Skip to main content
ARS Home » Plains Area » College Station, Texas » Southern Plains Agricultural Research Center » Insect Control and Cotton Disease Research » Research » Publications at this Location » Publication #302329

Title: Airborne multispectral detection of regrowth cotton fields

Author
item Westbrook, John
item Suh, Charles
item Yang, Chenghai
item Lan, Yubin
item Eyster, Ritchie

Submitted to: Journal of Applied Remote Sensing (JARS)
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/27/2015
Publication Date: 2/17/2015
Citation: Westbrook, J.K., Suh, C.P., Yang, C., Lan, Y., Eyster, R.S. 2015. Airborne multispectral detection of regrowth cotton fields. Journal of Applied Remote Sensing (JARS). 9(1):096081.

Interpretive Summary: Cotton stalks are shredded after harvest to remove potential reproductive habitat for boll weevils. However, shredded cotton plants may develop new growth that can supply a food and reproductive resource for boll weevils between crop production cycles. Timely areawide detection of regrowth cotton plants is critically needed to expedite eradication in south Texas. We acquired airborne multispectral images of experimental plots that contained various developmental stages and sizes of regrowth cotton plants. Airborne multispectral and ground-based hyperspectral reflectance measurements of cotton plants and soil were analyzed using the linear spectral unmixing technique to identify ‘pure’ image pixels of cotton and ‘fuzzy’ image pixels of cotton mixed with soil. The capability to accurately detect cotton plants from medium- or high-resolution images will result in earlier detection for subsequent management of regrowth cotton plants on an areawide basis.

Technical Abstract: Regrowth of cotton, Gossypium hirsutum L., can provide boll weevils, Anthonomus grandis Boheman, with an extended opportunity to feed and reproduce beyond the production season. Effective methods for timely areawide detection of these potential host plants are critically needed to achieve eradication in south Texas, where regrowth plants may persist year-round due to the sub-tropical climate. We acquired sequential airborne multispectral images of experimental cotton plots to quantify the size of regrowth cotton plants. Airborne multispectral and ground-based hyperspectral reflectance measurements of cotton plants and soil were analyzed using the linear spectral unmixing technique to identify ‘pure’ image pixels of cotton and mixed image pixels of cotton and soil. Linear spectral unmixing estimates of canopy cover of regrowth cotton plots correlated well (r2 = 0.81, F=596, p<0.0001) with the ratio of mean plant width to row spacing. The height and width of regrowth plants were each well correlated with accumulated degree-days (DD15.6 C) after shredding (r2 = 0.94, F=1002, p<0.0001 for height; r2 = 0.94, F=900, p<0.0001 for width). Application of airborne multispectral imagery and a degree-day cotton regrowth model could improve areawide assessments of the distribution of regrowth cotton plants. The capability to accurately detect cotton plants from high-resolution images, and possibly medium-resolution images, could result in earlier detection and subsequent management of regrowth cotton plants on an areawide basis.