Skip to main content
ARS Home » Midwest Area » Columbia, Missouri » Cropping Systems and Water Quality Research » Research » Publications at this Location » Publication #166806
Title: REMOTE SENSING OF SOILS: USING AIRBORNE HYPERSPECTRAL IMAGES TO ESTIMATE WITHIN-FIELD VARIATIONS IN SOIL PROPERTIES

Author
item HONG, SUK YOUNG - NIAST, S KOREA
item Sudduth, Kenneth - Ken
item Kitchen, Newell
item WIEBOLD, WILLIAM - U OF MO
item PALM, HARLAN - U OF MO
item RIM, SANG KYU - NIAST, S KOREA
item SONN, YEON KYU - NIAST, S KOREA
item KWAK, HAN KANG - NIAST, S KOREA

Submitted to: Meeting Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 7/2/2004
Publication Date: 9/14/2004
Citation: Hong, S.Y., Sudduth, K.A., Kitchen, N.R., Wiebold, W.J., Palm, H.L., Rim, S.K., Sonn, Y.K., Kwak, H.K. 2004. Remote sensing of soils: using airborne hyperspectral images to estimate within-field variations in soil properties. In: Proceedings Global Digital Soil Mapping Workshop, September 14-17, 2004, Montpellier, France. CD-ROM.

Interpretive Summary:

Technical Abstract: The ability of hyperspectral image (HSI) data to estimate soil chemical property, soil texture, and bulk soil electrical conductivity (ECa) levels without requiring extensive field data collection was investigated. Bare soil images were acquired using a prism grating pushbroom scanner in April 2000, May 2001, and June 2002 for a central Missouri experimental field in a minimum-tillage corn-soybean rotation. Data were converted to reflectance using chemically-treated reference tarps with known reflectance levels. Geometric distortion of the pushbroom sensor images was corrected with a rubber sheeting transformation. A 5-m image pixel size was used, based on analysis of short-range variations in five sub-field areas. Statistical analyses--correlation, principal component analysis (PCA), stepwise multiple linear regression (SMLR), and multiple regression (MR)--were used to relate HSI data and derived Landsat-like bands (LLBs) to field-measured soil properties. Blue wavelengths of the HSI and Landsat-like images showed the strongest correlations with soil properties. Clay, organic matter, exchangeable cations (Ca, Mg, K), cation exchange capacity (CEC), and ECa were negatively correlated with reflectance. SMLR models using HSI were more predictive of field-measured soil properties than were MR models using LLBs, demonstrating the value of HSI. However, the relationships between soil properties and Landsat-like images were still quite good, and may be more acceptable for practical application, considering data volume, efficiency, and overfitting concerns. Both creating LLBs and applying PCA reduced the volume of HSI data and showed potential for developing relationships with soil property variability. Bare soil images of dry soil (2000 and 2002) were more useful than moist soil data for estimating soil chemical properties and ECa. Moist soil reflectance (2001) was more strongly related to soil texture than was dry soil data. We conclude that HSI, calibrated with soil samples obtained in the fields of interest, is a promising approach for quantifying soil property variability.