Correction of hyperspectral reflectance measurements for surface objects and direct sun reflection on surface waters

• Published in: International journal of remote sensing Vol. 34; no. 19; pp. 6651 - 6667
• Main Authors: Busch, Julia A, Hedley, John D, Zielinski, Oliver
• Format: Journal Article
• Language: English
• Published: Abingdon Taylor & Francis 2013
• Subjects: absorption; algorithms; Animal, plant and microbial ecology; Applied geophysics; Automated; Biological and medical sciences; boats; data collection; Data points; Earth sciences; Earth, ocean, space; Exact sciences and technology; Fundamental and applied biological sciences. Psychology; General aspects. Techniques; Glint; Hydrology; Hydrology. Hydrogeology; Internal geophysics; oxygen; Reflectance; Reflection; Reflectivity; remote sensing; spatial data; Sun; surface water; Teledetection and vegetation maps; MED, Western Mediterranean; spectra; absorption; Space remote sensing; near infrared radiation; Curve; satellite measurements; subsurface; platforms; surface water; automation; depth; corrections; Spatial data; regression; Hyperspectral characteristic; transient phenomena; slope angle; Mediterranean region; methodology; Reflectance
• ISSN: 1366-5901; 0143-1161; 1366-5901
• DOI: 10.1080/01431161.2013.804226

Satellite, airborne, or platform-based remote sensing reflectance measurements of aquatic targets are frequently compromised by water-surface effects such as specular sun reflection (glint) or transient objects like buoys or boats. For temporal or spatial data series where sub-surface reflectance is of interest, the elimination of affected data may require time-consuming manual selection of spectra and substantial data loss. Here, we present a method for the automated elimination of data points containing surface objects or strong sun reflection, which is based on the spectral slope in the ultra-violet to blue (350 nm to 450 nm). To minimize data loss, an automated sun glint correction combining two previously published methods is also presented. The method operates by subtracting a glint spectrum by means of a regression curve characterized from low to medium glint data points and is further automated by selecting these low glint data on the basis of the oxygen absorption depth in the near infrared (NIR). The elimination and correction algorithms facilitate rapid automated processing of large bio-optical data sets for both spatial and temporally resolved remote-sensing reflectance data sets. Here we demonstrate their efficacy on a three-month data set of hourly light field measurements from a fixed platform in the northwest Mediterranean.