Remote sensing of woolly croton using manned and unmanned aerial imaging systems - A feasibility study

Authors: Yang, Chenghai; Suh, Charles; Eyster, Ritchie; HONG, JIANG - Fuzhou University; HAMONS, KRISTIN - Texas A&M University

Submitted to: National Cotton Council Beltwide Cotton Conference
Publication Type: Proceedings
Publication Acceptance Date: 3/5/2019
Publication Date: 5/20/2019
Citation: Yang, C., Suh, C.P., Eyster, R.S., Hong, J., Hamons, K. 2019. Remote sensing of woolly croton using manned and unmanned aerial imaging systems - A feasibility study. National Cotton Council Beltwide Cotton Conference. pp. 250.254.

Interpretive Summary: Woolly croton is an annual invasive weed that serves as an overwintering host for the cotton fleahopper. It is therefore important to determine the distribution and density of woolly croton for the management of the insect pest. This study evaluated the feasibility to detect and map woolly croton infestations using aerial imagery captured from two imaging systems on an unmanned aircraft system (UAS) and an imaging system on a manned aircraft. Aerial images from the three imaging systems over a pasture infested with woolly croton were mosaicked separately and then classified to distinguish woolly croton from associated plant species. Preliminary image classification results showed that high resolution imagery from the unmanned and manned aircraft has the potential for mapping woolly croton infestations, though more research is needed to develop effective remote sensing methods and image classification techniques for practical application.

Technical Abstract: Woolly croton is an annual invasive weed that serves as a host for the cotton fleahopper. It is therefore important to determine the distribution and density of woolly croton for the management of the insect pest. The objective of this study was to evaluate the feasibility to detect and map woolly croton infestations using aerial imagery captured from manned and unmanned aircraft. An unmanned aircraft system (UAS) and a manned aircraft were used for image acquisition from a pasture infested with woolly croton. The UAS was equipped with a pair of red-green-blue (RGB) and near-infrared (NIR) cameras and a five-band camera, while the manned aircraft carried a different pair of RGB and NIR cameras. UAS images were acquired at 400 ft above ground level (AGL), while manned aircraft images were captured at approximately 2200 ft. The two types of multispectral images from the UAS were mosaicked, respectively. The two mosaicked UAS images and a single multispectral image from the manned aircraft were then classified, and the three classification maps were compared for the identification of woolly croton infestations. Preliminary accuracy assessment based on training samples used for image classification showed that high resolution aerial imagery from both manned and unmanned aircraft has the potential for mapping woolly croton infestations, though other classification techniques need to be evaluated with more thorough and robust accuracy assessment.