Skip to main content
ARS Home » Southeast Area » Stoneville, Mississippi » Cotton Ginning Research » Research » Publications at this Location » Publication #369467
Research Project: Cotton Ginning Research to Improve Processing Efficiency and Product Quality in the Saw-Ginning of Picker-Harvested Cotton

Location: Cotton Ginning Research

Title: UAV-based multispectral detection of plastic debris in cotton fields

Author
item BLAKE, CODY - Oak Ridge Institute For Science And Education (ORISE)
item Sui, Ruixiu
item Yang, Chenghai

Submitted to: Meeting Abstract
Publication Type: Abstract Only
Publication Acceptance Date: 1/9/2020
Publication Date: 1/22/2020
Citation: Blake, C., Sui, R., Yang, C. 2020. UAV-based multispectral detection of plastic debris in cotton fields. Meeting Abstract. 560-564.

Interpretive Summary:

Technical Abstract: The US cotton industry has major concerns for plastic contamination in the harvesting, ginning and spinning portions of industry. The goal of this study was to detect plastic debris in cotton field using multispectral cameras on a UAV-based remote sensing technology. Plastic debris with different colors and sizes were introduced into a cotton field. Images of the field were acquired using multispectral cameras equipped on the UAV throughout the growing season to better understand the effect of plant canopy development on finding of plastic debris. The images were processed using Pix4Dmapper software. At early stage of this study, identifying the plastics relied heavily on visual observation of the images, which leaves a human error for false identification without multiple checks. Different colors in the plastics, such as tans and yellows, posed an issue for detection with the similarity of soil color and spectral response indices. While the results of the preliminary experiments were promising for detecting the plastic debris in a few colors, and the possibility of such use for the UAV-based sensing system could be beneficial, more work is required to create a computer program, or learning code, designed to distinguish between crop, ground and plastic debris reflectance signatures.