.)

Advisor

Susan Farnand

Advisor/Committee Member

Ben Zwickl

Advisor/Committee Member

Mark D. Fairchild

Campus

RIT – Main Campus

Share

COinS