Skip to main content
ARS Home » Plains Area » College Station, Texas » Southern Plains Agricultural Research Center » Aerial Application Technology Research » Research » Publications at this Location » Publication #337861
Title: Comparison of aerial imagery from manned and unmanned aircraft platforms for monitoring cotton growth

Author
item Yang, Chenghai
item LANDIVAR, JUAN - Texas A&M Agrilife
item MURILO, MAEDA - Texas A&M Agrilife
item JUNG, JINHA - Texas A&M University
item STAREK, MICHAEL - Texas A&M University
item CHU, TIANXING - Texas A&M University
item CHANG, ANJIN - Texas A&M University

Submitted to: National Cotton Council Beltwide Cotton Conference
Publication Type: Proceedings
Publication Acceptance Date: 1/30/2017
Publication Date: 5/1/2017
Citation: Yang, C., Landivar, J., Murilo, M., Jung, J., Starek, M., Chu, T., Chang, A. 2017. Comparison of aerial imagery from manned and unmanned aircraft platforms for monitoring cotton growth. National Cotton Council Beltwide Cotton Conference. pp. 361-367. CD-ROM.

Interpretive Summary: Unmanned aircraft systems (UAS) have emerged as a low-cost and versatile remote sensing platform in recent years, but little work has been done on comparing imagery from manned and unmanned platforms for crop assessment. This study evaluated imagery taken from two cameras on a manned aircraft and three cameras on three UAS platforms for monitoring cotton growth. Comparison of the images from the manned and unmanned platforms revealed that some subtle color differences between cotton genotypes may have been smoothed out in the mosaicked UAS imagery. Correlation analysis showed that the UAS normal color imagery was similar to that from the manned platform, but the UAS color-infrared imagery was different from that from the manned platform, indicating the low-cost color-infrared UAS cameras may not have appropriate spectral response for remote sensing. This preliminary work has identified some of the key issues that need to be addressed on the use of UAS for remote sensing.

Technical Abstract: Unmanned aircraft systems (UAS) have emerged as a low-cost and versatile remote sensing platform in recent years, but little work has been done on comparing imagery from manned and unmanned platforms for crop assessment. The objective of this study was to compare imagery taken from multiple cameras on manned and unmanned aircraft platforms for monitoring cotton growth. A manned aircraft equipped with a red-green-blue (RGB) camera, a modified near-infrared (NIR) camera, a thermal camera and a hyperspectral camera was used to take images from test plots. Two rotary wing UAS equipped with a RGB camera and a modified color-infrared (CIR) camera, respectively, and a fixed wing UAS equipped with a different modified CIR camera were used to acquire high resolution images from the plots. The RGB/NIR/CIR imagery taken by both the manned and unmanned platforms on June 23, 2016 was used. The RGB/NIR/CIR images from all the platforms were registered to each other and the images from the unmanned platforms were aggregated and resampled to the coarser resolution in the imagery from the manned platform. Correlation coefficients were calculated among all the visible and NIR bands. Normalized difference vegetation index (NDVI) images were also calculated from all the CIR images. Comparison of the RGB images from the manned and unmanned platforms revealed that some subtle color differences between cotton genotypes could have been smoothed out in the mosaicked UAS imagery possibly due to color balancing. Correlation analysis showed that the visible bands in the RGB imagery and the NIR band in the CIR imagery from the unmanned platforms were positively related to the respective bands in the RGB and NIR imagery from the manned platform. However, the red and green bands in the CIR imagery from the UAS platforms were poorly related to the respective bands in the RGB imagery from the manned platform. The preliminary results from this study indicate that the modified CIR cameras may not have appropriate spectral response in the visible bands and that some useful spectral differences may have been smoothed out in mosaicked UAS imagery.